[ { "ID": "6.0-0001", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "87xMI355X_22xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5439 ROUGE2: 22.1342 ROUGEL: 28.9124 TOKENS_PER_SAMPLE: 283.0", "Nodes": 11, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x87)", "a#": 8, "Total Accelerators": 88, "Location": "./closed/AMD/results/87xMI355X_22xEPYC_9575F/llama2-70b-99/Interactive/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/87xMI355X_22xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/87xMI355X_22xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 785522.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0001", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "87xMI355X_22xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.426 ROUGE2: 22.0612 ROUGEL: 28.834 TOKENS_PER_SAMPLE: 284.3", "Nodes": 11, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x87)", "a#": 8, "Total Accelerators": 88, "Location": "./closed/AMD/results/87xMI355X_22xEPYC_9575F/llama2-70b-99/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/87xMI355X_22xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/87xMI355X_22xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 1042110.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0001", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "87xMI355X_22xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5433 ROUGE2: 22.1372 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 11, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x87)", "a#": 8, "Total Accelerators": 88, "Location": "./closed/AMD/results/87xMI355X_22xEPYC_9575F/llama2-70b-99/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/87xMI355X_22xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/87xMI355X_22xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 1016380.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0001", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "87xMI355X_22xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5439 ROUGE2: 22.1342 ROUGEL: 28.9124 TOKENS_PER_SAMPLE: 283.0", "Nodes": 11, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x87)", "a#": 8, "Total Accelerators": 88, "Location": "./closed/AMD/results/87xMI355X_22xEPYC_9575F/llama2-70b-99.9/Interactive/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/87xMI355X_22xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/87xMI355X_22xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 785522.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0001", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "87xMI355X_22xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.426 ROUGE2: 22.0612 ROUGEL: 28.834 TOKENS_PER_SAMPLE: 284.3", "Nodes": 11, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x87)", "a#": 8, "Total Accelerators": 88, "Location": "./closed/AMD/results/87xMI355X_22xEPYC_9575F/llama2-70b-99.9/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/87xMI355X_22xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/87xMI355X_22xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 1042110.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0001", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "87xMI355X_22xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5433 ROUGE2: 22.1372 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 11, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x87)", "a#": 8, "Total Accelerators": 88, "Location": "./closed/AMD/results/87xMI355X_22xEPYC_9575F/llama2-70b-99.9/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/87xMI355X_22xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/87xMI355X_22xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 1016380.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.396", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/gpt-oss-120b/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 95004.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.813", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/gpt-oss-120b/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 82136.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5411 ROUGE2: 22.1345 ROUGEL: 28.9133 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/llama2-70b-99/Interactive/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 73608.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4495 ROUGE2: 22.0868 ROUGEL: 28.8592 TOKENS_PER_SAMPLE: 284.2", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/llama2-70b-99/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 103480.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5371 ROUGE2: 22.1319 ROUGEL: 28.9082 TOKENS_PER_SAMPLE: 283.0", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/llama2-70b-99/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 100282.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5411 ROUGE2: 22.1345 ROUGEL: 28.9133 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/llama2-70b-99.9/Interactive/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 73608.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4495 ROUGE2: 22.0868 ROUGEL: 28.8592 TOKENS_PER_SAMPLE: 284.2", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/llama2-70b-99.9/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 103480.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0002", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5371 ROUGE2: 22.1319 ROUGEL: 28.9082 TOKENS_PER_SAMPLE: 283.0", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/AMD/results/8xMI355X_2xEPYC_9575F/llama2-70b-99.9/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/8xMI355X_2xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 100282.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0003", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "94xMI355X_24xEPYC_9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.889", "Nodes": 12, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x94)", "a#": 8, "Total Accelerators": 96, "Location": "./closed/AMD/results/94xMI355X_24xEPYC_9575F/gpt-oss-120b/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/94xMI355X_24xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/94xMI355X_24xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 1031070.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0003", "Submitter": "AMD", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AMD MI355X Custom Cluster", "Platform": "94xMI355X_24xEPYC_9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 82.888", "Nodes": 12, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (x94)", "a#": 8, "Total Accelerators": 96, "Location": "./closed/AMD/results/94xMI355X_24xEPYC_9575F/gpt-oss-120b/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/results/94xMI355X_24xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/closed/AMD/94xMI355X_24xEPYC_9575F", "ColorKey": "availableAMD", "Performance_Result": 900054.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0004", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.7596 ROUGE2: 22.3266 ROUGEL: 29.1163 TOKENS_PER_SAMPLE: 272.8", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 6238.12, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0004", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7099 ROUGE2: 22.3189 ROUGEL: 29.083 TOKENS_PER_SAMPLE: 274.6", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 27812.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0004", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.765 ROUGE2: 22.33 ROUGEL: 29.1337 TOKENS_PER_SAMPLE: 274.9", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 28038.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0004", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.77 ROUGE2: 16.0196 ROUGEL: 24.5357 ROUGELSUM: 35.6853 GEN_LEN: 8178793", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 34241.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0004", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7666 ROUGE2: 15.9806 ROUGEL: 24.5188 ROUGELSUM: 35.6677 GEN_LEN: 8148259", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 48791.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0004", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7519 ROUGE2: 16.0012 ROUGEL: 24.5188 ROUGELSUM: 35.6782 GEN_LEN: 8178012", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 48531.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0004", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.83967947314827", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000A_E13P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 18338.3, "Performance_Units": "Samples/s" }, { "ID": "6.0-0005", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.8215 ROUGE2: 22.3657 ROUGEL: 29.1382 TOKENS_PER_SAMPLE: 272.1", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 6262.63, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0005", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7344 ROUGE2: 22.3064 ROUGEL: 29.0956 TOKENS_PER_SAMPLE: 274.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 27730.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0005", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8535 ROUGE2: 22.4326 ROUGEL: 29.2034 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 27113.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0005", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7575 ROUGE2: 16.0094 ROUGEL: 24.5303 ROUGELSUM: 35.6835 GEN_LEN: 8178438", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 48613.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0005", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8", "Platform": "ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.75 ROUGE2: 15.9983 ROUGEL: 24.5119 ROUGELSUM: 35.6724 GEN_LEN: 8178586", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "Data bandwidth for GPU-PCIe: 504 GB/s; PCIe-NIC: 126 GB/s", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/ESC8000_E12P_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 47581.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0006", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "XA NB3I-E12", "Platform": "XA_NB3I-E12_B300x8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Interactive", "Accuracy": "exact_match: 82.996", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT/gpt-oss-120b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/XA_NB3I-E12_B300x8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 26005.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0006", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "XA NB3I-E12", "Platform": "XA_NB3I-E12_B300x8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.151", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/XA_NB3I-E12_B300x8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 110077.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0006", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "XA NB3I-E12", "Platform": "XA_NB3I-E12_B300x8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.515", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/XA_NB3I-E12_B300x8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 100656.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0006", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "XA NB3I-E12", "Platform": "XA_NB3I-E12_B300x8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.799 ROUGE2: 22.3898 ROUGEL: 29.1767 TOKENS_PER_SAMPLE: 272.8", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/XA_NB3I-E12_B300x8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 109625.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0006", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "XA NB3I-E12", "Platform": "XA_NB3I-E12_B300x8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8309 ROUGE2: 22.3869 ROUGEL: 29.1766 TOKENS_PER_SAMPLE: 272.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/XA_NB3I-E12_B300x8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 102186.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0006", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "XA NB3I-E12", "Platform": "XA_NB3I-E12_B300x8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.4844 ROUGE2: 15.929 ROUGEL: 24.3386 ROUGELSUM: 35.6648 GEN_LEN: 8186789", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/XA_NB3I-E12_B300x8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 166745.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0006", "Submitter": "ASUSTeK", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "XA NB3I-E12", "Platform": "XA_NB3I-E12_B300x8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.4837 ROUGE2: 15.9227 ROUGEL: 24.3333 ROUGELSUM: 35.6577 GEN_LEN: 8185499", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/results/XA_NB3I-E12_B300x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ASUSTeK/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ASUSTeK/XA_NB3I-E12_B300x8_TRT", "ColorKey": "availableASUSTeK", "Performance_Result": 148067.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0007", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "C240M8-1-node-2S-GNR_86C", "Platform": "C240M8_1_node_2S_GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.8227 ROUGE2: 16.0784 ROUGEL: 24.5847 ROUGELSUM: 35.8974 GEN_LEN: 8148551", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Cisco/results/C240M8_1_node_2S_GNR_86C/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C240M8_1_node_2S_GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C240M8_1_node_2S_GNR_86C", "ColorKey": "availableCisco", "Performance_Result": 806.105, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0007", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "C240M8-1-node-2S-GNR_86C", "Platform": "C240M8_1_node_2S_GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.8579 ROUGE2: 16.0896 ROUGEL: 24.6038 ROUGELSUM: 35.9322 GEN_LEN: 8145938", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Cisco/results/C240M8_1_node_2S_GNR_86C/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C240M8_1_node_2S_GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C240M8_1_node_2S_GNR_86C", "ColorKey": "availableCisco", "Performance_Result": 258.538, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0007", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "C240M8-1-node-2S-GNR_86C", "Platform": "C240M8_1_node_2S_GNR_86C", "UsedModel": "rgat", "Model": "rgat", "Scenario": "Offline", "Accuracy": "acc: 71.803", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Cisco/results/C240M8_1_node_2S_GNR_86C/rgat/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C240M8_1_node_2S_GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C240M8_1_node_2S_GNR_86C", "ColorKey": "availableCisco", "Performance_Result": 14283.5, "Performance_Units": "Samples/s" }, { "ID": "6.0-0007", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "C240M8-1-node-2S-GNR_86C", "Platform": "C240M8_1_node_2S_GNR_86C", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.88576267638467", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Cisco/results/C240M8_1_node_2S_GNR_86C/whisper/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C240M8_1_node_2S_GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C240M8_1_node_2S_GNR_86C", "ColorKey": "availableCisco", "Performance_Result": 1422.99, "Performance_Units": "Samples/s" }, { "ID": "6.0-0008", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C845A M8 (8x H200-NVL-141GB, TensorRT)", "Platform": "C845A_M8_H200NVL_141GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.82730759347906", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C845A_M8_H200NVL_141GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "H200-NVL TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C845A_M8_H200NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C845A_M8_H200NVL_141GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 30195.2, "Performance_Units": "Samples/s" }, { "ID": "6.0-0009", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C845A M8 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "C845A_M8_RTXPro6000_96GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.81874244601579", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C845A_M8_RTXPro6000_96GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C845A_M8_RTXPro6000_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C845A_M8_RTXPro6000_96GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 18265.2, "Performance_Units": "Samples/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.22151321786691 TOKENS_PER_SAMPLE: 3770.239744758432", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 69251.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.42661804922516 TOKENS_PER_SAMPLE: 3726.4544211485872", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 58553.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 82.846", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 111496.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.393", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 110655.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7978 ROUGE2: 22.3733 ROUGEL: 29.1557 TOKENS_PER_SAMPLE: 273.1", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 115730.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8431 ROUGE2: 22.4093 ROUGEL: 29.1757 TOKENS_PER_SAMPLE: 272.7", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 114504.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.82114944226222 exact_match: 89.9662276637988 TOKENS_PER_SAMPLE: 637.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 1971.11, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Server", "Accuracy": "ROUGEL: 21.83555157124159 exact_match: 89.91636002647253 TOKENS_PER_SAMPLE: 659.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/llama3.1-405b/Server/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 1445.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 70.1617", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/wan-2.2-t2v-a14b/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 0.0573123, "Performance_Units": "Samples/s" }, { "ID": "6.0-0010", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)", "Platform": "C880A_M8_B300_288GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 70.3040", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/C880A_M8_B300_288GBx8_TRT/wan-2.2-t2v-a14b/SingleStream/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/C880A_M8_B300_288GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/C880A_M8_B300_288GBx8_TRT", "ColorKey": "availableCisco", "Performance_Result": 21.425919695, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0011", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS 16xMI350X", "Platform": "Cisco_UCS_16xMI350x", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.6085 ROUGE2: 22.1816 ROUGEL: 28.9644 TOKENS_PER_SAMPLE: 281.7", "Nodes": 2, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 16, "Location": "./closed/Cisco/results/Cisco_UCS_16xMI350x/llama2-70b-99/Interactive/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/Cisco_UCS_16xMI350x", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/Cisco_UCS_16xMI350x", "ColorKey": "availableCisco", "Performance_Result": 103633.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0011", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS 16xMI350X", "Platform": "Cisco_UCS_16xMI350x", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4637 ROUGE2: 22.1149 ROUGEL: 28.8866 TOKENS_PER_SAMPLE: 283.3", "Nodes": 2, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 16, "Location": "./closed/Cisco/results/Cisco_UCS_16xMI350x/llama2-70b-99/Offline/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/Cisco_UCS_16xMI350x", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/Cisco_UCS_16xMI350x", "ColorKey": "availableCisco", "Performance_Result": 159932.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0011", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS 16xMI350X", "Platform": "Cisco_UCS_16xMI350x", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.6075 ROUGE2: 22.1813 ROUGEL: 28.9635 TOKENS_PER_SAMPLE: 281.8", "Nodes": 2, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 16, "Location": "./closed/Cisco/results/Cisco_UCS_16xMI350x/llama2-70b-99/Server/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/Cisco_UCS_16xMI350x", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/Cisco_UCS_16xMI350x", "ColorKey": "availableCisco", "Performance_Result": 153022.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0012", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS 8xMI350X", "Platform": "Cisco_UCS_8xMI350x", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.6101 ROUGE2: 22.1826 ROUGEL: 28.9649 TOKENS_PER_SAMPLE: 281.8", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/Cisco_UCS_8xMI350x/llama2-70b-99/Interactive/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/Cisco_UCS_8xMI350x", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/Cisco_UCS_8xMI350x", "ColorKey": "availableCisco", "Performance_Result": 51616.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0012", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS 8xMI350X", "Platform": "Cisco_UCS_8xMI350x", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.467 ROUGE2: 22.1014 ROUGEL: 28.8752 TOKENS_PER_SAMPLE: 283.8", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/Cisco_UCS_8xMI350x/llama2-70b-99/Offline/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/Cisco_UCS_8xMI350x", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/Cisco_UCS_8xMI350x", "ColorKey": "availableCisco", "Performance_Result": 78936.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0012", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS 8xMI350X", "Platform": "Cisco_UCS_8xMI350x", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.6074 ROUGE2: 22.1814 ROUGEL: 28.9634 TOKENS_PER_SAMPLE: 281.8", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Cisco/results/Cisco_UCS_8xMI350x/llama2-70b-99/Server/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/Cisco_UCS_8xMI350x", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/Cisco_UCS_8xMI350x", "ColorKey": "availableCisco", "Performance_Result": 76617.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0013", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X210 M8 - X580P PCIe Node (4x H200-NVL-141GB, TensorRT)", "Platform": "X210M8_X580P_H200NVL_141GBx4_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7476 ROUGE2: 15.8938 ROUGEL: 24.4606 ROUGELSUM: 35.7751 GEN_LEN: 8209954", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X210M8_X580P_H200NVL_141GBx4_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "H200-NVL TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X210M8_X580P_H200NVL_141GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X210M8_X580P_H200NVL_141GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 24216.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0013", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X210 M8 - X580P PCIe Node (4x H200-NVL-141GB, TensorRT)", "Platform": "X210M8_X580P_H200NVL_141GBx4_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7502 ROUGE2: 15.8913 ROUGEL: 24.457 ROUGELSUM: 35.7785 GEN_LEN: 8209110", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X210M8_X580P_H200NVL_141GBx4_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "H200-NVL TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X210M8_X580P_H200NVL_141GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X210M8_X580P_H200NVL_141GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 25731.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0013", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X210 M8 - X580P PCIe Node (4x H200-NVL-141GB, TensorRT)", "Platform": "X210M8_X580P_H200NVL_141GBx4_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.82350086127316", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X210M8_X580P_H200NVL_141GBx4_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "H200-NVL TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X210M8_X580P_H200NVL_141GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X210M8_X580P_H200NVL_141GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 15338.6, "Performance_Units": "Samples/s" }, { "ID": "6.0-0014", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X215c M8 - X580P PCIe Node (4x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "X215M8_X580P_RTXPro6000_96GBx4_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7674 ROUGE2: 22.3465 ROUGEL: 29.1537 TOKENS_PER_SAMPLE: 273.2", "Nodes": 1, "Processor": "AMD EPYC 9845", "host_processors_per_node": 2, "host_processor_core_count": "160", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X215M8_X580P_RTXPro6000_96GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 13529.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0014", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X215c M8 - X580P PCIe Node (4x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "X215M8_X580P_RTXPro6000_96GBx4_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7585 ROUGE2: 22.3383 ROUGEL: 29.1509 TOKENS_PER_SAMPLE: 273.7", "Nodes": 1, "Processor": "AMD EPYC 9845", "host_processors_per_node": 2, "host_processor_core_count": "160", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X215M8_X580P_RTXPro6000_96GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 13029.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0014", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X215c M8 - X580P PCIe Node (4x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "X215M8_X580P_RTXPro6000_96GBx4_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5868 ROUGE2: 15.859 ROUGEL: 24.3111 ROUGELSUM: 35.538 GEN_LEN: 8181337", "Nodes": 1, "Processor": "AMD EPYC 9845", "host_processors_per_node": 2, "host_processor_core_count": "160", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X215M8_X580P_RTXPro6000_96GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 24013.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0014", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X215c M8 - X580P PCIe Node (4x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "X215M8_X580P_RTXPro6000_96GBx4_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.5862 ROUGE2: 15.8584 ROUGEL: 24.3129 ROUGELSUM: 35.5376 GEN_LEN: 8181285", "Nodes": 1, "Processor": "AMD EPYC 9845", "host_processors_per_node": 2, "host_processor_core_count": "160", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X215M8_X580P_RTXPro6000_96GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 23423.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0014", "Submitter": "Cisco", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Cisco UCS X215c M8 - X580P PCIe Node (4x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "X215M8_X580P_RTXPro6000_96GBx4_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.81683907991282", "Nodes": 1, "Processor": "AMD EPYC 9845", "host_processors_per_node": 2, "host_processor_core_count": "160", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/results/X215M8_X580P_RTXPro6000_96GBx4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Cisco/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Cisco/X215M8_X580P_RTXPro6000_96GBx4_TRT", "ColorKey": "availableCisco", "Performance_Result": 9099.08, "Performance_Units": "Samples/s" }, { "ID": "6.0-0015", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB200 NVL72 (64x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x64_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.54056517775751 TOKENS_PER_SAMPLE: 3759.982224247949", "Nodes": 16, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 64, "Location": "./closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 468665.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0015", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB200 NVL72 (64x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x64_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.26709206927985 TOKENS_PER_SAMPLE: 3818.1679580674568", "Nodes": 16, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 64, "Location": "./closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 311531.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0015", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB200 NVL72 (64x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x64_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.760", "Nodes": 16, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 64, "Location": "./closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 774539.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0015", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB200 NVL72 (64x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x64_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.085", "Nodes": 16, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 64, "Location": "./closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB200-NVL72_GB200-186GB_aarch64x64_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 743550.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0016", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB300 NVL72 (64x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x64_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.871", "Nodes": 16, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 64, "Location": "./closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x64_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x64_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB300-NVL72_GB300-288GB_aarch64x64_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 910468.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0016", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB300 NVL72 (64x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x64_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.043", "Nodes": 16, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 64, "Location": "./closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x64_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x64_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB300-NVL72_GB300-288GB_aarch64x64_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 956523.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0017", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB300 NVL72 (8x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 80.94804010938924 TOKENS_PER_SAMPLE: 3815.0455788514128", "Nodes": 2, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 8, "Location": "./closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 78564.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0017", "Submitter": "CoreWeave", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "CoreWeave GB300 NVL72 (8x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.17593436645396 TOKENS_PER_SAMPLE: 3780.3126709206927", "Nodes": 2, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 8, "Location": "./closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/closed/CoreWeave/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "ColorKey": "availableCoreWeave", "Performance_Result": 44589.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0018", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.8253 ROUGE2: 16.0656 ROUGEL: 24.58 ROUGELSUM: 35.9055 GEN_LEN: 8148260", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Dell/results/1-node-2S-GNR_86C/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04 LTS", "Notes": "Dell PowerEdge R770. INT8", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/1-node-2S-GNR_86C", "ColorKey": "availableDell", "Performance_Result": 807.675, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0018", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.8223 ROUGE2: 16.0515 ROUGEL: 24.5808 ROUGELSUM: 35.9033 GEN_LEN: 8145003", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Dell/results/1-node-2S-GNR_86C/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04 LTS", "Notes": "Dell PowerEdge R770. INT8", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/1-node-2S-GNR_86C", "ColorKey": "availableDell", "Performance_Result": 245.63, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0018", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "rgat", "Model": "rgat", "Scenario": "Offline", "Accuracy": "acc: 72.540", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Dell/results/1-node-2S-GNR_86C/rgat/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04 LTS", "Notes": "Dell PowerEdge R770. INT8", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/1-node-2S-GNR_86C", "ColorKey": "availableDell", "Performance_Result": 14283.7, "Performance_Units": "Samples/s" }, { "ID": "6.0-0018", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.88576267638467", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Dell/results/1-node-2S-GNR_86C/whisper/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04 LTS", "Notes": "Dell PowerEdge R770. INT8", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/1-node-2S-GNR_86C", "ColorKey": "availableDell", "Performance_Result": 1414.26, "Performance_Units": "Samples/s" }, { "ID": "6.0-0019", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9680 (8x MI300X)", "Platform": "8xMI300X_2xPlatinum_8470", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4421 ROUGE2: 22.0546 ROUGEL: 28.6036 TOKENS_PER_SAMPLE: 301.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8470", "host_processors_per_node": 2, "host_processor_core_count": "52", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI300X_2xPlatinum_8470/llama2-70b-99/Offline/performance/run_1", "Software": "vLLM 0.9.0.2.dev108+g71faa1880.d20260213, Pytorch 2.7.0+gitf717b2a, ROCm 6.4.1", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI300X_2xPlatinum_8470", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI300X_2xPlatinum_8470", "ColorKey": "availableDell", "Performance_Result": 27072.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0019", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9680 (8x MI300X)", "Platform": "8xMI300X_2xPlatinum_8470", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.4296 ROUGE2: 22.0707 ROUGEL: 28.6339 TOKENS_PER_SAMPLE: 301.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8470", "host_processors_per_node": 2, "host_processor_core_count": "52", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI300X_2xPlatinum_8470/llama2-70b-99/Server/performance/run_1", "Software": "vLLM 0.9.0.2.dev108+g71faa1880.d20260213, Pytorch 2.7.0+gitf717b2a, ROCm 6.4.1", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI300X_2xPlatinum_8470", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI300X_2xPlatinum_8470", "ColorKey": "availableDell", "Performance_Result": 24467.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0019", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9680 (8x MI300X)", "Platform": "8xMI300X_2xPlatinum_8470", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4421 ROUGE2: 22.0546 ROUGEL: 28.6036 TOKENS_PER_SAMPLE: 301.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8470", "host_processors_per_node": 2, "host_processor_core_count": "52", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI300X_2xPlatinum_8470/llama2-70b-99.9/Offline/performance/run_1", "Software": "vLLM 0.9.0.2.dev108+g71faa1880.d20260213, Pytorch 2.7.0+gitf717b2a, ROCm 6.4.1", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI300X_2xPlatinum_8470", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI300X_2xPlatinum_8470", "ColorKey": "availableDell", "Performance_Result": 27072.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0019", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9680 (8x MI300X)", "Platform": "8xMI300X_2xPlatinum_8470", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.4296 ROUGE2: 22.0707 ROUGEL: 28.6339 TOKENS_PER_SAMPLE: 301.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8470", "host_processors_per_node": 2, "host_processor_core_count": "52", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI300X_2xPlatinum_8470/llama2-70b-99.9/Server/performance/run_1", "Software": "vLLM 0.9.0.2.dev108+g71faa1880.d20260213, Pytorch 2.7.0+gitf717b2a, ROCm 6.4.1", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI300X_2xPlatinum_8470", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI300X_2xPlatinum_8470", "ColorKey": "availableDell", "Performance_Result": 24467.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.251", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/gpt-oss-120b/Offline/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 91395.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.122", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/gpt-oss-120b/Server/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 78603.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5403 ROUGE2: 22.1342 ROUGEL: 28.9108 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/llama2-70b-99/Interactive/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 70610.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.458 ROUGE2: 22.0859 ROUGEL: 28.8585 TOKENS_PER_SAMPLE: 284.2", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/llama2-70b-99/Offline/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 100008.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5361 ROUGE2: 22.132 ROUGEL: 28.9076 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/llama2-70b-99/Server/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 96734.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5403 ROUGE2: 22.1342 ROUGEL: 28.9108 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/llama2-70b-99.9/Interactive/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 70610.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.458 ROUGE2: 22.0859 ROUGEL: 28.8585 TOKENS_PER_SAMPLE: 284.2", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/llama2-70b-99.9/Offline/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 100008.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0020", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785L (8x MI355X)", "Platform": "8xMI355X_2xEPYC_9655", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5361 ROUGE2: 22.132 ROUGEL: 28.9076 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/8xMI355X_2xEPYC_9655/llama2-70b-99.9/Server/performance/run_1", "Software": "vLLM 0.14.0rc3.dev30+gb026cf14e.d20260212.rocm700, Pytorch 2.9.1+git8907517, ROCm 7.0.0.70000-38~22.04 ", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "MI355X TGP 1400W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/8xMI355X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/8xMI355X_2xEPYC_9655", "ColorKey": "availableDell", "Performance_Result": 96734.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5695 ROUGE2: 22.1191 ROUGEL: 28.7481 TOKENS_PER_SAMPLE: 291.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 16343.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.5303 ROUGE2: 22.094 ROUGEL: 28.7029 TOKENS_PER_SAMPLE: 293.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 32004.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5359 ROUGE2: 22.1023 ROUGEL: 28.7072 TOKENS_PER_SAMPLE: 294.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 29084.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5695 ROUGE2: 22.1191 ROUGEL: 28.7481 TOKENS_PER_SAMPLE: 291.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama2-70b-99.9/Interactive/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 16343.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.5303 ROUGE2: 22.094 ROUGEL: 28.7029 TOKENS_PER_SAMPLE: 293.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 32004.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5359 ROUGE2: 22.1023 ROUGEL: 28.7072 TOKENS_PER_SAMPLE: 294.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 29084.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7788 ROUGE2: 16.0073 ROUGEL: 24.5878 ROUGELSUM: 35.8159 GEN_LEN: 8207624", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 53193.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.8383 ROUGE2: 16.0476 ROUGEL: 24.5941 ROUGELSUM: 35.8728 GEN_LEN: 8204099", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 45075.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0021", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x H200-NVL-141GB, TensorRT)", "Platform": "XE7740_H200_NVL_141GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.8625198663837", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA H200-NVL-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.13, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "H200 TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_H200_NVL_141GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_H200_NVL_141GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 31969.2, "Performance_Units": "Samples/s" }, { "ID": "6.0-0022", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7740_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8818 ROUGE2: 22.4053 ROUGEL: 29.202 TOKENS_PER_SAMPLE: 272.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 27034.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0022", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7740_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8742 ROUGE2: 22.4134 ROUGEL: 29.2014 TOKENS_PER_SAMPLE: 272.8", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 25524.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0022", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7740_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8818 ROUGE2: 22.4053 ROUGEL: 29.202 TOKENS_PER_SAMPLE: 272.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 27034.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0022", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7740_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8742 ROUGE2: 22.4134 ROUGEL: 29.2014 TOKENS_PER_SAMPLE: 272.8", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 25524.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0022", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7740 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7740_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.85205135281745", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7740_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 18231.5, "Performance_Units": "Samples/s" }, { "ID": "6.0-0023", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7745 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7745_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8777 ROUGE2: 22.408 ROUGEL: 29.1982 TOKENS_PER_SAMPLE: 272.0", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 26829.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0023", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7745 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7745_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8697 ROUGE2: 22.4115 ROUGEL: 29.1961 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 25536.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0023", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7745 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7745_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8777 ROUGE2: 22.408 ROUGEL: 29.1982 TOKENS_PER_SAMPLE: 272.0", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 26829.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0023", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7745 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7745_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8697 ROUGE2: 22.4115 ROUGEL: 29.1961 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 25536.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0023", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7745 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7745_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7694 ROUGE2: 16.0138 ROUGEL: 24.531 ROUGELSUM: 35.6934 GEN_LEN: 8179110", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 49036.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0023", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7745 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7745_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7655 ROUGE2: 16.0173 ROUGEL: 24.5235 ROUGELSUM: 35.6909 GEN_LEN: 8178701", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 45008.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0023", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE7745 (8x RTXPro6000-PCIe-96GB, TensorRT)", "Platform": "XE7745_RTXPro6000_PCIe_96GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.86156818333222", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA RTX PRO 6000 Blackwell Server Edition TGP 600W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE7745_RTXPro6000_PCIe_96GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 18433.5, "Performance_Units": "Samples/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.63172288058341 TOKENS_PER_SAMPLE: 3764.0123062898815", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 66287.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.35824977210574 TOKENS_PER_SAMPLE: 3716.6032360984505", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 46510.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.8183 ROUGE2: 22.3646 ROUGEL: 29.1643 TOKENS_PER_SAMPLE: 274.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 70724.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7649 ROUGE2: 22.3597 ROUGEL: 29.1343 TOKENS_PER_SAMPLE: 272.1", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 108669.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7575 ROUGE2: 22.354 ROUGEL: 29.1306 TOKENS_PER_SAMPLE: 273.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 108392.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.8183 ROUGE2: 22.3646 ROUGEL: 29.1643 TOKENS_PER_SAMPLE: 274.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/llama2-70b-99.9/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 70724.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7649 ROUGE2: 22.3597 ROUGEL: 29.1343 TOKENS_PER_SAMPLE: 272.1", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 108669.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0024", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L-B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7575 ROUGE2: 22.354 ROUGEL: 29.1306 TOKENS_PER_SAMPLE: 273.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L-B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L-B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 108392.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0025", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780L (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780L_x86_B300-SXM-270GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.84063115619975", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6979P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780L_x86_B300-SXM-270GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780L_x86_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780L_x86_B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 45658.8, "Performance_Units": "Samples/s" }, { "ID": "6.0-0026", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780 (8x B200-SXM-180GB, TensorRT)", "Platform": "XE9780_x86_B200-SXM-180GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.1531449407475 TOKENS_PER_SAMPLE: 3727.768231540565", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780_x86_B200-SXM-180GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780_x86_B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780_x86_B200-SXM-180GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 58235.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0026", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780 (8x B200-SXM-180GB, TensorRT)", "Platform": "XE9780_x86_B200-SXM-180GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.0619872379216 TOKENS_PER_SAMPLE: 3741.5695077484047", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780_x86_B200-SXM-180GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780_x86_B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780_x86_B200-SXM-180GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 40949.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0027", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780 (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780_x86_B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.13035551504102 TOKENS_PER_SAMPLE: 3753.5875113947127", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780_x86_B300-SXM-270GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780_x86_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780_x86_B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 69021.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0027", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9780 (8x B300-SXM-270GB, TensorRT)", "Platform": "XE9780_x86_B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.13035551504102 TOKENS_PER_SAMPLE: 3712.944165907019", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE9780_x86_B300-SXM-270GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE9780_x86_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell/XE9780_x86_B300-SXM-270GBx8_TRT", "ColorKey": "availableDell", "Performance_Result": 46513.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0028", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Dell Pro Max (1x GB10)", "Platform": "Dell_Pro_Max_GB10x1", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "MultiStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "Cortex-X925", "host_processors_per_node": 20, "host_processor_core_count": "1", "Accelerator": "NVIDIA GB10", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Dell/results/Dell_Pro_Max_GB10x1/yolo-95/multistream/performance/run_1", "Software": "CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "Dell Pro Max with GB10. CPU only submission", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 20, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/Dell_Pro_Max_GB10x1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/Dell_Pro_Max_GB10x1", "ColorKey": "availableDell", "Performance_Result": 4891.380208, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0028", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Dell Pro Max (1x GB10)", "Platform": "Dell_Pro_Max_GB10x1", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "Offline", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "Cortex-X925", "host_processors_per_node": 20, "host_processor_core_count": "1", "Accelerator": "NVIDIA GB10", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Dell/results/Dell_Pro_Max_GB10x1/yolo-95/offline/performance/run_1", "Software": "CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "Dell Pro Max with GB10. CPU only submission", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 20, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/Dell_Pro_Max_GB10x1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/Dell_Pro_Max_GB10x1", "ColorKey": "availableDell", "Performance_Result": 1.74887, "Performance_Units": "Samples/s" }, { "ID": "6.0-0028", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Dell Pro Max (1x GB10)", "Platform": "Dell_Pro_Max_GB10x1", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "SingleStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "Cortex-X925", "host_processors_per_node": 20, "host_processor_core_count": "1", "Accelerator": "NVIDIA GB10", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Dell/results/Dell_Pro_Max_GB10x1/yolo-95/singlestream/performance/run_1", "Software": "CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "Dell Pro Max with GB10. CPU only submission", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 20, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/Dell_Pro_Max_GB10x1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/Dell_Pro_Max_GB10x1", "ColorKey": "availableDell", "Performance_Result": 611.926848, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0028", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Dell Pro Max (1x GB10)", "Platform": "Dell_Pro_Max_GB10x1", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "MultiStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "Cortex-X925", "host_processors_per_node": 20, "host_processor_core_count": "1", "Accelerator": "NVIDIA GB10", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Dell/results/Dell_Pro_Max_GB10x1/yolo-99/multistream/performance/run_1", "Software": "CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "Dell Pro Max with GB10. CPU only submission", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 20, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/Dell_Pro_Max_GB10x1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/Dell_Pro_Max_GB10x1", "ColorKey": "availableDell", "Performance_Result": 4891.380208, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0028", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Dell Pro Max (1x GB10)", "Platform": "Dell_Pro_Max_GB10x1", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "Offline", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "Cortex-X925", "host_processors_per_node": 20, "host_processor_core_count": "1", "Accelerator": "NVIDIA GB10", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Dell/results/Dell_Pro_Max_GB10x1/yolo-99/offline/performance/run_1", "Software": "CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "Dell Pro Max with GB10. CPU only submission", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 20, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/Dell_Pro_Max_GB10x1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/Dell_Pro_Max_GB10x1", "ColorKey": "availableDell", "Performance_Result": 1.74887, "Performance_Units": "Samples/s" }, { "ID": "6.0-0028", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Dell Pro Max (1x GB10)", "Platform": "Dell_Pro_Max_GB10x1", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "SingleStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "Cortex-X925", "host_processors_per_node": 20, "host_processor_core_count": "1", "Accelerator": "NVIDIA GB10", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Dell/results/Dell_Pro_Max_GB10x1/yolo-99/singlestream/performance/run_1", "Software": "CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "Dell Pro Max with GB10. CPU only submission", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 20, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/Dell_Pro_Max_GB10x1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/Dell_Pro_Max_GB10x1", "ColorKey": "availableDell", "Performance_Result": 611.926848, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0029", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (8x L40S, TensorRT)", "Platform": "XE7745_L40Sx8", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "MultiStream", "Accuracy": "mAP: 53.116", "Nodes": 1, "Processor": " AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA L40S", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_L40Sx8/yolo-95/multistream/performance/run_1", "Software": "TensorRT 10.14, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "L40S TGP 350W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L40Sx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L40Sx8", "ColorKey": "availableDell", "Performance_Result": 54.674812, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0029", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (8x L40S, TensorRT)", "Platform": "XE7745_L40Sx8", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "Offline", "Accuracy": "mAP: 53.116", "Nodes": 1, "Processor": " AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA L40S", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_L40Sx8/yolo-95/offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "L40S TGP 350W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L40Sx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L40Sx8", "ColorKey": "availableDell", "Performance_Result": 153.544, "Performance_Units": "Samples/s" }, { "ID": "6.0-0029", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (8x L40S, TensorRT)", "Platform": "XE7745_L40Sx8", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "SingleStream", "Accuracy": "mAP: 53.116", "Nodes": 1, "Processor": " AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA L40S", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_L40Sx8/yolo-95/singlestream/performance/run_1", "Software": "TensorRT 10.14, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "L40S TGP 350W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L40Sx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L40Sx8", "ColorKey": "availableDell", "Performance_Result": 7.124808, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0029", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (8x L40S, TensorRT)", "Platform": "XE7745_L40Sx8", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "MultiStream", "Accuracy": "mAP: 53.116", "Nodes": 1, "Processor": " AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA L40S", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_L40Sx8/yolo-99/multistream/performance/run_1", "Software": "TensorRT 10.14, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "L40S TGP 350W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L40Sx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L40Sx8", "ColorKey": "availableDell", "Performance_Result": 54.674812, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0029", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (8x L40S, TensorRT)", "Platform": "XE7745_L40Sx8", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "Offline", "Accuracy": "mAP: 53.116", "Nodes": 1, "Processor": " AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA L40S", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_L40Sx8/yolo-99/offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "L40S TGP 350W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L40Sx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L40Sx8", "ColorKey": "availableDell", "Performance_Result": 153.544, "Performance_Units": "Samples/s" }, { "ID": "6.0-0029", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (8x L40S, TensorRT)", "Platform": "XE7745_L40Sx8", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "SingleStream", "Accuracy": "mAP: 53.116", "Nodes": 1, "Processor": " AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA L40S", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell/results/XE7745_L40Sx8/yolo-99/singlestream/performance/run_1", "Software": "TensorRT 10.14, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "L40S TGP 350W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L40Sx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L40Sx8", "ColorKey": "availableDell", "Performance_Result": 7.124808, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0030", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (16x L4-PCIE-24GB, TensorRT)", "Platform": "XE7745_L4x16", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "MultiStream", "Accuracy": "mAP: 53.110", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 16, "Total Accelerators": 16, "Location": "./closed/Dell/results/XE7745_L4x16/yolo-95/multistream/performance/run_1", "Software": "CUDA 13.1", "operating_system": "Ubuntu 24.04.02", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L4x16", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L4x16", "ColorKey": "availableDell", "Performance_Result": 81.093026, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0030", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (16x L4-PCIE-24GB, TensorRT)", "Platform": "XE7745_L4x16", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "Offline", "Accuracy": "mAP: 53.110", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 16, "Total Accelerators": 16, "Location": "./closed/Dell/results/XE7745_L4x16/yolo-95/offline/performance/run_1", "Software": "CUDA 13.1", "operating_system": "Ubuntu 24.04.02", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L4x16", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L4x16", "ColorKey": "availableDell", "Performance_Result": 102.232, "Performance_Units": "Samples/s" }, { "ID": "6.0-0030", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (16x L4-PCIE-24GB, TensorRT)", "Platform": "XE7745_L4x16", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "SingleStream", "Accuracy": "mAP: 53.110", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 16, "Total Accelerators": 16, "Location": "./closed/Dell/results/XE7745_L4x16/yolo-95/singlestream/performance/run_1", "Software": "CUDA 13.1", "operating_system": "Ubuntu 24.04.02", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L4x16", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L4x16", "ColorKey": "availableDell", "Performance_Result": 10.54864, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0030", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (16x L4-PCIE-24GB, TensorRT)", "Platform": "XE7745_L4x16", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "MultiStream", "Accuracy": "mAP: 53.110", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 16, "Total Accelerators": 16, "Location": "./closed/Dell/results/XE7745_L4x16/yolo-99/multistream/performance/run_1", "Software": "CUDA 13.1", "operating_system": "Ubuntu 24.04.02", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L4x16", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L4x16", "ColorKey": "availableDell", "Performance_Result": 81.093026, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0030", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (16x L4-PCIE-24GB, TensorRT)", "Platform": "XE7745_L4x16", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "Offline", "Accuracy": "mAP: 53.110", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 16, "Total Accelerators": 16, "Location": "./closed/Dell/results/XE7745_L4x16/yolo-99/offline/performance/run_1", "Software": "CUDA 13.1", "operating_system": "Ubuntu 24.04.02", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L4x16", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L4x16", "ColorKey": "availableDell", "Performance_Result": 102.232, "Performance_Units": "Samples/s" }, { "ID": "6.0-0030", "Submitter": "Dell", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "PowerEdge XE7745 (16x L4-PCIE-24GB, TensorRT)", "Platform": "XE7745_L4x16", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "SingleStream", "Accuracy": "mAP: 53.110", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 16, "Total Accelerators": 16, "Location": "./closed/Dell/results/XE7745_L4x16/yolo-99/singlestream/performance/run_1", "Software": "CUDA 13.1", "operating_system": "Ubuntu 24.04.02", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/results/XE7745_L4x16", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell/src", "Unique ID (e.g. for Audit)": "edge/closed/Dell/XE7745_L4x16", "ColorKey": "availableDell", "Performance_Result": 10.54864, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0031", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Dell_MangoBoost Heterogeneous Cluster (8x MI300X + 8x MI325X + 8x MI355X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4975 ROUGE2: 22.0735 ROUGEL: 28.8141 TOKENS_PER_SAMPLE: 287.3", "Nodes": 3, "Processor": "AMD EPYC 9534 (on MI300X node), AMD EPYC 9655 (on MI325X node), AMD EPYC 9965 (on MI355X node)", "host_processors_per_node": 2, "host_processor_core_count": "64(AMD EPYC 9534) and 96(AMD EPYC 9655) and 192(AMD EPYC 9965)", "Accelerator": "AMD Instinct MI300X 192GB HBM3 (x8) and AMD Instinct MI325X 256GB HBM3e (x8) and AMD Instinct MI355X 288GB HBM3e (x8)", "a#": 8, "Total Accelerators": 24, "Location": "./closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965/llama2-70b-99/Offline/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 151843.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0031", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Dell_MangoBoost Heterogeneous Cluster (8x MI300X + 8x MI325X + 8x MI355X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.543 ROUGE2: 22.0969 ROUGEL: 28.8312 TOKENS_PER_SAMPLE: 287.0", "Nodes": 3, "Processor": "AMD EPYC 9534 (on MI300X node), AMD EPYC 9655 (on MI325X node), AMD EPYC 9965 (on MI355X node)", "host_processors_per_node": 2, "host_processor_core_count": "64(AMD EPYC 9534) and 96(AMD EPYC 9655) and 192(AMD EPYC 9965)", "Accelerator": "AMD Instinct MI300X 192GB HBM3 (x8) and AMD Instinct MI325X 256GB HBM3e (x8) and AMD Instinct MI355X 288GB HBM3e (x8)", "a#": 8, "Total Accelerators": 24, "Location": "./closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965/llama2-70b-99/Server/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 141521.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0031", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Dell_MangoBoost Heterogeneous Cluster (8x MI300X + 8x MI325X + 8x MI355X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4975 ROUGE2: 22.0735 ROUGEL: 28.8141 TOKENS_PER_SAMPLE: 287.3", "Nodes": 3, "Processor": "AMD EPYC 9534 (on MI300X node), AMD EPYC 9655 (on MI325X node), AMD EPYC 9965 (on MI355X node)", "host_processors_per_node": 2, "host_processor_core_count": "64(AMD EPYC 9534) and 96(AMD EPYC 9655) and 192(AMD EPYC 9965)", "Accelerator": "AMD Instinct MI300X 192GB HBM3 (x8) and AMD Instinct MI325X 256GB HBM3e (x8) and AMD Instinct MI355X 288GB HBM3e (x8)", "a#": 8, "Total Accelerators": 24, "Location": "./closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965/llama2-70b-99.9/Offline/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 151843.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0031", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Dell_MangoBoost Heterogeneous Cluster (8x MI300X + 8x MI325X + 8x MI355X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.543 ROUGE2: 22.0969 ROUGEL: 28.8312 TOKENS_PER_SAMPLE: 287.0", "Nodes": 3, "Processor": "AMD EPYC 9534 (on MI300X node), AMD EPYC 9655 (on MI325X node), AMD EPYC 9965 (on MI355X node)", "host_processors_per_node": 2, "host_processor_core_count": "64(AMD EPYC 9534) and 96(AMD EPYC 9655) and 192(AMD EPYC 9965)", "Accelerator": "AMD Instinct MI300X 192GB HBM3 (x8) and AMD Instinct MI325X 256GB HBM3e (x8) and AMD Instinct MI355X 288GB HBM3e (x8)", "a#": 8, "Total Accelerators": 24, "Location": "./closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965/llama2-70b-99.9/Server/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI300X_2xEPYC_9534_8xMI325X_2xEPYC_9655_8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 141521.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0032", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785", "Platform": "8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.5433 ROUGE2: 22.1372 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965/llama2-70b-99/Offline/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 97535.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0032", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785", "Platform": "8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5377 ROUGE2: 22.1338 ROUGEL: 28.9088 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965/llama2-70b-99/Server/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 93272.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0032", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785", "Platform": "8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.5433 ROUGE2: 22.1372 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965/llama2-70b-99.9/Offline/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 97535.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0032", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785", "Platform": "8xMI355X_2xEPYC_9965", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5377 ROUGE2: 22.1338 ROUGEL: 28.9088 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965/llama2-70b-99.9/Server/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 93272.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0033", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785 (MI355X Power Cap 1000W)", "Platform": "8xMI355X_2xEPYC_9965_Power_Cap_1000W", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.5433 ROUGE2: 22.1372 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (Power Cap 1000 W)", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W/llama2-70b-99/Offline/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 81870.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0033", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785 (MI355X Power Cap 1000W)", "Platform": "8xMI355X_2xEPYC_9965_Power_Cap_1000W", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5427 ROUGE2: 22.135 ROUGEL: 28.9148 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (Power Cap 1000 W)", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W/llama2-70b-99/Server/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 76359.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0033", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785 (MI355X Power Cap 1000W)", "Platform": "8xMI355X_2xEPYC_9965_Power_Cap_1000W", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.5433 ROUGE2: 22.1372 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (Power Cap 1000 W)", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W/llama2-70b-99.9/Offline/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 81870.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0033", "Submitter": "Dell_MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "PowerEdge XE9785 (MI355X Power Cap 1000W)", "Platform": "8xMI355X_2xEPYC_9965_Power_Cap_1000W", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5427 ROUGE2: 22.135 ROUGEL: 28.9148 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9965", "host_processors_per_node": 2, "host_processor_core_count": "192", "Accelerator": "AMD Instinct MI355X 288GB HBM3e (Power Cap 1000 W)", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W/llama2-70b-99.9/Server/performance/run_1", "Software": "ROCm", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/results/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Dell_MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Dell_MangoBoost/8xMI355X_2xEPYC_9965_Power_Cap_1000W", "ColorKey": "availableDell_MangoBoost", "Performance_Result": 76359.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0034", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Lenovo IdeaPad Gaming 3 15ARH05", "Platform": "Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "MultiStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "AMD Ryzen 5 4600H", "host_processors_per_node": 1, "host_processor_core_count": "6", "Accelerator": "NVIDIA GeForce GTX 1650 Ti", "a#": 1, "Total Accelerators": 1, "Location": "./closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126/yolo-95/multistream/performance/run_1", "Software": "pytorch v2.5.0a0", "operating_system": "Ubuntu 22.04 (linux-6.17.9-76061709-generic-glibc2.35)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "ColorKey": "availableGATEOverflow", "Performance_Result": 344.438863, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0034", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Lenovo IdeaPad Gaming 3 15ARH05", "Platform": "Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "Offline", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "AMD Ryzen 5 4600H", "host_processors_per_node": 1, "host_processor_core_count": "6", "Accelerator": "NVIDIA GeForce GTX 1650 Ti", "a#": 1, "Total Accelerators": 1, "Location": "./closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126/yolo-95/offline/performance/run_1", "Software": "pytorch v2.5.0a0", "operating_system": "Ubuntu 22.04 (linux-6.17.9-76061709-generic-glibc2.35)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "ColorKey": "availableGATEOverflow", "Performance_Result": 24.187, "Performance_Units": "Samples/s" }, { "ID": "6.0-0034", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Lenovo IdeaPad Gaming 3 15ARH05", "Platform": "Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "SingleStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "AMD Ryzen 5 4600H", "host_processors_per_node": 1, "host_processor_core_count": "6", "Accelerator": "NVIDIA GeForce GTX 1650 Ti", "a#": 1, "Total Accelerators": 1, "Location": "./closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126/yolo-95/singlestream/performance/run_1", "Software": "pytorch v2.5.0a0", "operating_system": "Ubuntu 22.04 (linux-6.17.9-76061709-generic-glibc2.35)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "ColorKey": "availableGATEOverflow", "Performance_Result": 43.150031, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0034", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Lenovo IdeaPad Gaming 3 15ARH05", "Platform": "Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "MultiStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "AMD Ryzen 5 4600H", "host_processors_per_node": 1, "host_processor_core_count": "6", "Accelerator": "NVIDIA GeForce GTX 1650 Ti", "a#": 1, "Total Accelerators": 1, "Location": "./closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126/yolo-99/multistream/performance/run_1", "Software": "pytorch v2.5.0a0", "operating_system": "Ubuntu 22.04 (linux-6.17.9-76061709-generic-glibc2.35)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "ColorKey": "availableGATEOverflow", "Performance_Result": 344.438863, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0034", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Lenovo IdeaPad Gaming 3 15ARH05", "Platform": "Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "Offline", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "AMD Ryzen 5 4600H", "host_processors_per_node": 1, "host_processor_core_count": "6", "Accelerator": "NVIDIA GeForce GTX 1650 Ti", "a#": 1, "Total Accelerators": 1, "Location": "./closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126/yolo-99/offline/performance/run_1", "Software": "pytorch v2.5.0a0", "operating_system": "Ubuntu 22.04 (linux-6.17.9-76061709-generic-glibc2.35)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "ColorKey": "availableGATEOverflow", "Performance_Result": 24.187, "Performance_Units": "Samples/s" }, { "ID": "6.0-0034", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "Lenovo IdeaPad Gaming 3 15ARH05", "Platform": "Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "SingleStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "AMD Ryzen 5 4600H", "host_processors_per_node": 1, "host_processor_core_count": "6", "Accelerator": "NVIDIA GeForce GTX 1650 Ti", "a#": 1, "Total Accelerators": 1, "Location": "./closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126/yolo-99/singlestream/performance/run_1", "Software": "pytorch v2.5.0a0", "operating_system": "Ubuntu 22.04 (linux-6.17.9-76061709-generic-glibc2.35)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/Lenovo_IdeaPad_Gaming_3_Ryzen_5_4600H-reference-gpu-pytorch_v2.5.0a0-cu126", "ColorKey": "availableGATEOverflow", "Performance_Result": 43.150031, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0035", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel Sapphire Rapids (2x RTX 4090)", "Platform": "d986d93afc0b-nvidia-gpu-TensorRT-default_config", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "MultiStream", "Accuracy": "acc: 76.064", "Nodes": 1, "Processor": "Intel(R) Xeon(R) w7-2495X", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "NVIDIA GeForce RTX 4090", "a#": 2, "Total Accelerators": 2, "Location": "./closed/GATEOverflow/results/d986d93afc0b-nvidia-gpu-TensorRT-default_config/resnet50/multistream/performance/run_1", "Software": "TensorRT", "operating_system": "Ubuntu 24.04 (linux-6.17.0-14-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/d986d93afc0b-nvidia-gpu-TensorRT-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/d986d93afc0b-nvidia-gpu-TensorRT-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 0.399625, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0035", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel Sapphire Rapids (2x RTX 4090)", "Platform": "d986d93afc0b-nvidia-gpu-TensorRT-default_config", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "Offline", "Accuracy": "acc: 76.078", "Nodes": 1, "Processor": "Intel(R) Xeon(R) w7-2495X", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "NVIDIA GeForce RTX 4090", "a#": 2, "Total Accelerators": 2, "Location": "./closed/GATEOverflow/results/d986d93afc0b-nvidia-gpu-TensorRT-default_config/resnet50/offline/performance/run_1", "Software": "TensorRT", "operating_system": "Ubuntu 24.04 (linux-6.17.0-14-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/d986d93afc0b-nvidia-gpu-TensorRT-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/d986d93afc0b-nvidia-gpu-TensorRT-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 87945.5, "Performance_Units": "Samples/s" }, { "ID": "6.0-0035", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel Sapphire Rapids (2x RTX 4090)", "Platform": "d986d93afc0b-nvidia-gpu-TensorRT-default_config", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "SingleStream", "Accuracy": "acc: 76.064", "Nodes": 1, "Processor": "Intel(R) Xeon(R) w7-2495X", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "NVIDIA GeForce RTX 4090", "a#": 2, "Total Accelerators": 2, "Location": "./closed/GATEOverflow/results/d986d93afc0b-nvidia-gpu-TensorRT-default_config/resnet50/singlestream/performance/run_1", "Software": "TensorRT", "operating_system": "Ubuntu 24.04 (linux-6.17.0-14-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/d986d93afc0b-nvidia-gpu-TensorRT-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/d986d93afc0b-nvidia-gpu-TensorRT-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 0.303116, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0036", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel SPR i9", "Platform": "intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "MultiStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "13th Gen Intel(R) Core(TM) i9-13900K", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config/yolo-95/multistream/performance/run_1", "Software": "pytorch v2.3.1", "operating_system": "Ubuntu 24.04 (linux-6.8.0-64-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 843.374268, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0036", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel SPR i9", "Platform": "intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "Offline", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "13th Gen Intel(R) Core(TM) i9-13900K", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config/yolo-95/offline/performance/run_1", "Software": "pytorch v2.3.1", "operating_system": "Ubuntu 24.04 (linux-6.8.0-64-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 12.0652, "Performance_Units": "Samples/s" }, { "ID": "6.0-0036", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel SPR i9", "Platform": "intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "UsedModel": "yolo-95", "Model": "yolo-95", "Scenario": "SingleStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "13th Gen Intel(R) Core(TM) i9-13900K", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config/yolo-95/singlestream/performance/run_1", "Software": "pytorch v2.3.1", "operating_system": "Ubuntu 24.04 (linux-6.8.0-64-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 90.105803, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0036", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel SPR i9", "Platform": "intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "MultiStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "13th Gen Intel(R) Core(TM) i9-13900K", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config/yolo-99/multistream/performance/run_1", "Software": "pytorch v2.3.1", "operating_system": "Ubuntu 24.04 (linux-6.8.0-64-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 843.374268, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0036", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel SPR i9", "Platform": "intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "Offline", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "13th Gen Intel(R) Core(TM) i9-13900K", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config/yolo-99/offline/performance/run_1", "Software": "pytorch v2.3.1", "operating_system": "Ubuntu 24.04 (linux-6.8.0-64-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 12.0652, "Performance_Units": "Samples/s" }, { "ID": "6.0-0036", "Submitter": "GATEOverflow", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "GATE Overflow Intel SPR i9", "Platform": "intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "UsedModel": "yolo-99", "Model": "yolo-99", "Scenario": "SingleStream", "Accuracy": "mAP: 53.108", "Nodes": 1, "Processor": "13th Gen Intel(R) Core(TM) i9-13900K", "host_processors_per_node": 1, "host_processor_core_count": "24", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config/yolo-99/singlestream/performance/run_1", "Software": "pytorch v2.3.1", "operating_system": "Ubuntu 24.04 (linux-6.8.0-64-generic-glibc2.39)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp32", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/results/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GATEOverflow/src", "Unique ID (e.g. for Audit)": "edge/closed/GATEOverflow/intel_spr_i9-reference-cpu-pytorch_v2.3.1-default_config", "ColorKey": "availableGATEOverflow", "Performance_Result": 90.105803, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.220", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/gpt-oss-120b/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 92332.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 82.853", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/gpt-oss-120b/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 80089.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5355 ROUGE2: 22.1293 ROUGEL: 28.91 TOKENS_PER_SAMPLE: 282.8", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/llama2-70b-99/Interactive/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 72547.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4514 ROUGE2: 22.0869 ROUGEL: 28.8592 TOKENS_PER_SAMPLE: 284.3", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/llama2-70b-99/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 100310.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5382 ROUGE2: 22.1338 ROUGEL: 28.9091 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/llama2-70b-99/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 98157.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5355 ROUGE2: 22.1293 ROUGEL: 28.91 TOKENS_PER_SAMPLE: 282.8", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/llama2-70b-99.9/Interactive/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 72547.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4514 ROUGE2: 22.0869 ROUGEL: 28.8592 TOKENS_PER_SAMPLE: 284.3", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/llama2-70b-99.9/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 100310.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0037", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G893-ZX1-AAX4", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5382 ROUGE2: 22.1338 ROUGEL: 28.9091 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/8xMI355X_2xEPYC-9575F/llama2-70b-99.9/Server/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0", "operating_system": "Ubuntu 22.04.5", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/8xMI355X_2xEPYC-9575F", "ColorKey": "availableGigaComputing", "Performance_Result": 98157.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0038", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "DLB2-CB3", "Platform": "DLB2-CB3_GB300-288GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7886741346065558", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/GigaComputing/results/DLB2-CB3_GB300-288GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/DLB2-CB3_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/DLB2-CB3_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 60.9525, "Performance_Units": "Samples/s" }, { "ID": "6.0-0038", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "DLB2-CB3", "Platform": "DLB2-CB3_GB300-288GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.7871305769336737", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/GigaComputing/results/DLB2-CB3_GB300-288GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/DLB2-CB3_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/DLB2-CB3_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 41.6342, "Performance_Units": "Queries/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Interactive", "Accuracy": "exact_match: 80.94804010938924 TOKENS_PER_SAMPLE: 3773.6034639927075", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/deepseek-r1/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 4934.65, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 80.67456700091158 TOKENS_PER_SAMPLE: 3756.651777575205", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 69617.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.22151321786691 TOKENS_PER_SAMPLE: 3733.6216955332725", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 35552.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.622", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 107696.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.855", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 106397.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.7219 ROUGE2: 22.309 ROUGEL: 29.0918 TOKENS_PER_SAMPLE: 275.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 65464.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7858 ROUGE2: 22.3212 ROUGEL: 29.1222 TOKENS_PER_SAMPLE: 276.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 109343.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7751 ROUGE2: 22.3149 ROUGEL: 29.1129 TOKENS_PER_SAMPLE: 277.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 104623.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.7219 ROUGE2: 22.309 ROUGEL: 29.0918 TOKENS_PER_SAMPLE: 275.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama2-70b-99.9/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 65464.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7858 ROUGE2: 22.3212 ROUGEL: 29.1222 TOKENS_PER_SAMPLE: 276.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 109343.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7751 ROUGE2: 22.3149 ROUGEL: 29.1129 TOKENS_PER_SAMPLE: 277.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 104623.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Interactive", "Accuracy": "ROUGEL: 21.835240799978806 exact_match: 90.10374586366643 TOKENS_PER_SAMPLE: 636.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama3.1-405b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 764.655, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 22.052524963008008 exact_match: 90.0969391131701 TOKENS_PER_SAMPLE: 652.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 1947.91, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Server", "Accuracy": "ROUGEL: 21.853317758658946 exact_match: 90.12565519523494 TOKENS_PER_SAMPLE: 636.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama3.1-405b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 1340.51, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.5225 ROUGE2: 15.8241 ROUGEL: 24.308 ROUGELSUM: 35.5704 GEN_LEN: 8194606", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 128633.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5224 ROUGE2: 15.8225 ROUGEL: 24.3072 ROUGELSUM: 35.5714 GEN_LEN: 8194322", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 150707.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.5265 ROUGE2: 15.8301 ROUGEL: 24.3121 ROUGELSUM: 35.5742 GEN_LEN: 8194725", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 141484.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 70.2730", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/wan-2.2-t2v-a14b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 0.0582849, "Performance_Units": "Samples/s" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 70.2024", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/wan-2.2-t2v-a14b/SingleStream/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 20.897894239, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0039", "Submitter": "GigaComputing", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G894-SD3-AAX7", "Platform": "G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.84824462061155", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 22.04.4", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/results/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/GigaComputing/src", "Unique ID (e.g. for Audit)": "datacenter/closed/GigaComputing/G894-SD3-AAX7_B300-SXM-270GBx8_TRT", "ColorKey": "availableGigaComputing", "Performance_Result": 48116.6, "Performance_Units": "Samples/s" }, { "ID": "6.0-0040", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.0619872379216 TOKENS_PER_SAMPLE: 3785.250455788514", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Google/results/B200-SXM-180GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/B200-SXM-180GBx8_TRT", "ColorKey": "availableGoogle", "Performance_Result": 57116.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0040", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.42661804922516 TOKENS_PER_SAMPLE: 3748.3705560619874", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Google/results/B200-SXM-180GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/B200-SXM-180GBx8_TRT", "ColorKey": "availableGoogle", "Performance_Result": 41027.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0040", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Offline", "Accuracy": "DLRM_NE: 0.8675504587396468 DLRM_ACC: 0.6964261896038318 DLRM_AUC: 0.7862891896349031", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Google/results/B200-SXM-180GBx8_TRT/dlrm-v3/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/B200-SXM-180GBx8_TRT", "ColorKey": "availableGoogle", "Performance_Result": 10447.3, "Performance_Units": "Samples/s" }, { "ID": "6.0-0040", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Server", "Accuracy": "DLRM_NE: 0.8675496448947934 DLRM_ACC: 0.6964279594720938 DLRM_AUC: 0.7862891760730971", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Google/results/B200-SXM-180GBx8_TRT/dlrm-v3/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/B200-SXM-180GBx8_TRT", "ColorKey": "availableGoogle", "Performance_Result": 10007.0, "Performance_Units": "Queries/s" }, { "ID": "6.0-0040", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.918853025027985 exact_match: 89.94308074123097 TOKENS_PER_SAMPLE: 654.5", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Google/results/B200-SXM-180GBx8_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/B200-SXM-180GBx8_TRT", "ColorKey": "availableGoogle", "Performance_Result": 1350.24, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0040", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Server", "Accuracy": "ROUGEL: 21.894180533777952 exact_match: 89.9182594308405 TOKENS_PER_SAMPLE: 654.1", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Google/results/B200-SXM-180GBx8_TRT/llama3.1-405b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/B200-SXM-180GBx8_TRT", "ColorKey": "availableGoogle", "Performance_Result": 983.626, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0040", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.85490640197189", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Google/results/B200-SXM-180GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/B200-SXM-180GBx8_TRT", "ColorKey": "availableGoogle", "Performance_Result": 49327.2, "Performance_Units": "Samples/s" }, { "ID": "6.0-0041", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (4x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7879394631035838", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "ColorKey": "availableGoogle", "Performance_Result": 42.4736, "Performance_Units": "Samples/s" }, { "ID": "6.0-0041", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (4x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.7874876598480153", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "ColorKey": "availableGoogle", "Performance_Result": 35.5625, "Performance_Units": "Queries/s" }, { "ID": "6.0-0041", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (4x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x4_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.8501479867145", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "ColorKey": "availableGoogle", "Performance_Result": 22051.4, "Performance_Units": "Samples/s" }, { "ID": "6.0-0042", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.22151321786691 TOKENS_PER_SAMPLE: 3723.8377392889697", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableGoogle", "Performance_Result": 307874.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0042", "Submitter": "Google", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 80.83409298085688 TOKENS_PER_SAMPLE: 3694.1697812215134", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Google/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Google/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableGoogle", "Performance_Result": 156781.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0043", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL145 Gen11 (1x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5879 ROUGE2: 15.8594 ROUGEL: 24.3118 ROUGELSUM: 35.5388 GEN_LEN: 8181331", "Nodes": 1, "Processor": "AMD EPYC 8434P", "host_processors_per_node": 1, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 1, "Total Accelerators": 1, "Location": "./closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT", "ColorKey": "availableHPE", "Performance_Result": 2340.75, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0043", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL145 Gen11 (1x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.587 ROUGE2: 15.8592 ROUGEL: 24.3112 ROUGELSUM: 35.538 GEN_LEN: 8181380", "Nodes": 1, "Processor": "AMD EPYC 8434P", "host_processors_per_node": 1, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 1, "Total Accelerators": 1, "Location": "./closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_DC_TRT", "ColorKey": "availableHPE", "Performance_Result": 2081.34, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0044", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Offline", "Accuracy": "ROUGE1: 45.5065 ROUGE2: 23.2835 ROUGEL: 30.3527 TOKENS_PER_SAMPLE: 144.3 gsm8k_accuracy: 73.92 mbxp_accuracy: 60.12", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT/mixtral-8x7b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 41242.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0044", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Server", "Accuracy": "ROUGE1: 45.4809 ROUGE2: 23.3483 ROUGEL: 30.4976 TOKENS_PER_SAMPLE: 145.7 gsm8k_accuracy: 73.86 mbxp_accuracy: 59.94", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT/mixtral-8x7b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 39234.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0044", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.86537491553813", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 18889.4, "Performance_Units": "Samples/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 84.096", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 19040.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.746", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 17737.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8722 ROUGE2: 22.4111 ROUGEL: 29.1997 TOKENS_PER_SAMPLE: 271.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 36710.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.87 ROUGE2: 22.409 ROUGEL: 29.197 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 35590.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8721 ROUGE2: 22.411 ROUGEL: 29.1995 TOKENS_PER_SAMPLE: 271.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 36704.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8738 ROUGE2: 22.4117 ROUGEL: 29.1998 TOKENS_PER_SAMPLE: 272.8", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 35522.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.7806 ROUGE2: 16.0229 ROUGEL: 24.5438 ROUGELSUM: 35.7089 GEN_LEN: 8179467", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 53054.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7667 ROUGE2: 16.0132 ROUGEL: 24.5324 ROUGELSUM: 35.6922 GEN_LEN: 8178916", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 60679.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0045", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7625 ROUGE2: 16.0093 ROUGEL: 24.5327 ROUGELSUM: 35.6872 GEN_LEN: 8179131", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 10, "Total Accelerators": 10, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx10_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 59632.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0046", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Offline", "Accuracy": "ROUGE1: 45.4424 ROUGE2: 23.2859 ROUGEL: 30.3282 TOKENS_PER_SAMPLE: 145.1 gsm8k_accuracy: 73.34 mbxp_accuracy: 60.28", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT/mixtral-8x7b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 33637.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0046", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Server", "Accuracy": "ROUGE1: 45.5238 ROUGE2: 23.3068 ROUGEL: 30.4805 TOKENS_PER_SAMPLE: 144.9 gsm8k_accuracy: 73.72 mbxp_accuracy: 59.96", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT/mixtral-8x7b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 32151.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.644", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 15189.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.633", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 14258.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8698 ROUGE2: 22.4084 ROUGEL: 29.1974 TOKENS_PER_SAMPLE: 271.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 29908.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.873 ROUGE2: 22.4095 ROUGEL: 29.199 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 28727.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8712 ROUGE2: 22.4093 ROUGEL: 29.1983 TOKENS_PER_SAMPLE: 271.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 29900.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8687 ROUGE2: 22.4095 ROUGEL: 29.1984 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 28717.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.7729 ROUGE2: 16.0189 ROUGEL: 24.5391 ROUGELSUM: 35.699 GEN_LEN: 8178708", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 44087.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7725 ROUGE2: 16.0187 ROUGEL: 24.5371 ROUGELSUM: 35.6975 GEN_LEN: 8178941", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 49580.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7506 ROUGE2: 16.0037 ROUGEL: 24.5269 ROUGELSUM: 35.6755 GEN_LEN: 8178736", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 48794.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0047", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.82254917822169", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "172", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_PROLIANT_DL380A_RTX_PRO_6000_PCIE_96GBx8_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 18709.0, "Performance_Units": "Samples/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.7586 ROUGE2: 22.3423 ROUGEL: 29.1305 TOKENS_PER_SAMPLE: 274.4", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 61644.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7474 ROUGE2: 22.3568 ROUGEL: 29.1367 TOKENS_PER_SAMPLE: 273.3", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 103954.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7537 ROUGE2: 22.361 ROUGEL: 29.1392 TOKENS_PER_SAMPLE: 274.3", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 103635.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.7627 ROUGE2: 22.343 ROUGEL: 29.1316 TOKENS_PER_SAMPLE: 274.4", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama2-70b-99.9/Interactive/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 61300.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7474 ROUGE2: 22.3568 ROUGEL: 29.1367 TOKENS_PER_SAMPLE: 273.3", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 104572.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7454 ROUGE2: 22.3541 ROUGEL: 29.1337 TOKENS_PER_SAMPLE: 274.3", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 103627.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.4119 ROUGE2: 15.7662 ROUGEL: 24.2994 ROUGELSUM: 35.4992 GEN_LEN: 8201032", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 126143.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.4165 ROUGE2: 15.7673 ROUGEL: 24.3028 ROUGELSUM: 35.5051 GEN_LEN: 8201311", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 152625.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.4148 ROUGE2: 15.7669 ROUGEL: 24.3024 ROUGELSUM: 35.5024 GEN_LEN: 8201083", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 131270.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Offline", "Accuracy": "ROUGE1: 45.4127 ROUGE2: 23.1305 ROUGEL: 30.2484 TOKENS_PER_SAMPLE: 144.4 gsm8k_accuracy: 73.42 mbxp_accuracy: 60.34", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/mixtral-8x7b/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 126468.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Server", "Accuracy": "ROUGE1: 45.4142 ROUGE2: 23.2853 ROUGEL: 30.2919 TOKENS_PER_SAMPLE: 146.0 gsm8k_accuracy: 73.02 mbxp_accuracy: 60.16", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/mixtral-8x7b/Server/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 127209.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0048", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x NVIDIA B200 180GB, TensorRT)", "Platform": "HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.8377761070453", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_B200_SXM_192GBx8_TRT", "ColorKey": "availableHPE", "Performance_Result": 46033.9, "Performance_Units": "Samples/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.347", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/gpt-oss-120b/Offline/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 91846.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 82.871", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/gpt-oss-120b/Server/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 75354.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5404 ROUGE2: 22.1352 ROUGEL: 28.9123 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/llama2-70b-99/Interactive/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 70954.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4435 ROUGE2: 22.0815 ROUGEL: 28.8533 TOKENS_PER_SAMPLE: 284.2", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/llama2-70b-99/Offline/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 100019.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5391 ROUGE2: 22.1346 ROUGEL: 28.91 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/llama2-70b-99/Server/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 95544.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5404 ROUGE2: 22.1352 ROUGEL: 28.9123 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/llama2-70b-99.9/Interactive/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 70954.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4435 ROUGE2: 22.0815 ROUGEL: 28.8533 TOKENS_PER_SAMPLE: 284.2", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/llama2-70b-99.9/Offline/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 100019.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0049", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)", "Platform": "HPE_ProLiant_XD685_MI355X_288GBx8", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5391 ROUGE2: 22.1346 ROUGEL: 28.91 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8/llama2-70b-99.9/Server/performance/run_1", "Software": "vLLM", "operating_system": "Ubuntu 22.04", "Notes": "256 CUs per GPU, gfx950 architecture", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_ProLiant_XD685_MI355X_288GBx8", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_ProLiant_XD685_MI355X_288GBx8", "ColorKey": "availableHPE", "Performance_Result": 95544.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0050", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380 Gen12 (5x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5883 ROUGE2: 15.86 ROUGEL: 24.3131 ROUGELSUM: 35.5397 GEN_LEN: 8181251", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 5, "Total Accelerators": 5, "Location": "./closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT", "ColorKey": "availableHPE", "Performance_Result": 11665.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0050", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant Compute DL380 Gen12 (5x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.587 ROUGE2: 15.8597 ROUGEL: 24.3106 ROUGELSUM: 35.5382 GEN_LEN: 8181395", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 5, "Total Accelerators": 5, "Location": "./closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_DC_TRT", "ColorKey": "availableHPE", "Performance_Result": 11352.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0051", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7589 ROUGE2: 22.3391 ROUGEL: 29.1509 TOKENS_PER_SAMPLE: 272.7", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 6862.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0051", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7582 ROUGE2: 22.3381 ROUGEL: 29.1502 TOKENS_PER_SAMPLE: 273.8", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 6662.33, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0051", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.5865 ROUGE2: 15.8596 ROUGEL: 24.3107 ROUGELSUM: 35.538 GEN_LEN: 8181239", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 10459.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0051", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5864 ROUGE2: 15.8586 ROUGEL: 24.3112 ROUGELSUM: 35.5374 GEN_LEN: 8181316", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 12433.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0051", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.5854 ROUGE2: 15.8591 ROUGEL: 24.3111 ROUGELSUM: 35.5366 GEN_LEN: 8181448", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 11898.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0051", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Offline", "Accuracy": "ROUGE1: 45.5367 ROUGE2: 23.3779 ROUGEL: 30.4572 TOKENS_PER_SAMPLE: 144.5 gsm8k_accuracy: 73.34 mbxp_accuracy: 60.06", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT/mixtral-8x7b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 8198.25, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0051", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Server", "Accuracy": "ROUGE1: 45.3405 ROUGE2: 23.2703 ROUGEL: 30.3729 TOKENS_PER_SAMPLE: 146.0 gsm8k_accuracy: 73.52 mbxp_accuracy: 59.78", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT/mixtral-8x7b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 7742.84, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8714 ROUGE2: 22.41 ROUGEL: 29.1982 TOKENS_PER_SAMPLE: 271.8", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 6901.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8708 ROUGE2: 22.4088 ROUGEL: 29.1982 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 6952.97, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8714 ROUGE2: 22.41 ROUGEL: 29.1982 TOKENS_PER_SAMPLE: 271.8", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 6893.36, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8727 ROUGE2: 22.4101 ROUGEL: 29.1991 TOKENS_PER_SAMPLE: 272.9", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 6953.48, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.7609 ROUGE2: 15.9725 ROUGEL: 24.5197 ROUGELSUM: 35.6577 GEN_LEN: 8148454", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 10589.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7643 ROUGE2: 15.9777 ROUGEL: 24.5156 ROUGELSUM: 35.6656 GEN_LEN: 8148206", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 12495.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7631 ROUGE2: 15.9835 ROUGEL: 24.5211 ROUGELSUM: 35.6664 GEN_LEN: 8148571", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 11899.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0052", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (2x NVIDIA RTX PRO 6000 Blackwell Server Edition-600W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.82254917822169", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 2, "Total Accelerators": 2, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 600W / 600W (full power)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx2_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 4712.66, "Performance_Units": "Samples/s" }, { "ID": "6.0-0053", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.5864 ROUGE2: 15.8598 ROUGEL: 24.3121 ROUGELSUM: 35.5377 GEN_LEN: 8181396", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 19991.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0053", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5869 ROUGE2: 15.8583 ROUGEL: 24.3109 ROUGELSUM: 35.5384 GEN_LEN: 8181409", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 23265.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0053", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.5867 ROUGE2: 15.8599 ROUGEL: 24.3123 ROUGELSUM: 35.5373 GEN_LEN: 8181397", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 22916.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0053", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Offline", "Accuracy": "ROUGE1: 45.378 ROUGE2: 23.1949 ROUGEL: 30.2877 TOKENS_PER_SAMPLE: 144.9 gsm8k_accuracy: 73.8 mbxp_accuracy: 60.12", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT/mixtral-8x7b/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 14828.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0053", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "UsedModel": "mixtral-8x7b", "Model": "mixtral-8x7b", "Scenario": "Server", "Accuracy": "ROUGE1: 45.3775 ROUGE2: 23.2622 ROUGEL: 30.3288 TOKENS_PER_SAMPLE: 145.9 gsm8k_accuracy: 73.52 mbxp_accuracy: 59.98", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT/mixtral-8x7b/Server/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_11_0_33_TRT", "ColorKey": "availableHPE", "Performance_Result": 13846.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.972", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 7297.05, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 84.122", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 6687.25, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8707 ROUGE2: 22.409 ROUGEL: 29.198 TOKENS_PER_SAMPLE: 271.9", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 13361.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8723 ROUGE2: 22.4108 ROUGEL: 29.1999 TOKENS_PER_SAMPLE: 272.8", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 12822.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8711 ROUGE2: 22.4097 ROUGEL: 29.1987 TOKENS_PER_SAMPLE: 271.9", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 13200.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8739 ROUGE2: 22.4139 ROUGEL: 29.2026 TOKENS_PER_SAMPLE: 272.8", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 12572.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.7669 ROUGE2: 16.003 ROUGEL: 24.5363 ROUGELSUM: 35.6914 GEN_LEN: 8180254", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 20118.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7715 ROUGE2: 16.0183 ROUGEL: 24.5357 ROUGELSUM: 35.696 GEN_LEN: 8178815", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 23556.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7528 ROUGE2: 16.008 ROUGEL: 24.5234 ROUGELSUM: 35.6804 GEN_LEN: 8178500", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 23227.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0054", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)", "Platform": "HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.81493571380987", "Nodes": 1, "Processor": "AMD EPYC 9475F", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 4, "Total Accelerators": 4, "Location": "./closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14.1.48, CUDA 13.0", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "GPU 450W / 600W (reduced)", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/HPE/HPE_Proliant_DL385_RTX_PRO_6000_PCIe_96GBx4_TRT_10_14_1_48_TRT", "ColorKey": "availableHPE", "Performance_Result": 9228.36, "Performance_Units": "Samples/s" }, { "ID": "6.0-0055", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant DL145 Gen11 (1x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "UsedModel": "resnet", "Model": "resnet", "Scenario": "MultiStream", "Accuracy": "acc: 76.138", "Nodes": 1, "Processor": "AMD EPYC 8434P", "host_processors_per_node": 1, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 1, "Total Accelerators": 1, "Location": "./closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT/resnet/MultiStream/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "ColorKey": "availableHPE", "Performance_Result": 10.570551, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0055", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant DL145 Gen11 (1x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "UsedModel": "resnet", "Model": "resnet", "Scenario": "Offline", "Accuracy": "acc: 76.138", "Nodes": 1, "Processor": "AMD EPYC 8434P", "host_processors_per_node": 1, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 1, "Total Accelerators": 1, "Location": "./closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT/resnet/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "ColorKey": "availableHPE", "Performance_Result": 19210.7, "Performance_Units": "Samples/s" }, { "ID": "6.0-0055", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant DL145 Gen11 (1x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "UsedModel": "resnet", "Model": "resnet", "Scenario": "SingleStream", "Accuracy": "acc: 76.138", "Nodes": 1, "Processor": "AMD EPYC 8434P", "host_processors_per_node": 1, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 1, "Total Accelerators": 1, "Location": "./closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT/resnet/SingleStream/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "ColorKey": "availableHPE", "Performance_Result": 10.561798, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0055", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant DL145 Gen11 (1x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.8425345223027", "Nodes": 1, "Processor": "AMD EPYC 8434P", "host_processors_per_node": 1, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 1, "Total Accelerators": 1, "Location": "./closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_PROLIANT_DL145_RTX_PRO_4500_PCIE_32GBx1_TRT_TRT", "ColorKey": "availableHPE", "Performance_Result": 916.743, "Performance_Units": "Samples/s" }, { "ID": "6.0-0056", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant Compute DL380 Gen12 (5x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "UsedModel": "resnet", "Model": "resnet", "Scenario": "MultiStream", "Accuracy": "acc: 76.138", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 5, "Total Accelerators": 5, "Location": "./closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT/resnet/MultiStream/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "ColorKey": "availableHPE", "Performance_Result": 11.077851, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0056", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant Compute DL380 Gen12 (5x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "UsedModel": "resnet", "Model": "resnet", "Scenario": "Offline", "Accuracy": "acc: 76.138", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 5, "Total Accelerators": 5, "Location": "./closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT/resnet/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "ColorKey": "availableHPE", "Performance_Result": 94650.5, "Performance_Units": "Samples/s" }, { "ID": "6.0-0056", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant Compute DL380 Gen12 (5x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "UsedModel": "resnet", "Model": "resnet", "Scenario": "SingleStream", "Accuracy": "acc: 76.138", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 5, "Total Accelerators": 5, "Location": "./closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT/resnet/SingleStream/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "ColorKey": "availableHPE", "Performance_Result": 10.991706, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0056", "Submitter": "HPE", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "HPE ProLiant Compute DL380 Gen12 (5x NVIDIA RTX PRO 4500 Blackwell, TensorRT)", "Platform": "HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.80922561550102", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA RTX PRO 4500 Blackwell", "a#": 5, "Total Accelerators": 5, "Location": "./closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.11.0.33, CUDA 13.0", "operating_system": "Ubuntu 24.04.3", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/results/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/HPE/src", "Unique ID (e.g. for Audit)": "edge/closed/HPE/HPE_Proliant_DL380_RTX_PRO_4500_PCIe_32GBx5_TRT", "ColorKey": "availableHPE", "Performance_Result": 4513.09, "Performance_Units": "Samples/s" }, { "ID": "6.0-0057", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.8062 ROUGE2: 16.0222 ROUGEL: 24.5645 ROUGELSUM: 35.8805 GEN_LEN: 8167163", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Intel/results/1-node-2S-GNR_128C/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "CentOS Stream 9", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-2S-GNR_128C", "ColorKey": "availableIntel", "Performance_Result": 1229.56, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0057", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.792 ROUGE2: 15.9934 ROUGEL: 24.5338 ROUGELSUM: 35.8672 GEN_LEN: 8163459", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Intel/results/1-node-2S-GNR_128C/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch", "operating_system": "CentOS Stream 9", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-2S-GNR_128C", "ColorKey": "availableIntel", "Performance_Result": 471.541, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0057", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "rgat", "Model": "rgat", "Scenario": "Offline", "Accuracy": "acc: 72.565", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Intel/results/1-node-2S-GNR_128C/rgat/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "CentOS Stream 9", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-2S-GNR_128C", "ColorKey": "availableIntel", "Performance_Result": 24149.8, "Performance_Units": "Samples/s" }, { "ID": "6.0-0057", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.89147168805961", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Intel/results/1-node-2S-GNR_128C/whisper/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "CentOS Stream 9", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-2S-GNR_128C", "ColorKey": "availableIntel", "Performance_Result": 2200.8, "Performance_Units": "Samples/s" }, { "ID": "6.0-0058", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B50", "Platform": "1-node-4x-BMG-B50", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7265 ROUGE2: 16.0947 ROUGEL: 24.6051 ROUGELSUM: 35.8654 GEN_LEN: 8149276", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8480C", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "Intel(R) Arc Pro(R) B50", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B50/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B50", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B50", "ColorKey": "availableIntel", "Performance_Result": 1721.84, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0058", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B50", "Platform": "1-node-4x-BMG-B50", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7283 ROUGE2: 16.0975 ROUGEL: 24.6103 ROUGELSUM: 35.8691 GEN_LEN: 8149111", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8480C", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "Intel(R) Arc Pro(R) B50", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B50/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B50", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B50", "ColorKey": "availableIntel", "Performance_Result": 1187.13, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0058", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B50", "Platform": "1-node-4x-BMG-B50", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.94194045614135", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8480C", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "Intel(R) Arc Pro(R) B50", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B50/whisper/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B50", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B50", "ColorKey": "availableIntel", "Performance_Result": 2112.35, "Performance_Units": "Samples/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.405", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/gpt-oss-120b/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "MXFP4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 841.038, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.582", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/gpt-oss-120b/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "MXFP4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 452.191, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8085 ROUGE2: 22.3222 ROUGEL: 29.0232 TOKENS_PER_SAMPLE: 284.5", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/llama2-70b-99/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 1697.66, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7707 ROUGE2: 22.2808 ROUGEL: 28.9988 TOKENS_PER_SAMPLE: 283.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/llama2-70b-99/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 1106.26, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8085 ROUGE2: 22.3222 ROUGEL: 29.0232 TOKENS_PER_SAMPLE: 284.5", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/llama2-70b-99.9/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 1697.66, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7707 ROUGE2: 22.2808 ROUGEL: 28.9988 TOKENS_PER_SAMPLE: 283.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/llama2-70b-99.9/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 1106.26, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7291 ROUGE2: 16.0967 ROUGEL: 24.6066 ROUGELSUM: 35.8688 GEN_LEN: 8149281", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 3347.43, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7301 ROUGE2: 16.0968 ROUGEL: 24.6083 ROUGELSUM: 35.8698 GEN_LEN: 8149242", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 3145.22, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0059", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B60", "Platform": "1-node-4x-BMG-B60", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.45197476712876", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B60", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B60/whisper/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B60", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B60", "ColorKey": "availableIntel", "Performance_Result": 3995.63, "Performance_Units": "Samples/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 82.528", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/gpt-oss-120b/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "MXFP4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 1601.91, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.413", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/gpt-oss-120b/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "MXFP4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 884.238, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7862 ROUGE2: 22.3133 ROUGEL: 29.0124 TOKENS_PER_SAMPLE: 284.7", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/llama2-70b-99/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 3270.66, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7636 ROUGE2: 22.2774 ROUGEL: 28.9958 TOKENS_PER_SAMPLE: 283.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/llama2-70b-99/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 2199.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7862 ROUGE2: 22.3133 ROUGEL: 29.0124 TOKENS_PER_SAMPLE: 284.7", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/llama2-70b-99.9/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 3270.66, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7636 ROUGE2: 22.2774 ROUGEL: 28.9958 TOKENS_PER_SAMPLE: 283.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/llama2-70b-99.9/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 2199.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7302 ROUGE2: 16.0981 ROUGEL: 24.6087 ROUGELSUM: 35.8702 GEN_LEN: 8149144", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 6762.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0060", "Submitter": "Intel", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-Pro-B60-Dual", "Platform": "1-node-4x-BMG-Pro-B60-Dual", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7315 ROUGE2: 16.0989 ROUGEL: 24.6089 ROUGELSUM: 35.8705 GEN_LEN: 8149219", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "MS-Intel Arc Pro B60 Dual 48G Turbo", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 25.04", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-Pro-B60-Dual", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-Pro-B60-Dual", "ColorKey": "availableIntel", "Performance_Result": 6293.89, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0061", "Submitter": "Krai", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE Cray XD670 (8x H200-SXM-141GB, KISS-V)", "Platform": "h200_n1-kissv", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 69.78", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "32", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Krai/results/h200_n1-kissv/wan-2.2-t2v-a14b/offline/performance/run_1", "Software": "KISS-V: Krai Inference Serving Solution for Video", "operating_system": "Red Hat Enterprise Linux release 9.4 (Plow) (Linux kernel: 5.14.0-427.13.1.el9_4.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Apr 10 10:29:16 EDT 2024 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "Powered by the Krai axs automation technology", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "bfloat16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Krai/results/h200_n1-kissv", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Krai/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Krai/h200_n1-kissv", "ColorKey": "availableKrai", "Performance_Result": 0.0181337, "Performance_Units": "Samples/s" }, { "ID": "6.0-0061", "Submitter": "Krai", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "HPE Cray XD670 (8x H200-SXM-141GB, KISS-V)", "Platform": "h200_n1-kissv", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 69.89", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "32", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Krai/results/h200_n1-kissv/wan-2.2-t2v-a14b/singlestream/performance/run_1", "Software": "KISS-V: Krai Inference Serving Solution for Video", "operating_system": "Red Hat Enterprise Linux release 9.4 (Plow) (Linux kernel: 5.14.0-427.13.1.el9_4.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Apr 10 10:29:16 EDT 2024 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "Powered by the Krai axs automation technology", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "bfloat16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Krai/results/h200_n1-kissv", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Krai/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Krai/h200_n1-kissv", "ColorKey": "availableKrai", "Performance_Result": 57.401991155, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0062", "Submitter": "Lambda_SIT", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Interactive", "Accuracy": "ROUGE1: 38.6433 ROUGE2: 15.8347 ROUGEL: 24.3517 ROUGELSUM: 35.6515 GEN_LEN: 8192621", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lambda_SIT/results/B200-SXM-180GBx8_TRT/llama3.1-8b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lambda_SIT/B200-SXM-180GBx8_TRT", "ColorKey": "availableLambda_SIT", "Performance_Result": 128750.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0062", "Submitter": "Lambda_SIT", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7787 ROUGE2: 16.1964 ROUGEL: 24.6085 ROUGELSUM: 35.8317 GEN_LEN: 8187884", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lambda_SIT/results/B200-SXM-180GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lambda_SIT/B200-SXM-180GBx8_TRT", "ColorKey": "availableLambda_SIT", "Performance_Result": 160403.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0062", "Submitter": "Lambda_SIT", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7757 ROUGE2: 16.1933 ROUGEL: 24.6056 ROUGELSUM: 35.8283 GEN_LEN: 8187968", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lambda_SIT/results/B200-SXM-180GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lambda_SIT/B200-SXM-180GBx8_TRT", "ColorKey": "availableLambda_SIT", "Performance_Result": 130008.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0063", "Submitter": "Lambda_SIT", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.601", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Lambda_SIT/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lambda_SIT/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableLambda_SIT", "Performance_Result": 60220.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0063", "Submitter": "Lambda_SIT", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.008", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Lambda_SIT/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lambda_SIT/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lambda_SIT/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableLambda_SIT", "Performance_Result": 53462.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0064", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7771 ROUGE2: 16.0352 ROUGEL: 24.5468 ROUGELSUM: 35.8518 GEN_LEN: 8146140", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Lenovo/results/1-node-2S-GNR_86C/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "ThinkSystem SR650 V4. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/1-node-2S-GNR_86C", "ColorKey": "availableLenovo", "Performance_Result": 775.982, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0064", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.851 ROUGE2: 16.086 ROUGEL: 24.6018 ROUGELSUM: 35.9286 GEN_LEN: 8145738", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Lenovo/results/1-node-2S-GNR_86C/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "ThinkSystem SR650 V4. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/1-node-2S-GNR_86C", "ColorKey": "availableLenovo", "Performance_Result": 281.779, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0064", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "rgat", "Model": "rgat", "Scenario": "Offline", "Accuracy": "acc: 72.574", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Lenovo/results/1-node-2S-GNR_86C/rgat/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "ThinkSystem SR650 V4. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/1-node-2S-GNR_86C", "ColorKey": "availableLenovo", "Performance_Result": 13570.3, "Performance_Units": "Samples/s" }, { "ID": "6.0-0064", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.88576267638467", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Lenovo/results/1-node-2S-GNR_86C/whisper/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "ThinkSystem SR650 V4. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/1-node-2S-GNR_86C", "ColorKey": "availableLenovo", "Performance_Result": 1411.42, "Performance_Units": "Samples/s" }, { "ID": "6.0-0065", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR675i V3 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "SR675iV3", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7302 ROUGE2: 22.3372 ROUGEL: 29.1145 TOKENS_PER_SAMPLE: 273.4", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR675iV3/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR675iV3", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR675iV3", "ColorKey": "availableLenovo", "Performance_Result": 26889.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0065", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR675i V3 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "SR675iV3", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.787 ROUGE2: 22.329 ROUGEL: 29.1181 TOKENS_PER_SAMPLE: 274.1", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR675iV3/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR675iV3", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR675iV3", "ColorKey": "availableLenovo", "Performance_Result": 27160.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0065", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR675i V3 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "SR675iV3", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7302 ROUGE2: 22.3372 ROUGEL: 29.1145 TOKENS_PER_SAMPLE: 273.4", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR675iV3/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR675iV3", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR675iV3", "ColorKey": "availableLenovo", "Performance_Result": 26889.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0065", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR675i V3 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "SR675iV3", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.787 ROUGE2: 22.329 ROUGEL: 29.1181 TOKENS_PER_SAMPLE: 274.1", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR675iV3/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR675iV3", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR675iV3", "ColorKey": "availableLenovo", "Performance_Result": 27160.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0065", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR675i V3 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)", "Platform": "SR675iV3", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.84158283925122", "Nodes": 1, "Processor": "AMD EPYC 9535", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR675iV3/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR675iV3", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR675iV3", "ColorKey": "availableLenovo", "Performance_Result": 18401.1, "Performance_Units": "Samples/s" }, { "ID": "6.0-0066", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR680a V4 (8x B300-SXM-270GB, TensorRT)", "Platform": "SR680aV4", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.731", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR680aV4/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR680aV4", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR680aV4", "ColorKey": "availableLenovo", "Performance_Result": 109141.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0066", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR680a V4 (8x B300-SXM-270GB, TensorRT)", "Platform": "SR680aV4", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.520", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR680aV4/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR680aV4", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR680aV4", "ColorKey": "availableLenovo", "Performance_Result": 100687.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0066", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "ThinkSystem SR680a V4 (8x B300-SXM-270GB, TensorRT)", "Platform": "SR680aV4", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.85109966976599", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Lenovo/results/SR680aV4/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SR680aV4", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Lenovo/SR680aV4", "ColorKey": "availableLenovo", "Performance_Result": 47268.7, "Performance_Units": "Samples/s" }, { "ID": "6.0-0067", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE100", "Platform": "SE100", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "MultiStream", "Accuracy": "acc: 76.218", "Nodes": 1, "Processor": "Intel(R) Core(TM) Ultra 7 255H [GPU]", "host_processors_per_node": 1, "host_processor_core_count": "16", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Lenovo/results/SE100/resnet50/multistream/performance/run_1", "Software": "OpenVINO", "operating_system": "Ubuntu 24.04.3 LTS (Noble Numbat)", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE100", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE100", "ColorKey": "availableLenovo", "Performance_Result": 9.918944, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0067", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE100", "Platform": "SE100", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "Offline", "Accuracy": "acc: 76.218", "Nodes": 1, "Processor": "Intel(R) Core(TM) Ultra 7 255H [GPU]", "host_processors_per_node": 1, "host_processor_core_count": "16", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Lenovo/results/SE100/resnet50/offline/performance/run_1", "Software": "OpenVINO", "operating_system": "Ubuntu 24.04.3 LTS (Noble Numbat)", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE100", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE100", "ColorKey": "availableLenovo", "Performance_Result": 834.804, "Performance_Units": "Samples/s" }, { "ID": "6.0-0067", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE100", "Platform": "SE100", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "SingleStream", "Accuracy": "acc: 76.218", "Nodes": 1, "Processor": "Intel(R) Core(TM) Ultra 7 255H [GPU]", "host_processors_per_node": 1, "host_processor_core_count": "16", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Lenovo/results/SE100/resnet50/SingleStream/performance/run_1", "Software": "OpenVINO", "operating_system": "Ubuntu 24.04.3 LTS (Noble Numbat)", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE100", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE100", "ColorKey": "availableLenovo", "Performance_Result": 1.203352, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "3d-unet-99", "Model": "3d-unet-99", "Scenario": "Offline", "Accuracy": "DICE: 0.90862", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/3d-unet-99/Offline/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 2.18794, "Performance_Units": "Samples/s" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "3d-unet-99", "Model": "3d-unet-99", "Scenario": "SingleStream", "Accuracy": "DICE: 0.90850", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/3d-unet-99/SingleStream/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 1880.729881, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "3d-unet-99.9", "Model": "3d-unet-99.9", "Scenario": "Offline", "Accuracy": "DICE: 0.90862", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/3d-unet-99.9/Offline/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 2.18794, "Performance_Units": "Samples/s" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "3d-unet-99.9", "Model": "3d-unet-99.9", "Scenario": "SingleStream", "Accuracy": "DICE: 0.90850", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/3d-unet-99.9/SingleStream/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 1880.729881, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "bert-99", "Model": "bert-99", "Scenario": "Offline", "Accuracy": "F1: 90.279", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/bert-99/Offline/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 1521.11, "Performance_Units": "Samples/s" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "bert-99", "Model": "bert-99", "Scenario": "SingleStream", "Accuracy": "F1: 90.267", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/bert-99/SingleStream/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 2.166524, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "MultiStream", "Accuracy": "acc: 76.078", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/resnet50/MultiStream/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 0.80682, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "Offline", "Accuracy": "acc: 76.078", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/resnet50/Offline/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 25200.5, "Performance_Units": "Samples/s" }, { "ID": "6.0-0068", "Submitter": "Lenovo", "Availability": "available", "Category": "closed", "Suite": "edge", "System": "ThinkEdge SE455i V3(2xL4, TensorRT", "Platform": "SE455I_V3_TRT", "UsedModel": "resnet50", "Model": "resnet", "Scenario": "SingleStream", "Accuracy": "acc: 76.064", "Nodes": 1, "Processor": "AMD EPYC 8534P", "host_processors_per_node": 1, "host_processor_core_count": "64", "Accelerator": "NVIDIA L4-PCIe-24GB", "a#": 2, "Total Accelerators": 2, "Location": "./closed/Lenovo/results/SE455I_V3_TRT/resnet50/SingleStream/performance/run_1", "Software": "TensorRT 10.8, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "int8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/results/SE455I_V3_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Lenovo/src", "Unique ID (e.g. for Audit)": "edge/closed/Lenovo/SE455I_V3_TRT", "ColorKey": "availableLenovo", "Performance_Result": 0.400845, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0069", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi300X (8x MI300X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4213 ROUGE2: 22.0278 ROUGEL: 28.6475 TOKENS_PER_SAMPLE: 296.1", "Nodes": 1, "Processor": "AMD EPYC 9534", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI300X_2xEPYC_9534/llama2-70b-99/Offline/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI300X_2xEPYC_9534", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI300X_2xEPYC_9534", "ColorKey": "availableMangoBoost", "Performance_Result": 27598.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0069", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi300X (8x MI300X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.4718 ROUGE2: 22.0358 ROUGEL: 28.668 TOKENS_PER_SAMPLE: 295.1", "Nodes": 1, "Processor": "AMD EPYC 9534", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI300X_2xEPYC_9534/llama2-70b-99/Server/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI300X_2xEPYC_9534", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI300X_2xEPYC_9534", "ColorKey": "availableMangoBoost", "Performance_Result": 25463.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0069", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi300X (8x MI300X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4213 ROUGE2: 22.0278 ROUGEL: 28.6475 TOKENS_PER_SAMPLE: 296.1", "Nodes": 1, "Processor": "AMD EPYC 9534", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI300X_2xEPYC_9534/llama2-70b-99.9/Offline/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI300X_2xEPYC_9534", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI300X_2xEPYC_9534", "ColorKey": "availableMangoBoost", "Performance_Result": 27598.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0069", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi300X (8x MI300X, LLMBoost)", "Platform": "8xMI300X_2xEPYC_9534", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.4718 ROUGE2: 22.0358 ROUGEL: 28.668 TOKENS_PER_SAMPLE: 295.1", "Nodes": 1, "Processor": "AMD EPYC 9534", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI300X 192GB HBM3", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI300X_2xEPYC_9534/llama2-70b-99.9/Server/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI300X_2xEPYC_9534", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI300X_2xEPYC_9534", "ColorKey": "availableMangoBoost", "Performance_Result": 25463.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0070", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi325X (8x MI325X, LLMBoost)", "Platform": "8xMI325X_2xEPYC_9655", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4375 ROUGE2: 22.0426 ROUGEL: 28.5893 TOKENS_PER_SAMPLE: 300.6", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI325X 256GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI325X_2xEPYC_9655/llama2-70b-99/Offline/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI325X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI325X_2xEPYC_9655", "ColorKey": "availableMangoBoost", "Performance_Result": 33880.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0070", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi325X (8x MI325X, LLMBoost)", "Platform": "8xMI325X_2xEPYC_9655", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.4754 ROUGE2: 22.0607 ROUGEL: 28.6289 TOKENS_PER_SAMPLE: 300.1", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI325X 256GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI325X_2xEPYC_9655/llama2-70b-99/Server/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI325X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI325X_2xEPYC_9655", "ColorKey": "availableMangoBoost", "Performance_Result": 31888.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0070", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi325X (8x MI325X, LLMBoost)", "Platform": "8xMI325X_2xEPYC_9655", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4375 ROUGE2: 22.0426 ROUGEL: 28.5893 TOKENS_PER_SAMPLE: 300.6", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI325X 256GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI325X_2xEPYC_9655/llama2-70b-99.9/Offline/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI325X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI325X_2xEPYC_9655", "ColorKey": "availableMangoBoost", "Performance_Result": 33880.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0070", "Submitter": "MangoBoost", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "MangoBoost Mi325X (8x MI325X, LLMBoost)", "Platform": "8xMI325X_2xEPYC_9655", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.4754 ROUGE2: 22.0607 ROUGEL: 28.6289 TOKENS_PER_SAMPLE: 300.1", "Nodes": 1, "Processor": "AMD EPYC 9655", "host_processors_per_node": 2, "host_processor_core_count": "96", "Accelerator": "AMD Instinct MI325X 256GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MangoBoost/results/8xMI325X_2xEPYC_9655/llama2-70b-99.9/Server/performance/run_1", "Software": "ROCm 6.10.5", "operating_system": "Ubuntu 22.04.5 LTS (jammy)", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/results/8xMI325X_2xEPYC_9655", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MangoBoost/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MangoBoost/8xMI325X_2xEPYC_9655", "ColorKey": "availableMangoBoost", "Performance_Result": 31888.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0071", "Submitter": "MiTAC", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G8825Z5", "Platform": "8xMI350X_2xEPYC-9755", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.6043 ROUGE2: 22.1796 ROUGEL: 28.9629 TOKENS_PER_SAMPLE: 281.8", "Nodes": 1, "Processor": "AMD EPYC 9755", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MiTAC/results/8xMI350X_2xEPYC-9755/llama2-70b-99/Interactive/performance/run_1", "Software": "Pytorch 2.9.1 , ROCm 7.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/results/8xMI350X_2xEPYC-9755", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MiTAC/8xMI350X_2xEPYC-9755", "ColorKey": "availableMiTAC", "Performance_Result": 54730.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0071", "Submitter": "MiTAC", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G8825Z5", "Platform": "8xMI350X_2xEPYC-9755", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4576 ROUGE2: 22.0944 ROUGEL: 28.8678 TOKENS_PER_SAMPLE: 284.0", "Nodes": 1, "Processor": "AMD EPYC 9755", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MiTAC/results/8xMI350X_2xEPYC-9755/llama2-70b-99/Offline/performance/run_1", "Software": "Pytorch 2.9.1 , ROCm 7.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/results/8xMI350X_2xEPYC-9755", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MiTAC/8xMI350X_2xEPYC-9755", "ColorKey": "availableMiTAC", "Performance_Result": 78988.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0071", "Submitter": "MiTAC", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G8825Z5", "Platform": "8xMI350X_2xEPYC-9755", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.607 ROUGE2: 22.1794 ROUGEL: 28.9613 TOKENS_PER_SAMPLE: 281.7", "Nodes": 1, "Processor": "AMD EPYC 9755", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MiTAC/results/8xMI350X_2xEPYC-9755/llama2-70b-99/Server/performance/run_1", "Software": "Pytorch 2.9.1 , ROCm 7.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/results/8xMI350X_2xEPYC-9755", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MiTAC/8xMI350X_2xEPYC-9755", "ColorKey": "availableMiTAC", "Performance_Result": 77377.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0071", "Submitter": "MiTAC", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G8825Z5", "Platform": "8xMI350X_2xEPYC-9755", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.6043 ROUGE2: 22.1796 ROUGEL: 28.9629 TOKENS_PER_SAMPLE: 281.8", "Nodes": 1, "Processor": "AMD EPYC 9755", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MiTAC/results/8xMI350X_2xEPYC-9755/llama2-70b-99.9/Interactive/performance/run_1", "Software": "Pytorch 2.9.1 , ROCm 7.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/results/8xMI350X_2xEPYC-9755", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MiTAC/8xMI350X_2xEPYC-9755", "ColorKey": "availableMiTAC", "Performance_Result": 54730.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0071", "Submitter": "MiTAC", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G8825Z5", "Platform": "8xMI350X_2xEPYC-9755", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4576 ROUGE2: 22.0944 ROUGEL: 28.8678 TOKENS_PER_SAMPLE: 284.0", "Nodes": 1, "Processor": "AMD EPYC 9755", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MiTAC/results/8xMI350X_2xEPYC-9755/llama2-70b-99.9/Offline/performance/run_1", "Software": "Pytorch 2.9.1 , ROCm 7.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/results/8xMI350X_2xEPYC-9755", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MiTAC/8xMI350X_2xEPYC-9755", "ColorKey": "availableMiTAC", "Performance_Result": 78988.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0071", "Submitter": "MiTAC", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "G8825Z5", "Platform": "8xMI350X_2xEPYC-9755", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.607 ROUGE2: 22.1794 ROUGEL: 28.9613 TOKENS_PER_SAMPLE: 281.7", "Nodes": 1, "Processor": "AMD EPYC 9755", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/MiTAC/results/8xMI350X_2xEPYC-9755/llama2-70b-99.9/Server/performance/run_1", "Software": "Pytorch 2.9.1 , ROCm 7.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/results/8xMI350X_2xEPYC-9755", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/MiTAC/src", "Unique ID (e.g. for Audit)": "datacenter/closed/MiTAC/8xMI350X_2xEPYC-9755", "ColorKey": "availableMiTAC", "Performance_Result": 77377.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0072", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Offline", "Accuracy": "DLRM_NE: 0.8675504587396435 DLRM_ACC: 0.6964261896038318 DLRM_AUC: 0.7862891896349031", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B200-SXM-180GBx8_TRT/dlrm-v3/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B200-SXM-180GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 10547.4, "Performance_Units": "Samples/s" }, { "ID": "6.0-0072", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Server", "Accuracy": "DLRM_NE: 0.8675504587396445 DLRM_ACC: 0.6964261896038318 DLRM_AUC: 0.7862891896349031", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B200-SXM-180GBx8_TRT/dlrm-v3/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B200-SXM-180GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 10007.0, "Performance_Units": "Queries/s" }, { "ID": "6.0-0072", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 69.8098", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B200-SXM-180GBx8_TRT/wan-2.2-t2v-a14b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B200-SXM-180GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 0.0464402, "Performance_Units": "Samples/s" }, { "ID": "6.0-0072", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 69.8853", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B200-SXM-180GBx8_TRT/wan-2.2-t2v-a14b/SingleStream/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B200-SXM-180GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 25.458383354, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0072", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.83967947314827", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B200-SXM-180GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B200-SXM-180GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 47433.9, "Performance_Units": "Samples/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.10756608933455 TOKENS_PER_SAMPLE: 3775.6280765724705", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 70325.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.19872379216044 TOKENS_PER_SAMPLE: 3741.4772105742936", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 42721.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.565", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 103961.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.582", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 100328.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7714 ROUGE2: 22.3739 ROUGEL: 29.1483 TOKENS_PER_SAMPLE: 273.1", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 112954.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.856 ROUGE2: 22.4112 ROUGEL: 29.1899 TOKENS_PER_SAMPLE: 271.7", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 107318.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7714 ROUGE2: 22.3739 ROUGEL: 29.1483 TOKENS_PER_SAMPLE: 273.1", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 112954.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.856 ROUGE2: 22.4112 ROUGEL: 29.1899 TOKENS_PER_SAMPLE: 271.7", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 107318.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.84372254775757 exact_match: 90.04178689609532 TOKENS_PER_SAMPLE: 636.0", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 1951.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Server", "Accuracy": "ROUGEL: 21.822118834475546 exact_match: 90.01338186631372 TOKENS_PER_SAMPLE: 638.6", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama3.1-405b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 1460.25, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5159 ROUGE2: 15.9339 ROUGEL: 24.3491 ROUGELSUM: 35.6714 GEN_LEN: 8185045", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 165432.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.5059 ROUGE2: 15.9356 ROUGEL: 24.3441 ROUGELSUM: 35.6683 GEN_LEN: 8185542", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/llama3.1-8b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 141565.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 70.3453", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/wan-2.2-t2v-a14b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 0.0585234, "Performance_Units": "Samples/s" }, { "ID": "6.0-0073", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "B300-SXM-270GBx8_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 70.7974", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/NVIDIA/results/B300-SXM-270GBx8_TRT/wan-2.2-t2v-a14b/SingleStream/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/B300-SXM-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/B300-SXM-270GBx8_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 23.778912625, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0074", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (4x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7884930210985771", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 43.7662, "Performance_Units": "Samples/s" }, { "ID": "6.0-0074", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (4x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.7883566298697722", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 38.0313, "Performance_Units": "Queries/s" }, { "ID": "6.0-0074", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (4x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x4_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 69.8883", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT/wan-2.2-t2v-a14b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 0.0269817, "Performance_Units": "Samples/s" }, { "ID": "6.0-0074", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (4x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x4_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 70.6549", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT/wan-2.2-t2v-a14b/SingleStream/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 40.243719116, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Interactive", "Accuracy": "exact_match: 81.51777575205105 TOKENS_PER_SAMPLE: 3778.798541476755", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/deepseek-r1/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 240318.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.51777575205105 TOKENS_PER_SAMPLE: 3761.302415679125", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 486141.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.74567000911577 TOKENS_PER_SAMPLE: 3737.3990428441202", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 336106.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Offline", "Accuracy": "DLRM_NE: 0.8675504587396562 DLRM_ACC: 0.6964261896038318 DLRM_AUC: 0.7862891896349031", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/dlrm-v3/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 104637.0, "Performance_Units": "Samples/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Server", "Accuracy": "DLRM_NE: 0.8675504587396344 DLRM_ACC: 0.6964261896038318 DLRM_AUC: 0.7862891896349031", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/dlrm-v3/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 99996.8, "Performance_Units": "Queries/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Interactive", "Accuracy": "exact_match: 83.609", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/gpt-oss-120b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 624929.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.630", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 879542.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.181", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 899218.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.8415 ROUGE2: 22.4159 ROUGEL: 29.2129 TOKENS_PER_SAMPLE: 271.5", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 754855.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8311 ROUGE2: 22.4118 ROUGEL: 29.2118 TOKENS_PER_SAMPLE: 271.5", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 888054.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8281 ROUGE2: 22.407 ROUGEL: 29.2059 TOKENS_PER_SAMPLE: 271.6", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 810104.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.8415 ROUGE2: 22.4159 ROUGEL: 29.2129 TOKENS_PER_SAMPLE: 271.5", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama2-70b-99.9/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 754855.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8311 ROUGE2: 22.4118 ROUGEL: 29.2118 TOKENS_PER_SAMPLE: 271.5", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 888054.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8281 ROUGE2: 22.407 ROUGEL: 29.2059 TOKENS_PER_SAMPLE: 271.6", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 810104.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Interactive", "Accuracy": "ROUGEL: 21.784306887727258 exact_match: 89.97819986763734 TOKENS_PER_SAMPLE: 638.7", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama3.1-405b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 14010.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.791103518649138 exact_match: 89.965337524818 TOKENS_PER_SAMPLE: 636.9", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 15461.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0075", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)", "Platform": "GB200-NVL72_GB200-186GB_aarch64x72_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Server", "Accuracy": "ROUGEL: 21.78378090513782 exact_match: 90.00418266048975 TOKENS_PER_SAMPLE: 640.0", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB200", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT/llama3.1-405b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB200 NVL72, 1200W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB200-NVL72_GB200-186GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 14133.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0076", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x288_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.60893345487693 TOKENS_PER_SAMPLE: 3773.291704649043", "Nodes": 72, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 288, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x288_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x288_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x288_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 2494310.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0076", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x288_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.56335460346399 TOKENS_PER_SAMPLE: 3753.950774840474", "Nodes": 72, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 288, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x288_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x288_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x288_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 1555110.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0077", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7882200623476316", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 61.2443, "Performance_Units": "Samples/s" }, { "ID": "6.0-0077", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.7883219387216303", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 41.6351, "Performance_Units": "Queries/s" }, { "ID": "6.0-0077", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 70.0891", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/wan-2.2-t2v-a14b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 0.036622, "Performance_Units": "Samples/s" }, { "ID": "6.0-0077", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 70.2488", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/wan-2.2-t2v-a14b/SingleStream/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 30.910508163, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Interactive", "Accuracy": "exact_match: 81.54056517775751 TOKENS_PER_SAMPLE: 3744.0936645396537", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/deepseek-r1/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 250634.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.10756608933455 TOKENS_PER_SAMPLE: 3775.2474931631723", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 647076.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.70009115770283 TOKENS_PER_SAMPLE: 3771.458979033728", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 400437.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Interactive", "Accuracy": "exact_match: 83.361", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/gpt-oss-120b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 677199.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.503", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 1042980.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 82.918", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 1072250.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.8316 ROUGE2: 22.3762 ROUGEL: 29.1984 TOKENS_PER_SAMPLE: 272.2", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama2-70b-99/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 814128.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8406 ROUGE2: 22.4168 ROUGEL: 29.2108 TOKENS_PER_SAMPLE: 272.0", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 1126850.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8072 ROUGE2: 22.3641 ROUGEL: 29.1722 TOKENS_PER_SAMPLE: 272.7", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 868278.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.8316 ROUGE2: 22.3762 ROUGEL: 29.1984 TOKENS_PER_SAMPLE: 272.2", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama2-70b-99.9/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 814128.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.8406 ROUGE2: 22.4168 ROUGEL: 29.2108 TOKENS_PER_SAMPLE: 272.0", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 1126850.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.8072 ROUGE2: 22.3641 ROUGEL: 29.1722 TOKENS_PER_SAMPLE: 272.7", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 868278.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Interactive", "Accuracy": "ROUGEL: 21.756418050547687 exact_match: 90.05151555261416 TOKENS_PER_SAMPLE: 641.5", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama3.1-405b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 18365.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.728067830579086 exact_match: 89.94727332892126 TOKENS_PER_SAMPLE: 638.8", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 19511.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0078", "Submitter": "NVIDIA", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Server", "Accuracy": "ROUGEL: 21.801876398806673 exact_match: 90.001455989411 TOKENS_PER_SAMPLE: 642.7", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/llama3.1-405b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72, 1400W TGP Per GPU", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/NVIDIA/src", "Unique ID (e.g. for Audit)": "datacenter/closed/NVIDIA/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNVIDIA", "Performance_Result": 18627.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0079", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x1_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.529", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x1_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x1_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x1_TRT", "ColorKey": "availableNebius", "Performance_Result": 15371.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0079", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x1_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.915", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 1, "Total Accelerators": 1, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x1_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x1_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x1_TRT", "ColorKey": "availableNebius", "Performance_Result": 14973.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0080", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7876592354860111", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableNebius", "Performance_Result": 60.6154, "Performance_Units": "Samples/s" }, { "ID": "6.0-0080", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x4_TRT", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.787752047599809", "Nodes": 1, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT/qwen3-vl-235b-a22b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp32", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x4_TRT", "ColorKey": "availableNebius", "Performance_Result": 43.1538, "Performance_Units": "Queries/s" }, { "ID": "6.0-0081", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.44940747493163 TOKENS_PER_SAMPLE: 3806.084092980857", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNebius", "Performance_Result": 673936.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0081", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 80.78851412944394 TOKENS_PER_SAMPLE: 3751.325205104831", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNebius", "Performance_Result": 575580.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0081", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.036", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNebius", "Performance_Result": 1046150.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0081", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x72_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.217", "Nodes": 18, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 72, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x72_TRT", "ColorKey": "availableNebius", "Performance_Result": 1096770.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0082", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.76845943482225 TOKENS_PER_SAMPLE: 3735.315861440292", "Nodes": 2, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 8, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "ColorKey": "availableNebius", "Performance_Result": 76346.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0082", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.40382862351869 TOKENS_PER_SAMPLE: 3766.282133090246", "Nodes": 2, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 8, "Location": "./closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, Driver 580.105.08, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "NVIDIA GB300 NVL72", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/GB300-NVL72_GB300-288GB_aarch64x8_TRT", "ColorKey": "availableNebius", "Performance_Result": 64510.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0083", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B200 n1 (8x B200-SXM-180GB, TensorRT)", "Platform": "nebius_b200_n1", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.79124886052871 TOKENS_PER_SAMPLE: 3765.5856882406565", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8580", "host_processors_per_node": 2, "host_processor_core_count": "60", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b200_n1/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b200_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b200_n1", "ColorKey": "availableNebius", "Performance_Result": 58581.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0083", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B200 n1 (8x B200-SXM-180GB, TensorRT)", "Platform": "nebius_b200_n1", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.42661804922516 TOKENS_PER_SAMPLE: 3735.2408842297173", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8580", "host_processors_per_node": 2, "host_processor_core_count": "60", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b200_n1/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b200_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b200_n1", "ColorKey": "availableNebius", "Performance_Result": 51692.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0083", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B200 n1 (8x B200-SXM-180GB, TensorRT)", "Platform": "nebius_b200_n1", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Interactive", "Accuracy": "exact_match: 83.295", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8580", "host_processors_per_node": 2, "host_processor_core_count": "60", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b200_n1/gpt-oss-120b/Interactive/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b200_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b200_n1", "ColorKey": "availableNebius", "Performance_Result": 13155.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0083", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B200 n1 (8x B200-SXM-180GB, TensorRT)", "Platform": "nebius_b200_n1", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.750", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8580", "host_processors_per_node": 2, "host_processor_core_count": "60", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b200_n1/gpt-oss-120b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b200_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b200_n1", "ColorKey": "availableNebius", "Performance_Result": 85921.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0083", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B200 n1 (8x B200-SXM-180GB, TensorRT)", "Platform": "nebius_b200_n1", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.833", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8580", "host_processors_per_node": 2, "host_processor_core_count": "60", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b200_n1/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b200_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b200_n1", "ColorKey": "availableNebius", "Performance_Result": 87444.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0084", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B300 n1 (8x B300-SXM-270GB, TensorRT)", "Platform": "nebius_b300_n1", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.58614402917047 TOKENS_PER_SAMPLE: 3722.1116681859617", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6776P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b300_n1/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b300_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b300_n1", "ColorKey": "availableNebius", "Performance_Result": 69318.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0084", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B300 n1 (8x B300-SXM-270GB, TensorRT)", "Platform": "nebius_b300_n1", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.58614402917047 TOKENS_PER_SAMPLE: 3721.7894257064722", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6776P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b300_n1/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b300_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b300_n1", "ColorKey": "availableNebius", "Performance_Result": 60413.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0084", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B300 n1 (8x B300-SXM-270GB, TensorRT)", "Platform": "nebius_b300_n1", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 82.959", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6776P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b300_n1/gpt-oss-120b/offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b300_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b300_n1", "ColorKey": "availableNebius", "Performance_Result": 106885.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0084", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B300 n1 (8x B300-SXM-270GB, TensorRT)", "Platform": "nebius_b300_n1", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.337", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6776P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b300_n1/gpt-oss-120b/server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b300_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b300_n1", "ColorKey": "availableNebius", "Performance_Result": 100437.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0084", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B300 n1 (8x B300-SXM-270GB, TensorRT)", "Platform": "nebius_b300_n1", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7880721729948121", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6776P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b300_n1/qwen3-vl-235b-a22b/offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "BF16/F32/F8_E4M3/U8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b300_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b300_n1", "ColorKey": "availableNebius", "Performance_Result": 78.2775, "Performance_Units": "Samples/s" }, { "ID": "6.0-0084", "Submitter": "Nebius", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Nebius B300 n1 (8x B300-SXM-270GB, TensorRT)", "Platform": "nebius_b300_n1", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.7865237675409595", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6776P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Nebius/results/nebius_b300_n1/qwen3-vl-235b-a22b/server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04.3 LTS (Linux kernel: 6.11.0-1016-nvidia #16-Ubuntu SMP PREEMPT_DYNAMIC Sun Sep 21 17:06:33 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "BF16/F32/F8_E4M3/U8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/results/nebius_b300_n1", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Nebius/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Nebius/nebius_b300_n1", "ColorKey": "availableNebius", "Performance_Result": 45.1513, "Performance_Units": "Queries/s" }, { "ID": "6.0-0085", "Submitter": "Netweb_technologies", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Camarero PDI200A2HG-810 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Offline", "Accuracy": "DLRM_NE: 0.8675504587396444 DLRM_ACC: 0.6964261896038318 DLRM_AUC: 0.7862891896349031", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Netweb_technologies/results/B200-SXM-180GBx8_TRT/dlrm-v3/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Netweb_technologies/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Netweb_technologies/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Netweb_technologies/B200-SXM-180GBx8_TRT", "ColorKey": "availableNetweb_technologies", "Performance_Result": 10737.2, "Performance_Units": "Samples/s" }, { "ID": "6.0-0085", "Submitter": "Netweb_technologies", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Camarero PDI200A2HG-810 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "dlrm-v3", "Model": "dlrm-v3", "Scenario": "Server", "Accuracy": "DLRM_NE: 0.8675504587396445 DLRM_ACC: 0.6964261896038318 DLRM_AUC: 0.7862891896349031", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Netweb_technologies/results/B200-SXM-180GBx8_TRT/dlrm-v3/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Netweb_technologies/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Netweb_technologies/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Netweb_technologies/B200-SXM-180GBx8_TRT", "ColorKey": "availableNetweb_technologies", "Performance_Result": 10007.0, "Performance_Units": "Queries/s" }, { "ID": "6.0-0085", "Submitter": "Netweb_technologies", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Camarero PDI200A2HG-810 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.86632659858961", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Netweb_technologies/results/B200-SXM-180GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1, cuDNN 9.17, TensorRT-LLM feat/1.2-mlpinf, NVIDIA Dynamo mlperf-v6.0-dynamo-v0.8.0, vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Netweb_technologies/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Netweb_technologies/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Netweb_technologies/B200-SXM-180GBx8_TRT", "ColorKey": "availableNetweb_technologies", "Performance_Result": 45083.1, "Performance_Units": "Samples/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.801", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/gpt-oss-120b/Offline/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 91706.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 82.597", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/gpt-oss-120b/Server/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 78186.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5404 ROUGE2: 22.1362 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/llama2-70b-99/Interactive/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 72116.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4665 ROUGE2: 22.092 ROUGEL: 28.8666 TOKENS_PER_SAMPLE: 283.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/llama2-70b-99/Offline/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 99874.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5401 ROUGE2: 22.1345 ROUGEL: 28.9098 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/llama2-70b-99/Server/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 98239.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5404 ROUGE2: 22.1362 ROUGEL: 28.9134 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/llama2-70b-99.9/Interactive/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 72116.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4665 ROUGE2: 22.092 ROUGEL: 28.8666 TOKENS_PER_SAMPLE: 283.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/llama2-70b-99.9/Offline/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 99874.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0086", "Submitter": "ORACLE", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "BM.GPU.MI355X.8", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5401 ROUGE2: 22.1345 ROUGEL: 28.9098 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/ORACLE/results/8xMI355X_2xEPYC_9575F/llama2-70b-99.9/Server/performance/run_1", "Software": "", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. System typically deployed with ROCm stack and standard OCI GPU images for AI/HPC workloads.", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/ORACLE/src", "Unique ID (e.g. for Audit)": "datacenter/closed/ORACLE/8xMI355X_2xEPYC_9575F", "ColorKey": "availableORACLE", "Performance_Result": 98239.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0087", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7399 ROUGE2: 15.8699 ROUGEL: 24.3997 ROUGELSUM: 35.7855 GEN_LEN: 8144909", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Rocky Linux 9.i5", "Notes": "QuantaGrid D54Q-2U. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/1-node-2S-GNR_86C", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 777.671, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0087", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7265 ROUGE2: 15.8924 ROUGEL: 24.395 ROUGELSUM: 35.7669 GEN_LEN: 8146962", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch", "operating_system": "Rocky Linux 9.i5", "Notes": "QuantaGrid D54Q-2U. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/1-node-2S-GNR_86C", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 284.359, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0087", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "rgat", "Model": "rgat", "Scenario": "Offline", "Accuracy": "acc: 72.567", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C/rgat/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Rocky Linux 9.i5", "Notes": "QuantaGrid D54Q-2U. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/1-node-2S-GNR_86C", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 13151.7, "Performance_Units": "Samples/s" }, { "ID": "6.0-0087", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_86C", "Platform": "1-node-2S-GNR_86C", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.88576267638467", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6787P", "host_processors_per_node": 2, "host_processor_core_count": "86", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C/whisper/Offline/performance/run_1", "Software": "PyTorch", "operating_system": "Rocky Linux 9.i5", "Notes": "QuantaGrid D54Q-2U. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/1-node-2S-GNR_86C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/1-node-2S-GNR_86C", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 1407.49, "Performance_Units": "Samples/s" }, { "ID": "6.0-0088", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D75E-4U (8x RTX PRO6000 96GB, TensorRT)", "Platform": "D75E-4U_RTX_PRO_6000_PCIE_96GBx8_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.8387277900968", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA RTX PRO 6000 Blackwell Server Edition", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Quanta_Cloud_Technology/results/D75E-4U_RTX_PRO_6000_PCIE_96GBx8_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Rocky Linux 9.5", "Notes": "QuantaGrid D75E-4U", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/D75E-4U_RTX_PRO_6000_PCIE_96GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/D75E-4U_RTX_PRO_6000_PCIE_96GBx8_TRT", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 17838.3, "Performance_Units": "Samples/s" }, { "ID": "6.0-0089", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D75H-10U (8x B300-SXM-270GB, TensorRT)", "Platform": "D75H-10U_B300-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 80.92525068368278 TOKENS_PER_SAMPLE: 3795.107793983592", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Rocky Linux 9.5", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/D75H-10U_B300-270GBx8_TRT", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 69263.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0089", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D75H-10U (8x B300-SXM-270GB, TensorRT)", "Platform": "D75H-10U_B300-270GBx8_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Server", "Accuracy": "exact_match: 81.38103919781221 TOKENS_PER_SAMPLE: 3759.358477666363", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT/deepseek-r1/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Rocky Linux 9.5", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/D75H-10U_B300-270GBx8_TRT", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 35435.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0089", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D75H-10U (8x B300-SXM-270GB, TensorRT)", "Platform": "D75H-10U_B300-270GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.9243802693401 exact_match: 90.1098841826605 TOKENS_PER_SAMPLE: 648.6", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Rocky Linux 9.5", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/D75H-10U_B300-270GBx8_TRT", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 1953.77, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0089", "Submitter": "Quanta_Cloud_Technology", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D75H-10U (8x B300-SXM-270GB, TensorRT)", "Platform": "D75H-10U_B300-270GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Server", "Accuracy": "ROUGEL: 21.924447125381427 exact_match: 90.12480476505624 TOKENS_PER_SAMPLE: 647.6", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6747P", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT/llama3.1-405b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Rocky Linux 9.5", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/results/D75H-10U_B300-270GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Quanta_Cloud_Technology/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Quanta_Cloud_Technology/D75H-10U_B300-270GBx8_TRT", "ColorKey": "availableQuanta_Cloud_Technology", "Performance_Result": 1484.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0090", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "IBM Cloud gx3-48x240x2l40s instance,2xL40S-PCIE-48GB,vLLM 0.15.1", "Platform": "2xL40S_RHEL9", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 98.1099574597676", "Nodes": 1, "Processor": "Intel Xeon Processor (SapphireRapids)", "host_processors_per_node": 2, "host_processor_core_count": "48", "Accelerator": "NVIDIA L40S", "a#": 2, "Total Accelerators": 2, "Location": "./closed/RedHat/results/2xL40S_RHEL9/whisper/Offline/performance/run_1", "Software": "vLLM 0.15.1, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "IBM Cloud gx3-48x240x2l40s instance with 2xL40S-PCIE-48G. RHEL 9.6, vllm 0.15.1", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "bfloat16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/2xL40S_RHEL9", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/2xL40S_RHEL9", "ColorKey": "availableRedHat", "Performance_Result": 3646.91, "Performance_Units": "Samples/s" }, { "ID": "6.0-0091", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "LLM-D v0.5.0,Openshift 4.20.12,NVIDIA 8xB200-SXM-180GB", "Platform": "8xB200-LLM-D-Openshift", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 84.028", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "112", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xB200-LLM-D-Openshift/gpt-oss-120b/Offline/performance/run_1", "Software": "LLM-D v0.5.0 ,vLLM 0.14.1, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "NVIDIA 8xB200-SXM-180G. LLM-D v0.5.0,Openshift 4.20.12,RHEL 9.6,vllm 0.14.1", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "mxfp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xB200-LLM-D-Openshift", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xB200-LLM-D-Openshift", "ColorKey": "availableRedHat", "Performance_Result": 93070.7, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0091", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "LLM-D v0.5.0,Openshift 4.20.12,NVIDIA 8xB200-SXM-180GB", "Platform": "8xB200-LLM-D-Openshift", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.724", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "112", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xB200-LLM-D-Openshift/gpt-oss-120b/Server/performance/run_1", "Software": "LLM-D v0.5.0 ,vLLM 0.14.1, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "NVIDIA 8xB200-SXM-180G. LLM-D v0.5.0,Openshift 4.20.12,RHEL 9.6,vllm 0.14.1", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "mxfp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xB200-LLM-D-Openshift", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xB200-LLM-D-Openshift", "ColorKey": "availableRedHat", "Performance_Result": 71588.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0092", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "LLM-D v0.5.0,OpenShift 4.21,IBM cloud gx3d-160x1792x8h200,NVIDIA 8xH200-SXM-141GB", "Platform": "8xH200-LLM-D-Openshift", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 84.206", "Nodes": 1, "Processor": "Intel Xeon Processor (SapphireRapids)", "host_processors_per_node": 2, "host_processor_core_count": "80", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xH200-LLM-D-Openshift/gpt-oss-120b/Offline/performance/run_1", "Software": "LLM-D v0.5.0 ,OpenShift 4.21, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "IBM Cloud gx3d-160x1792x8h200 instance with 8xH200-SXM-141G. LLM-D 0.5.0, RHEL 9.6, vllm 0.14.1", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "mxfp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xH200-LLM-D-Openshift", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xH200-LLM-D-Openshift", "ColorKey": "availableRedHat", "Performance_Result": 28680.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0092", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "LLM-D v0.5.0,OpenShift 4.21,IBM cloud gx3d-160x1792x8h200,NVIDIA 8xH200-SXM-141GB", "Platform": "8xH200-LLM-D-Openshift", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.680", "Nodes": 1, "Processor": "Intel Xeon Processor (SapphireRapids)", "host_processors_per_node": 2, "host_processor_core_count": "80", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xH200-LLM-D-Openshift/gpt-oss-120b/Server/performance/run_1", "Software": "LLM-D v0.5.0 ,OpenShift 4.21, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "IBM Cloud gx3d-160x1792x8h200 instance with 8xH200-SXM-141G. LLM-D 0.5.0, RHEL 9.6, vllm 0.14.1", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "mxfp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xH200-LLM-D-Openshift", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xH200-LLM-D-Openshift", "ColorKey": "availableRedHat", "Performance_Result": 24103.2, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0093", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "IBM Cloud qx3d-160x1792x8h200,NVIDIA 8xH200-SXM-141GB,vLLM 0.15.1", "Platform": "8xH200_whisper", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 98.11090914281908", "Nodes": 1, "Processor": "Intel Xeon Processor (Sapphire Rapids)", "host_processors_per_node": 2, "host_processor_core_count": "80", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xH200_whisper/whisper/Offline/performance/run_1", "Software": "vLLM 0.15.1, PyTorch 2.9.1+cu128, CUDA 12.9.86", "operating_system": "Red Hat Enterprise Linux 9.6 (Plow), Kernel 5.14.0-570.79.1.el9_6.x86_64", "Notes": "IBM cloud H200 system with 8x NVIDIA H200 GPUs (141GB HBM3e each), 2-socket Intel Sapphire Rapids CPUs (80 cores total), 1.7TB DDR5 ECC memory, and 56TB NVMe storage. Using vLLM 0.15.1 with Flash-Attention, torch.compile enabled, 8 parallel workers (1 per GPU), GPU memory utilization 95%, max batch tokens 32000, max sequences 256", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "bfloat16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xH200_whisper", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xH200_whisper", "ColorKey": "availableRedHat", "Performance_Result": 36395.7, "Performance_Units": "Samples/s" }, { "ID": "6.0-0094", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Dell B200,8xB200-SXM-180GB,RHEL 10.1,vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "Platform": "8xb200_NV_vllm_qwen", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7890600786539296", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "112", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xb200_NV_vllm_qwen/qwen3-vl-235b-a22b/offline/performance/run_1", "Software": "vLLM + NVIDIA-dynamo + MLPerf Inference Qwen3-VL harness, RHEL 10.1", "operating_system": "Red Hat Enterprise Linux 10.1 (Coughlan)", "Notes": "1 node with 8x NVIDIA B200; all-to-all NVLink (NV18); 2-socket Intel Xeon Platinum 8570; NUMA=2; ~2.0 TiB RAM; 4x 7.68TB Dell NVMe ISE PS1010 RI U.2 + Dell BOSS-N1 OS disk. RHEL 10.1, vLLM via Dynamo backend; async scheduling enabled; SJF enabled with --scheduling-policy=sjf; CUDA Graph capture enabled via compilation config", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xb200_NV_vllm_qwen", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xb200_NV_vllm_qwen", "ColorKey": "availableRedHat", "Performance_Result": 79.0373, "Performance_Units": "Samples/s" }, { "ID": "6.0-0094", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "Dell B200,8xB200-SXM-180GB,RHEL 10.1,vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "Platform": "8xb200_NV_vllm_qwen", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.7887584531969497", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "112", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xb200_NV_vllm_qwen/qwen3-vl-235b-a22b/server/performance/run_1", "Software": "vLLM + NVIDIA-dynamo + MLPerf Inference Qwen3-VL harness, RHEL 10.1", "operating_system": "Red Hat Enterprise Linux 10.1 (Coughlan)", "Notes": "1 node with 8x NVIDIA B200; all-to-all NVLink (NV18); 2-socket Intel Xeon Platinum 8570; NUMA=2; ~2.0 TiB RAM; 4x 7.68TB Dell NVMe ISE PS1010 RI U.2 + Dell BOSS-N1 OS disk. RHEL 10.1, vLLM via Dynamo backend; async scheduling enabled; SJF enabled with --scheduling-policy=sjf; CUDA Graph capture enabled via compilation config", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xb200_NV_vllm_qwen", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xb200_NV_vllm_qwen", "ColorKey": "availableRedHat", "Performance_Result": 67.8642, "Performance_Units": "Queries/s" }, { "ID": "6.0-0095", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "IBM Cloud gx3d-160x1792x8h200,8xH200-SXM-141GB,RHEL 9.6,vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "Platform": "8xh200_NV_vllm_qwen", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Offline", "Accuracy": "F1_HIERARCHICAL: 0.7900114757860913", "Nodes": 1, "Processor": "Intel Xeon Processor (SapphireRapids)", "host_processors_per_node": 2, "host_processor_core_count": "80", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xh200_NV_vllm_qwen/qwen3-vl-235b-a22b/offline/performance/run_1", "Software": "vLLM 0.10.0, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "IBM Cloud instance gx3d-160x1792x8h200 8xH200-SXM-141GB. RHEL 9.6, vllm 6.0rc2+cu130", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xh200_NV_vllm_qwen", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xh200_NV_vllm_qwen", "ColorKey": "availableRedHat", "Performance_Result": 18.218, "Performance_Units": "Samples/s" }, { "ID": "6.0-0095", "Submitter": "RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "IBM Cloud gx3d-160x1792x8h200,8xH200-SXM-141GB,RHEL 9.6,vLLM CentML:mlperf-inf-mm-q3vl-v6.0", "Platform": "8xh200_NV_vllm_qwen", "UsedModel": "qwen3-vl-235b-a22b", "Model": "qwen3-vl-235b-a22b", "Scenario": "Server", "Accuracy": "F1_HIERARCHICAL: 0.7898129513572557", "Nodes": 1, "Processor": "Intel Xeon Processor (SapphireRapids)", "host_processors_per_node": 2, "host_processor_core_count": "80", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/RedHat/results/8xh200_NV_vllm_qwen/qwen3-vl-235b-a22b/server/performance/run_1", "Software": "vLLM 0.10.0, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "IBM Cloud instance gx3d-160x1792x8h200 8xH200-SXM-141GB. RHEL 9.6, vllm 6.0rc2+cu130", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/results/8xh200_NV_vllm_qwen", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/RedHat/8xh200_NV_vllm_qwen", "ColorKey": "availableRedHat", "Performance_Result": 11.0498, "Performance_Units": "Queries/s" }, { "ID": "6.0-0096", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7203 ROUGE2: 15.8576 ROUGEL: 24.3722 ROUGELSUM: 35.7664 GEN_LEN: 8142319", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Supermicro/results/1-node-2S-GNR_128C/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch: rgat: 2.7.0+cpu, whipser: 2.7.0+cpu, llama3.1-8b: 2.8.0.dev20250627+cpu", "operating_system": "Ubuntu 22.04.5", "Notes": "SYS-822GA-NGR3. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/1-node-2S-GNR_128C", "ColorKey": "availableSupermicro", "Performance_Result": 1185.98, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0096", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.7405 ROUGE2: 15.8949 ROUGEL: 24.4131 ROUGELSUM: 35.7831 GEN_LEN: 8147194", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Supermicro/results/1-node-2S-GNR_128C/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch: rgat: 2.7.0+cpu, whipser: 2.7.0+cpu, llama3.1-8b: 2.8.0.dev20250627+cpu", "operating_system": "Ubuntu 22.04.5", "Notes": "SYS-822GA-NGR3. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/1-node-2S-GNR_128C", "ColorKey": "availableSupermicro", "Performance_Result": 497.339, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0096", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "rgat", "Model": "rgat", "Scenario": "Offline", "Accuracy": "acc: 72.557", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Supermicro/results/1-node-2S-GNR_128C/rgat/Offline/performance/run_1", "Software": "PyTorch: rgat: 2.7.0+cpu, whipser: 2.7.0+cpu, llama3.1-8b: 2.8.0.dev20250627+cpu", "operating_system": "Ubuntu 22.04.5", "Notes": "SYS-822GA-NGR3. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/1-node-2S-GNR_128C", "ColorKey": "availableSupermicro", "Performance_Result": 25192.7, "Performance_Units": "Samples/s" }, { "ID": "6.0-0096", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "1-node-2S-GNR_128C", "Platform": "1-node-2S-GNR_128C", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.85666881011248", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6980P", "host_processors_per_node": 2, "host_processor_core_count": "128", "Accelerator": "", "a#": "", "Total Accelerators": 0, "Location": "./closed/Supermicro/results/1-node-2S-GNR_128C/whisper/Offline/performance/run_1", "Software": "PyTorch: rgat: 2.7.0+cpu, whipser: 2.7.0+cpu, llama3.1-8b: 2.8.0.dev20250627+cpu", "operating_system": "Ubuntu 22.04.5", "Notes": "SYS-822GA-NGR3. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/1-node-2S-GNR_128C", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/1-node-2S-GNR_128C", "ColorKey": "availableSupermicro", "Performance_Result": 2204.09, "Performance_Units": "Samples/s" }, { "ID": "6.0-0097", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D74A-7U", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5421 ROUGE2: 22.1361 ROUGEL: 28.9128 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/8xMI355X_2xEPYC-9575F/llama2-70b-99/Interactive/performance/run_1", "Software": "TODO", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/8xMI355X_2xEPYC-9575F", "ColorKey": "availableSupermicro", "Performance_Result": 71295.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0097", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D74A-7U", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4357 ROUGE2: 22.0707 ROUGEL: 28.8441 TOKENS_PER_SAMPLE: 284.6", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/8xMI355X_2xEPYC-9575F/llama2-70b-99/Offline/performance/run_1", "Software": "TODO", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/8xMI355X_2xEPYC-9575F", "ColorKey": "availableSupermicro", "Performance_Result": 100754.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0097", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D74A-7U", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5381 ROUGE2: 22.1334 ROUGEL: 28.9085 TOKENS_PER_SAMPLE: 283.0", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/8xMI355X_2xEPYC-9575F/llama2-70b-99/Server/performance/run_1", "Software": "TODO", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/8xMI355X_2xEPYC-9575F", "ColorKey": "availableSupermicro", "Performance_Result": 97587.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0097", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D74A-7U", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Interactive", "Accuracy": "ROUGE1: 44.5421 ROUGE2: 22.1361 ROUGEL: 28.9128 TOKENS_PER_SAMPLE: 282.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/8xMI355X_2xEPYC-9575F/llama2-70b-99.9/Interactive/performance/run_1", "Software": "TODO", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/8xMI355X_2xEPYC-9575F", "ColorKey": "availableSupermicro", "Performance_Result": 71295.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0097", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D74A-7U", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4357 ROUGE2: 22.0707 ROUGEL: 28.8441 TOKENS_PER_SAMPLE: 284.6", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/8xMI355X_2xEPYC-9575F/llama2-70b-99.9/Offline/performance/run_1", "Software": "TODO", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/8xMI355X_2xEPYC-9575F", "ColorKey": "availableSupermicro", "Performance_Result": 100754.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0097", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "QuantaGrid D74A-7U", "Platform": "8xMI355X_2xEPYC-9575F", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.5381 ROUGE2: 22.1334 ROUGEL: 28.9085 TOKENS_PER_SAMPLE: 283.0", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/8xMI355X_2xEPYC-9575F/llama2-70b-99.9/Server/performance/run_1", "Software": "TODO", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/8xMI355X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/8xMI355X_2xEPYC-9575F", "ColorKey": "availableSupermicro", "Performance_Result": 97587.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0098", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "NVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)", "Platform": "AS-8126GS-NB3RT_B300-SXM-288GB_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.88155352741323", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B300-SXM-270GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/AS-8126GS-NB3RT_B300-SXM-288GB_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B300 TGP 1100W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/AS-8126GS-NB3RT_B300-SXM-288GB_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/AS-8126GS-NB3RT_B300-SXM-288GB_TRT", "ColorKey": "availableSupermicro", "Performance_Result": 50562.2, "Performance_Units": "Samples/s" }, { "ID": "6.0-0099", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "SYS-822GS-NBRT (8x B200-SXM-180GB, TensorRT)", "Platform": "SYS-822GS-NBRT_B200-SXM-180GB_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7467 ROUGE2: 22.3565 ROUGEL: 29.1362 TOKENS_PER_SAMPLE: 273.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "ColorKey": "availableSupermicro", "Performance_Result": 101595.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0099", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "SYS-822GS-NBRT (8x B200-SXM-180GB, TensorRT)", "Platform": "SYS-822GS-NBRT_B200-SXM-180GB_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.744 ROUGE2: 22.3567 ROUGEL: 29.1286 TOKENS_PER_SAMPLE: 274.4", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT/llama2-70b-99/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "ColorKey": "availableSupermicro", "Performance_Result": 99301.5, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0099", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "SYS-822GS-NBRT (8x B200-SXM-180GB, TensorRT)", "Platform": "SYS-822GS-NBRT_B200-SXM-180GB_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7465 ROUGE2: 22.356 ROUGEL: 29.1359 TOKENS_PER_SAMPLE: 273.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT/llama2-70b-99.9/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "ColorKey": "availableSupermicro", "Performance_Result": 101526.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0099", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "SYS-822GS-NBRT (8x B200-SXM-180GB, TensorRT)", "Platform": "SYS-822GS-NBRT_B200-SXM-180GB_TRT", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7501 ROUGE2: 22.3589 ROUGEL: 29.1383 TOKENS_PER_SAMPLE: 274.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT/llama2-70b-99.9/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "ColorKey": "availableSupermicro", "Performance_Result": 99304.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0099", "Submitter": "Supermicro", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "SYS-822GS-NBRT (8x B200-SXM-180GB, TensorRT)", "Platform": "SYS-822GS-NBRT_B200-SXM-180GB_TRT", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.86918164774403", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 6767P", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT/whisper/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/results/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro/SYS-822GS-NBRT_B200-SXM-180GB_TRT", "ColorKey": "availableSupermicro", "Performance_Result": 42389.7, "Performance_Units": "Samples/s" }, { "ID": "6.0-0100", "Submitter": "Supermicro_RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AS-8126GS-TNMR,vLLM", "Platform": "8XMI350x_RHEL9", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.4988 ROUGE2: 22.1091 ROUGEL: 28.8905 TOKENS_PER_SAMPLE: 282.7", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro_RedHat/results/8XMI350x_RHEL9/llama2-70b-99/Offline/performance/run_1", "Software": "vLLM 0.14.0, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "Supermicro MI350x. RHEL 9.6, vllm 0.14.0", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro_RedHat/results/8XMI350x_RHEL9", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro_RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro_RedHat/8XMI350x_RHEL9", "ColorKey": "availableSupermicro_RedHat", "Performance_Result": 80478.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0100", "Submitter": "Supermicro_RedHat", "Availability": "available", "Category": "closed", "Suite": "datacenter", "System": "AS-8126GS-TNMR,vLLM", "Platform": "8XMI350x_RHEL9", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.6049 ROUGE2: 22.178 ROUGEL: 28.9591 TOKENS_PER_SAMPLE: 281.8", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI350X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./closed/Supermicro_RedHat/results/8XMI350x_RHEL9/llama2-70b-99/Server/performance/run_1", "Software": "vLLM 0.14.0, RHEL 9.6", "operating_system": "Red Hat Enterprise Linux 9.6", "Notes": "Supermicro MI350x. RHEL 9.6, vllm 0.14.0", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro_RedHat/results/8XMI350x_RHEL9", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Supermicro_RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Supermicro_RedHat/8XMI350x_RHEL9", "ColorKey": "availableSupermicro_RedHat", "Performance_Result": 76393.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.632", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/gpt-oss-120b/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "MXFP4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 1536.9, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 82.520", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/gpt-oss-120b/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "MXFP4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 951.674, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7654 ROUGE2: 22.2793 ROUGEL: 28.9971 TOKENS_PER_SAMPLE: 283.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/llama2-70b-99/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 2459.18, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7712 ROUGE2: 22.2839 ROUGEL: 29.0009 TOKENS_PER_SAMPLE: 283.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/llama2-70b-99/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 1698.57, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7654 ROUGE2: 22.2793 ROUGEL: 28.9971 TOKENS_PER_SAMPLE: 283.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/llama2-70b-99.9/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 2459.18, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "llama2-70b-99.9", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 44.7712 ROUGE2: 22.2839 ROUGEL: 29.0009 TOKENS_PER_SAMPLE: 283.3", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/llama2-70b-99.9/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 1698.57, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.7303 ROUGE2: 16.0979 ROUGEL: 24.6095 ROUGELSUM: 35.8705 GEN_LEN: 8149202", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/llama3_1-8b/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 4617.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "llama3_1-8b", "Model": "llama3.1-8b", "Scenario": "Server", "Accuracy": "ROUGE1: 38.711 ROUGE2: 16.0552 ROUGEL: 24.5943 ROUGELSUM: 35.8304 GEN_LEN: 8147007", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/llama3_1-8b/Server/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT8", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 4170.13, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0101", "Submitter": "Intel", "Availability": "preview", "Category": "closed", "Suite": "datacenter", "System": "1-node-4x-BMG-B70", "Platform": "1-node-4x-BMG-B70", "UsedModel": "whisper", "Model": "whisper", "Scenario": "Offline", "Accuracy": "ACCURACY: 97.46051379638439", "Nodes": 1, "Processor": "Intel(R) Xeon(R) 698X", "host_processors_per_node": 1, "host_processor_core_count": "86", "Accelerator": "Intel(R) Arc Pro(R) B70", "a#": 4, "Total Accelerators": 4, "Location": "./closed/Intel/results/1-node-4x-BMG-B70/whisper/Offline/performance/run_1", "Software": "PyTorch vLLM", "operating_system": "Ubuntu 24.04.3 LTS", "Notes": "N/A. N/A", "compliance": "closed", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "INT4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/results/1-node-4x-BMG-B70", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/Intel/src", "Unique ID (e.g. for Audit)": "datacenter/closed/Intel/1-node-4x-BMG-B70", "ColorKey": "previewIntel", "Performance_Result": 5596.6, "Performance_Units": "Samples/s" }, { "ID": "6.0-0102", "Submitter": "AMD", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Interactive", "Accuracy": "ROUGEL: 21.51398069082574 exact_match: 86.51147584381204 TOKENS_PER_SAMPLE: 756.7", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./open/AMD/results/8xMI355X_2xEPYC_9575F/llama3.1-405b/Interactive/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0, PyTorch 2.9.0a0+git1c57644, ROCm 7.1.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/open/AMD/8xMI355X_2xEPYC_9575F/llama3.1-405b", "ColorKey": "availableAMD", "Performance_Result": 793.994, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0102", "Submitter": "AMD", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "smci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)", "Platform": "8xMI355X_2xEPYC_9575F", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.545066368325006 exact_match: 82.66624751819988 TOKENS_PER_SAMPLE: 758.9", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 8, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3e", "a#": 8, "Total Accelerators": 8, "Location": "./open/AMD/results/8xMI355X_2xEPYC_9575F/llama3.1-405b/Offline/performance/run_1", "Software": "PyTorch 2.9.1+git8907517, ROCm 7.0.0, PyTorch 2.9.0a0+git1c57644, ROCm 7.1.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 8, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/AMD/results/8xMI355X_2xEPYC_9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/open/AMD/8xMI355X_2xEPYC_9575F/llama3.1-405b", "ColorKey": "availableAMD", "Performance_Result": 2011.17, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0103", "Submitter": "AMD", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "SuperMicro AS-4126GS-NMR-LCC", "Platform": "8xMI355X_2xEPYC_9597F", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 69.98", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "AMD Instinct MI355X 288GB HBM3E", "a#": 8, "Total Accelerators": 8, "Location": "./open/AMD/results/8xMI355X_2xEPYC_9597F/wan-2.2-t2v-a14b/SingleStream/performance/run_1", "Software": "xDiT, xfuser 0.4.5+git6ffc551, Pytorch 2.10.0+gitd9556b0, ROCm 7.11.0", "operating_system": "Ubuntu 22.04.5 LTS", "Notes": "", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp8,bf16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/AMD/results/8xMI355X_2xEPYC_9597F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/AMD/src", "Unique ID (e.g. for Audit)": "datacenter/open/AMD/8xMI355X_2xEPYC_9597F/wan-2.2-t2v-a14b", "ColorKey": "availableAMD", "Performance_Result": 27.384519297, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0104", "Submitter": "CoreWeave", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "CoreWeave GB300 NVL72 (64x GB300-288GB_aarch64, TensorRT)", "Platform": "GB300-NVL72_GB300-288GB_aarch64x64_TRT", "UsedModel": "deepseek-r1", "Model": "deepseek-r1", "Scenario": "Offline", "Accuracy": "exact_match: 81.42661804922516 TOKENS_PER_SAMPLE: 3784.834092980857", "Nodes": 16, "Processor": "NVIDIA Grace CPU", "host_processors_per_node": 2, "host_processor_core_count": "144", "Accelerator": "NVIDIA GB300", "a#": 4, "Total Accelerators": 64, "Location": "./open/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x64_TRT/deepseek-r1/Offline/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "GB300 NVL72", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/CoreWeave/results/GB300-NVL72_GB300-288GB_aarch64x64_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/CoreWeave/src", "Unique ID (e.g. for Audit)": "datacenter/open/CoreWeave/GB300-NVL72_GB300-288GB_aarch64x64_TRT/deepseek-r1", "ColorKey": "availableCoreWeave", "Performance_Result": 627653.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0105", "Submitter": "Inventec", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "Inventec P9000AG7 (8x B200-SXM-180GB, TensorRT)", "Platform": "P9000AG7_B200-SXM-180GBx8_TRT", "UsedModel": "llama2-70b-99", "Model": "llama2-70b-99", "Scenario": "Offline", "Accuracy": "ROUGE1: 44.7474 ROUGE2: 22.3568 ROUGEL: 29.1367 TOKENS_PER_SAMPLE: 273.3", "Nodes": 1, "Processor": "AMD EPYC 9A75", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Inventec/results/P9000AG7_B200-SXM-180GBx8_TRT/llama2-70b-99/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Inventec/results/P9000AG7_B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Inventec/src", "Unique ID (e.g. for Audit)": "datacenter/open/Inventec/P9000AG7_B200-SXM-180GBx8_TRT/llama2-70b-99", "ColorKey": "availableInventec", "Performance_Result": 100291.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0106", "Submitter": "Inventec", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "Inventec P9000AG7 (8x B200-SXM-180GB, TensorRT)", "Platform": "P9000AG7_B200-SXM-180GBx8_TRT", "UsedModel": "llama3.1-405b", "Model": "llama3.1-405b", "Scenario": "Offline", "Accuracy": "ROUGEL: 21.979457539013907 exact_match: 90.03752812706816 TOKENS_PER_SAMPLE: 652.0", "Nodes": 1, "Processor": "AMD EPYC 9A75", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Inventec/results/P9000AG7_B200-SXM-180GBx8_TRT/llama3.1-405b/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Inventec/results/P9000AG7_B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Inventec/src", "Unique ID (e.g. for Audit)": "datacenter/open/Inventec/P9000AG7_B200-SXM-180GBx8_TRT/llama3.1-405b", "ColorKey": "availableInventec", "Performance_Result": 1632.1, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0107", "Submitter": "Inventec", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "Inventec P9000AG7 (8x B200-SXM-180GB, TensorRT)", "Platform": "P9000AG7_B200-SXM-180GBx8_TRT", "UsedModel": "llama3.1-8b", "Model": "llama3.1-8b", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.4843 ROUGE2: 15.9098 ROUGEL: 24.3232 ROUGELSUM: 35.6313 GEN_LEN: 8151646", "Nodes": 1, "Processor": "AMD EPYC 9A75", "host_processors_per_node": 2, "host_processor_core_count": "64", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Inventec/results/P9000AG7_B200-SXM-180GBx8_TRT/llama3.1-8b/Offline/performance/run_1", "Software": "TensorRT 10.11, CUDA 12.9", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Inventec/results/P9000AG7_B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Inventec/src", "Unique ID (e.g. for Audit)": "datacenter/open/Inventec/P9000AG7_B200-SXM-180GBx8_TRT/llama3.1-8b", "ColorKey": "availableInventec", "Performance_Result": 142508.0, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0108", "Submitter": "Krai", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "HPE Cray XD670 (8x H200-SXM-141GB, KISS-V Fast Mode)", "Platform": "h200_n1-kissv", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "Offline", "Accuracy": "vbench_score: 70.09", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "32", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Krai/results/h200_n1-kissv/wan-2.2-t2v-a14b/offline/performance/run_1", "Software": "KISS-V: Krai Inference Serving Solution for Video (Fast Mode)", "operating_system": "Red Hat Enterprise Linux release 9.4 (Plow) (Linux kernel: 5.14.0-427.13.1.el9_4.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Apr 10 10:29:16 EDT 2024 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "Powered by the Krai axs automation technology", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "bfloat16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/results/h200_n1-kissv", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/src", "Unique ID (e.g. for Audit)": "datacenter/open/Krai/h200_n1-kissv/wan-2.2-t2v-a14b", "ColorKey": "availableKrai", "Performance_Result": 0.0241194, "Performance_Units": "Samples/s" }, { "ID": "6.0-0108", "Submitter": "Krai", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "HPE Cray XD670 (8x H200-SXM-141GB, KISS-V Fast Mode)", "Platform": "h200_n1-kissv", "UsedModel": "wan-2.2-t2v-a14b", "Model": "wan-2.2-t2v-a14b", "Scenario": "SingleStream", "Accuracy": "vbench_score: 69.78999999999999", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "32", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Krai/results/h200_n1-kissv/wan-2.2-t2v-a14b/singlestream/performance/run_1", "Software": "KISS-V: Krai Inference Serving Solution for Video (Fast Mode)", "operating_system": "Red Hat Enterprise Linux release 9.4 (Plow) (Linux kernel: 5.14.0-427.13.1.el9_4.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Apr 10 10:29:16 EDT 2024 x86_64 x86_64 x86_64 GNU/Linux)", "Notes": "Powered by the Krai axs automation technology", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "bfloat16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/results/h200_n1-kissv", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/src", "Unique ID (e.g. for Audit)": "datacenter/open/Krai/h200_n1-kissv/wan-2.2-t2v-a14b", "ColorKey": "availableKrai", "Performance_Result": 43.936000879, "Performance_Units": "Latency (s)" }, { "ID": "6.0-0109", "Submitter": "Krai", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "HPE Cray XD670 (8x H200-SXM-141GB, vLLM)", "Platform": "h200_n1-vllm", "UsedModel": "llama3_1-70b-fp8_dyn", "Model": "llama2-70b-99.9", "Scenario": "Offline", "Accuracy": "ROUGE1: 47.6127 ROUGE2: 25.255 ROUGEL: 31.5076 TOKENS_PER_SAMPLE: 298.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "32", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Krai/results/h200_n1-vllm/llama3_1-70b-fp8_dyn/offline/performance/run_1", "Software": "vLLM v0.15.0-dev", "operating_system": "host: Red Hat Enterprise Linux 9.4, container: Ubuntu 22.04.5 LTS", "Notes": "Powered by the Krai axs automation technology", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/results/h200_n1-vllm", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/src", "Unique ID (e.g. for Audit)": "datacenter/open/Krai/h200_n1-vllm/llama3_1-70b-fp8_dyn", "ColorKey": "availableKrai", "Performance_Result": 25939.8, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0109", "Submitter": "Krai", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "HPE Cray XD670 (8x H200-SXM-141GB, vLLM)", "Platform": "h200_n1-vllm", "UsedModel": "llama3_1-70b-fp8_dyn", "Model": "llama2-70b-99.9", "Scenario": "Server", "Accuracy": "ROUGE1: 47.6215 ROUGE2: 25.2626 ROUGEL: 31.5143 TOKENS_PER_SAMPLE: 298.2", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8562Y+", "host_processors_per_node": 2, "host_processor_core_count": "32", "Accelerator": "NVIDIA H200-SXM-141GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Krai/results/h200_n1-vllm/llama3_1-70b-fp8_dyn/server/performance/run_1", "Software": "vLLM v0.15.0-dev", "operating_system": "host: Red Hat Enterprise Linux 9.4, container: Ubuntu 22.04.5 LTS", "Notes": "Powered by the Krai axs automation technology", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 1, "has_power": false, "weight_data_types": "fp16", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/results/h200_n1-vllm", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Krai/src", "Unique ID (e.g. for Audit)": "datacenter/open/Krai/h200_n1-vllm/llama3_1-70b-fp8_dyn", "ColorKey": "availableKrai", "Performance_Result": 24821.6, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0110", "Submitter": "Lambda_SIT", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "NVIDIA DGX B200 (8x B200-SXM-180GB, TensorRT)", "Platform": "B200-SXM-180GBx8_TRT", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.661", "Nodes": 1, "Processor": "Intel(R) Xeon(R) Platinum 8570", "host_processors_per_node": 2, "host_processor_core_count": "56", "Accelerator": "NVIDIA B200-SXM-180GB", "a#": 8, "Total Accelerators": 8, "Location": "./open/Lambda_SIT/results/B200-SXM-180GBx8_TRT/gpt-oss-120b/Server/performance/run_1", "Software": "TensorRT 10.14, CUDA 13.1", "operating_system": "Ubuntu 24.04", "Notes": "B200 TGP 1000W", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Lambda_SIT/results/B200-SXM-180GBx8_TRT", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Lambda_SIT/src", "Unique ID (e.g. for Audit)": "datacenter/open/Lambda_SIT/B200-SXM-180GBx8_TRT/gpt-oss-120b", "ColorKey": "availableLambda_SIT", "Performance_Result": 16167.4, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0111", "Submitter": "NVIDIA", "Availability": "available", "Category": "open", "Suite": "edge", "System": "NVIDIA Jetson AGX Thor Developer Kit 128G (TensorRT)", "Platform": "Jetson-Thor", "UsedModel": "llama3.1-8b-edge", "Model": "llama3.1-8b-edge", "Scenario": "Offline", "Accuracy": "ROUGE1: 38.5299 ROUGE2: 16.2005 ROUGEL: 24.5013 ROUGELSUM: 35.6685 GEN_LEN: 3089315", "Nodes": 1, "Processor": "14-core ARM Poseidon-AE CPU", "host_processors_per_node": 1, "host_processor_core_count": "14", "Accelerator": "NVIDIA Jetson AGX Thor 128G", "a#": 1, "Total Accelerators": 1, "Location": "./open/NVIDIA/results/Jetson-Thor/llama3.1-8b-edge/Offline/performance/run_1", "Software": "TensorRT 10.8, CUDA 12.8", "operating_system": "Jetson L4T", "Notes": "", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/NVIDIA/results/Jetson-Thor", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/NVIDIA/src", "Unique ID (e.g. for Audit)": "edge/open/NVIDIA/Jetson-Thor/llama3.1-8b-edge", "ColorKey": "availableNVIDIA", "Performance_Result": 350.605, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0111", "Submitter": "NVIDIA", "Availability": "available", "Category": "open", "Suite": "edge", "System": "NVIDIA Jetson AGX Thor Developer Kit 128G (TensorRT)", "Platform": "Jetson-Thor", "UsedModel": "llama3.1-8b-edge", "Model": "llama3.1-8b-edge", "Scenario": "SingleStream", "Accuracy": "ROUGE1: 38.3796 ROUGE2: 15.9406 ROUGEL: 24.3855 ROUGELSUM: 35.5168 GEN_LEN: 3112369", "Nodes": 1, "Processor": "14-core ARM Poseidon-AE CPU", "host_processors_per_node": 1, "host_processor_core_count": "14", "Accelerator": "NVIDIA Jetson AGX Thor 128G", "a#": 1, "Total Accelerators": 1, "Location": "./open/NVIDIA/results/Jetson-Thor/llama3.1-8b-edge/SingleStream/performance/run_1", "Software": "TensorRT 10.8, CUDA 12.8", "operating_system": "Jetson L4T", "Notes": "", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 1, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/NVIDIA/results/Jetson-Thor", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/NVIDIA/src", "Unique ID (e.g. for Audit)": "edge/open/NVIDIA/Jetson-Thor/llama3.1-8b-edge", "ColorKey": "availableNVIDIA", "Performance_Result": 1357.552837, "Performance_Units": "Latency (ms)" }, { "ID": "6.0-0112", "Submitter": "Supermicro_RedHat", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "AS-8126GS-TNMR,vLLM", "Platform": "8xMI350X_2xEPYC-9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Offline", "Accuracy": "exact_match: 83.568", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "1", "Accelerator": "AMD Instinct MI350X", "a#": 8, "Total Accelerators": 8, "Location": "./open/Supermicro_RedHat/results/8xMI350X_2xEPYC-9575F/gpt-oss-120b/Offline/performance/run_1", "Software": "", "operating_system": "", "Notes": "SMC MI350X AS -8126GS-TNMR. vLLM 0.14.0, RHEL 9.6", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Supermicro_RedHat/results/8xMI350X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Supermicro_RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/open/Supermicro_RedHat/8xMI350X_2xEPYC-9575F/gpt-oss-120b", "ColorKey": "availableSupermicro_RedHat", "Performance_Result": 64293.3, "Performance_Units": "Tokens/s" }, { "ID": "6.0-0112", "Submitter": "Supermicro_RedHat", "Availability": "available", "Category": "open", "Suite": "datacenter", "System": "AS-8126GS-TNMR,vLLM", "Platform": "8xMI350X_2xEPYC-9575F", "UsedModel": "gpt-oss-120b", "Model": "gpt-oss-120b", "Scenario": "Server", "Accuracy": "exact_match: 83.001", "Nodes": 1, "Processor": "AMD EPYC 9575F", "host_processors_per_node": 2, "host_processor_core_count": "1", "Accelerator": "AMD Instinct MI350X", "a#": 8, "Total Accelerators": 8, "Location": "./open/Supermicro_RedHat/results/8xMI350X_2xEPYC-9575F/gpt-oss-120b/Server/performance/run_1", "Software": "", "operating_system": "", "Notes": "SMC MI350X AS -8126GS-TNMR. vLLM 0.14.0, RHEL 9.6", "compliance": "open", "errors": 0, "version": "v6.0", "inferred": 0, "has_power": false, "weight_data_types": "fp4", "p#": 2, "Details": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Supermicro_RedHat/results/8xMI350X_2xEPYC-9575F", "Code": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/open/Supermicro_RedHat/src", "Unique ID (e.g. for Audit)": "datacenter/open/Supermicro_RedHat/8xMI350X_2xEPYC-9575F/gpt-oss-120b", "ColorKey": "availableSupermicro_RedHat", "Performance_Result": 58373.3, "Performance_Units": "Tokens/s" } ]