# OpenROAD Physical Design Results Run date: 2026-07-02 Design: `openMSP430` Platform: `sky130hd` Flow: OpenROAD-flow-scripts Docker image `openroad/orfs:latest` ## Closed Baseline Variant The clean baseline physical-design run is `util30_ar10_pd040_pad1`, using: - `CORE_UTILIZATION=30` - `CORE_ASPECT_RATIO=1.0` - `CORE_MARGIN=8` - `PLACE_DENSITY=0.40` - `CELL_PAD_IN_SITES_GLOBAL_PLACEMENT=1` - `CELL_PAD_IN_SITES_DETAIL_PLACEMENT=1` This run completed synthesis, floorplan, placement, CTS, global route, detailed route, filler insertion, final report generation, and DEF-to-GDS merge. | Metric | Value | | --- | ---: | | Final setup WNS at 50 ns | 35.555 ns | | Final hold WNS | 0.231 ns | | Final TNS | 0.000 ns | | Estimated Fmax | 69.226 MHz | | Die area | 203,527 um2 | | Core area | 187,999 um2 | | Placed standard-cell area | 66,966.7 um2 | | Final utilization | 35.621% | | Global-route wirelength | 496,869 um | | Detailed-route wirelength | 362,584 um | | Detailed-route DRC | 0 | | Antenna violations | 0 net / 0 pin | | Final total power | 0.004505 W | | Worst VDD IR drop | 0.00008075 V | Generated final artifacts are checked in under: - `physical-design/results/final/util30_ar10_pd040_pad1/6_final.gds` - `physical-design/results/final/util30_ar10_pd040_pad1/6_final.def.gz` - `physical-design/results/final/util30_ar10_pd040_pad1/6_final.v.gz` - `physical-design/results/final/util30_ar10_pd040_pad1/6_final.spef.gz` - `physical-design/results/final/util30_ar10_pd040_pad1/6_final.sdc` ## Routability Exploration The initial baseline and intermediate floorplans met timing and global-route congestion targets, but detailed routing exposed met1/met2 pin-access and short violations. Lowering utilization alone improved the DRC count but did not fully close the design. Adding one-site placement padding at 30% target utilization gave the router enough local whitespace to converge. | Variant | Outcome | | --- | --- | | `base` | Global route completed with 0 global-route violations; detailed route was congested at the default 65% target utilization. | | `util45_ar10_pd050` | Global route completed; detailed route still produced thousands of DRC violations. | | `util35_ar10_pd045` | Improved detailed-route DRC compared with 45%, but did not close cleanly. | | `util35_ar10_pd045_m2` | Raising the minimum routing layer to met2 worsened placement/routability pressure, so it was dropped. | | `util30_ar10_pd040_pad1` | Completed final GDS with 0 DRC and 0 antenna violations. | OpenROAD screenshots for the clean final run are in `physical-design/results/images/util30_ar10_pd040_pad1/`. ![Final placement](images/util30_ar10_pd040_pad1/final_placement.webp) ![Final routing](images/util30_ar10_pd040_pad1/final_routing.webp) Blank cells in `results.csv` mean the flow did not reach the stage that emits that metric for the variant (e.g. no `drc_count` for runs whose detailed route was aborted); values are never guessed. ## Implementation Notes - The SDC uses `dco_clk`, the real top-level clock input, rather than generated downstream clock outputs. - ORFS Kepler LEC is disabled in this repo config because the current Docker image can auto-enable it when the binary is present; under the local amd64 Docker emulation setup it crashed before CTS. The physical-design PPA and routing closure flow is otherwise complete. - Raw ORFS work directories are kept under `physical-design/orfs-work/` and are gitignored. The repository tracks the design config, run scripts, summary CSV, selected reports/images, and compact final artifacts.