25 June 2010 11:54:25 AM FFT_OPEN_MP C++/OpenMP version Demonstrate an implementation of the Fast Fourier Transform of a complex data vector, using OpenMP for parallel execution. Number of processors available = 8 Number of threads = 1 Accuracy check: FFT ( FFT ( X(1:N) ) ) == N * X(1:N) N NITS Error Time Time/Call MFLOPS 2 10000 7.85908e-17 1 6.51253e-07 15.355 4 10000 1.20984e-16 1 1.25894e-06 31.7727 8 10000 6.8208e-17 1 1.51742e-06 79.0816 16 10000 1.43867e-16 1 2.60906e-06 122.649 32 1000 1.31163e-16 1 4.50722e-06 177.493 64 1000 1.79181e-16 1 9.07248e-06 211.629 128 1000 1.89871e-16 1 1.71339e-05 261.469 256 1000 2.07082e-16 1 3.73111e-05 274.449 512 100 1.83841e-16 1 7.64694e-05 301.297 1024 100 2.30037e-16 1 0.000172903 296.12 2048 100 2.41757e-16 1 0.000361061 311.969 4096 100 2.49034e-16 1 0.000805384 305.147 8192 10 2.576e-16 1 0.00168413 316.176 16384 10 2.73634e-16 1 0.00371468 308.742 32768 10 2.90122e-16 1 0.00779025 315.471 65536 10 2.86232e-16 1 0.0168233 311.644 131072 1 3.13444e-16 1 0.0348283 319.887 262144 1 3.17945e-16 1 0.0762147 309.559 524288 1 3.27674e-16 1 0.158222 314.793 1048576 1 3.29127e-16 1 0.339533 308.829 2097152 1 3.4937e-16 1 0.696257 316.264 4194304 1 3.57571e-16 1 1.48358 310.987 8388608 1 3.69676e-16 1 3.04008 317.323 16777216 1 3.66367e-16 1 6.44959 312.154 33554432 1 3.81862e-16 1 13.1821 318.181 FFT_OPEN_MP: Normal end of execution. 25 June 2010 11:56:21 AM 25 June 2010 11:56:21 AM FFT_OPEN_MP C++/OpenMP version Demonstrate an implementation of the Fast Fourier Transform of a complex data vector, using OpenMP for parallel execution. Number of processors available = 8 Number of threads = 2 Accuracy check: FFT ( FFT ( X(1:N) ) ) == N * X(1:N) N NITS Error Time Time/Call MFLOPS 2 10000 7.85908e-17 1 2.13519e-05 0.468343 4 10000 1.20984e-16 1 4.41584e-05 0.905829 8 10000 6.8208e-17 1 5.69487e-05 2.10716 16 10000 1.43867e-16 1 8.00575e-05 3.99713 32 1000 1.31163e-16 1 9.30744e-05 8.59528 64 1000 1.79181e-16 1 0.000111446 17.2281 128 1000 1.89871e-16 1 0.000135141 33.1505 256 1000 2.07082e-16 1 0.000152492 67.1513 512 100 1.83841e-16 1 0.000188309 122.352 1024 100 2.30037e-16 1 0.000256434 199.661 2048 100 2.41757e-16 1 0.000369352 304.967 4096 100 2.49034e-16 1 0.0008638 284.51 8192 10 2.576e-16 1 0.00143353 371.448 16384 10 2.73634e-16 1 0.00270206 424.447 32768 10 2.90122e-16 1 0.00496684 494.801 65536 10 2.86232e-16 1 0.0103317 507.455 131072 1 3.13444e-16 1 0.0202274 550.794 262144 1 3.17945e-16 1 0.0438087 538.545 524288 1 3.27674e-16 1 0.0877284 567.745 1048576 1 3.29127e-16 1 0.199114 526.622 2097152 1 3.4937e-16 1 0.376194 585.339 4194304 1 3.57571e-16 1 0.815331 565.873 8388608 1 3.69676e-16 1 1.64387 586.84 16777216 1 3.66367e-16 1 3.52529 571.092 33554432 1 3.81862e-16 1 7.03497 596.208 FFT_OPEN_MP: Normal end of execution. 25 June 2010 11:57:33 AM 25 June 2010 11:57:33 AM FFT_OPEN_MP C++/OpenMP version Demonstrate an implementation of the Fast Fourier Transform of a complex data vector, using OpenMP for parallel execution. Number of processors available = 8 Number of threads = 4 Accuracy check: FFT ( FFT ( X(1:N) ) ) == N * X(1:N) N NITS Error Time Time/Call MFLOPS 2 10000 7.85908e-17 1 2.81734e-05 0.354945 4 10000 1.20984e-16 1 6.07309e-05 0.658643 8 10000 6.8208e-17 1 8.65341e-05 1.38674 16 10000 1.43867e-16 1 0.000127123 2.51726 32 1000 1.31163e-16 1 0.000176689 4.52774 64 1000 1.79181e-16 1 0.000176928 10.8519 128 1000 1.89871e-16 1 0.000193024 23.2096 256 1000 2.07082e-16 1 0.000238461 42.9421 512 100 1.83841e-16 1 0.00029021 79.3909 1024 100 2.30037e-16 1 0.000369605 138.526 2048 100 2.41757e-16 1 0.000558664 201.624 4096 100 2.49034e-16 1 0.000837055 293.601 8192 10 2.576e-16 1 0.00122439 434.893 16384 10 2.73634e-16 1 0.00204062 562.026 32768 10 2.90122e-16 1 0.00331053 742.358 65536 10 2.86232e-16 1 0.00668689 784.053 131072 1 3.13444e-16 1 0.0125207 889.817 262144 1 3.17945e-16 1 0.0273608 862.29 524288 1 3.27674e-16 1 0.0524647 949.35 1048576 1 3.29127e-16 1 0.120074 873.277 2097152 1 3.4937e-16 1 0.224076 982.707 4194304 1 3.57571e-16 1 0.507378 909.329 8388608 1 3.69676e-16 1 0.937157 1029.38 16777216 1 3.66367e-16 1 2.07935 968.218 33554432 1 3.81862e-16 1 3.95356 1060.89 FFT_OPEN_MP: Normal end of execution. 25 June 2010 11:58:23 AM 25 June 2010 11:58:23 AM FFT_OPEN_MP C++/OpenMP version Demonstrate an implementation of the Fast Fourier Transform of a complex data vector, using OpenMP for parallel execution. Number of processors available = 8 Number of threads = 8 Accuracy check: FFT ( FFT ( X(1:N) ) ) == N * X(1:N) N NITS Error Time Time/Call MFLOPS 2 10000 7.85908e-17 1 3.76028e-05 0.265938 4 10000 1.20984e-16 1 6.29345e-05 0.635581 8 10000 6.8208e-17 1 9.99141e-05 1.20103 16 10000 1.43867e-16 1 0.000182598 1.75248 32 1000 1.31163e-16 1 0.000192952 4.14611 64 1000 1.79181e-16 1 0.000242489 7.91789 128 1000 1.89871e-16 1 0.000306251 14.6285 256 1000 2.07082e-16 1 0.000350655 29.2025 512 100 1.83841e-16 1 0.000400068 57.5902 1024 100 2.30037e-16 1 0.000492428 103.975 2048 100 2.41757e-16 1 0.000569026 197.952 4096 100 2.49034e-16 1 0.000759611 323.534 8192 10 2.576e-16 1 0.00109841 484.775 16384 10 2.73634e-16 1 0.00208653 549.659 32768 10 2.90122e-16 1 0.00481199 510.724 65536 10 2.86232e-16 1 0.00887547 590.716 131072 1 3.13444e-16 1 0.0122166 911.966 262144 1 3.17945e-16 1 0.0301132 783.476 524288 1 3.27674e-16 1 0.0561447 887.126 1048576 1 3.29127e-16 1 0.0819924 1278.87 2097152 1 3.4937e-16 1 0.167378 1315.59 4194304 1 3.57571e-16 1 0.363582 1268.97 8388608 1 3.69676e-16 1 0.646349 1492.52 16777216 1 3.66367e-16 1 1.4501 1388.36 33554432 1 3.81862e-16 1 2.74365 1528.73 FFT_OPEN_MP: Normal end of execution. 25 June 2010 11:59:06 AM