Skip to content

Instantly share code, notes, and snippets.

@asi1024

asi1024/8d6317313

Created December 4, 2019 09:25
Show Gist options
  • Save asi1024/2b5d8d119a6ef82421e62893157dd468 to your computer and use it in GitHub Desktop.
Save asi1024/2b5d8d119a6ef82421e62893157dd468 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
8d6317313
cupy.sum (shape = ( 1, 16777216), axis=0): 14.256 us +/- 0.705 (min: 13.582 / max: 24.281) us 188.497 us +/- 1.241 (min: 179.552 / max: 198.432) us
cupy.sum (shape = ( 2, 8388608), axis=0): 14.646 us +/- 0.579 (min: 14.041 / max: 20.914) us 178.568 us +/- 0.724 (min: 168.384 / max: 184.640) us
cupy.sum (shape = ( 4, 4194304), axis=0): 14.906 us +/- 0.518 (min: 14.368 / max: 20.649) us 187.801 us +/- 0.709 (min: 177.792 / max: 193.312) us
cupy.sum (shape = ( 8, 2097152), axis=0): 14.831 us +/- 0.871 (min: 14.219 / max: 25.082) us 199.579 us +/- 0.969 (min: 189.184 / max: 209.696) us
cupy.sum (shape = ( 16, 1048576), axis=0): 14.718 us +/- 0.541 (min: 14.216 / max: 22.016) us 213.216 us +/- 0.704 (min: 203.104 / max: 220.000) us
cupy.sum (shape = ( 32, 524288), axis=0): 14.761 us +/- 1.614 (min: 14.130 / max: 48.321) us 143.380 us +/- 1.984 (min: 133.248 / max: 182.848) us
cupy.sum (shape = ( 64, 262144), axis=0): 14.711 us +/- 0.346 (min: 14.253 / max: 18.220) us 119.760 us +/- 0.717 (min: 110.432 / max: 122.784) us
cupy.sum (shape = ( 128, 131072), axis=0): 14.759 us +/- 0.507 (min: 14.257 / max: 22.780) us 108.374 us +/- 0.774 (min: 97.920 / max: 115.968) us
cupy.sum (shape = ( 256, 65536), axis=0): 14.732 us +/- 0.360 (min: 14.264 / max: 17.457) us 103.740 us +/- 0.682 (min: 93.536 / max: 106.432) us
cupy.sum (shape = ( 512, 32768), axis=0): 14.862 us +/- 0.407 (min: 14.424 / max: 19.711) us 106.879 us +/- 0.771 (min: 97.152 / max: 113.024) us
cupy.sum (shape = ( 1024, 16384), axis=0): 14.735 us +/- 0.328 (min: 14.264 / max: 17.941) us 105.881 us +/- 0.753 (min: 95.584 / max: 108.416) us
cupy.sum (shape = ( 2048, 8192), axis=0): 14.893 us +/- 0.805 (min: 14.395 / max: 27.005) us 103.560 us +/- 0.990 (min: 93.440 / max: 112.640) us
cupy.sum (shape = ( 4096, 4096), axis=0): 14.816 us +/- 0.444 (min: 14.283 / max: 21.789) us 146.516 us +/- 1.114 (min: 135.392 / max: 151.232) us
cupy.sum (shape = ( 8192, 2048), axis=0): 15.312 us +/-10.241 (min: 14.257 / max: 243.574) us 231.259 us +/-10.416 (min: 222.048 / max: 461.504) us
cupy.sum (shape = ( 16384, 1024), axis=0): 14.940 us +/- 0.634 (min: 14.301 / max: 22.478) us 429.440 us +/- 1.566 (min: 418.880 / max: 437.568) us
cupy.sum (shape = ( 32768, 512), axis=0): 15.064 us +/- 0.552 (min: 14.309 / max: 18.973) us 816.746 us +/- 2.045 (min: 811.488 / max: 824.032) us
cupy.sum (shape = ( 65536, 256), axis=0): 15.262 us +/- 0.657 (min: 14.435 / max: 21.007) us 1585.463 us +/- 2.595 (min: 1576.000 / max: 1596.608) us
cupy.sum (shape = ( 131072, 128), axis=0): 16.076 us +/- 9.189 (min: 14.499 / max: 220.366) us 3125.623 us +/-10.385 (min: 3114.336 / max: 3336.640) us
cupy.sum (shape = ( 262144, 64), axis=0): 16.205 us +/- 0.648 (min: 15.263 / max: 23.484) us 6169.057 us +/- 8.066 (min: 6139.584 / max: 6187.744) us
cupy.sum (shape = ( 524288, 32), axis=0): 17.037 us +/- 9.697 (min: 14.886 / max: 233.240) us 12186.044 us +/-19.018 (min:12134.528 / max:12402.816) us
cupy.sum (shape = ( 1048576, 16), axis=0): 17.044 us +/- 8.992 (min: 15.263 / max: 217.155) us 12184.717 us +/-19.006 (min:12141.824 / max:12379.808) us
cupy.sum (shape = ( 2097152, 8), axis=0): 16.625 us +/- 0.394 (min: 15.974 / max: 21.089) us 12185.818 us +/-15.279 (min:12128.480 / max:12230.560) us
cupy.sum (shape = ( 4194304, 4), axis=0): 31.137 us +/-162.523 (min: 15.065 / max: 3599.126) us 12266.655 us +/-282.072 (min:12134.240 / max:15707.616) us
cupy.sum (shape = ( 8388608, 2), axis=0): 18.097 us +/- 9.320 (min: 15.210 / max: 224.676) us 12736.687 us +/-501.499 (min:12148.800 / max:13444.320) us
cupy.sum (shape = (16777216, 1), axis=0): 17.324 us +/- 1.085 (min: 15.285 / max: 27.154) us 12350.245 us +/-381.815 (min:12144.640 / max:13429.536) us
cupy.sum (shape = ( 1, 16777216), axis=1): 18.443 us +/- 1.060 (min: 16.022 / max: 27.265) us 12186.097 us +/-16.402 (min:12135.296 / max:12234.752) us
cupy.sum (shape = ( 2, 8388608), axis=1): 17.000 us +/- 0.754 (min: 15.929 / max: 24.583) us 7414.979 us +/- 8.550 (min: 7388.800 / max: 7447.840) us
cupy.sum (shape = ( 4, 4194304), axis=1): 16.543 us +/- 0.736 (min: 15.374 / max: 25.202) us 3737.531 us +/- 4.364 (min: 3722.944 / max: 3752.800) us
cupy.sum (shape = ( 8, 2097152), axis=1): 16.274 us +/- 1.029 (min: 15.169 / max: 23.991) us 1894.360 us +/- 3.323 (min: 1880.768 / max: 1905.856) us
cupy.sum (shape = ( 16, 1048576), axis=1): 15.802 us +/- 0.573 (min: 15.110 / max: 19.826) us 962.597 us +/- 2.177 (min: 951.008 / max: 970.624) us
cupy.sum (shape = ( 32, 524288), axis=1): 15.733 us +/- 0.841 (min: 14.957 / max: 26.934) us 499.898 us +/- 1.676 (min: 488.864 / max: 509.216) us
cupy.sum (shape = ( 64, 262144), axis=1): 15.448 us +/- 0.554 (min: 14.775 / max: 21.849) us 262.028 us +/- 1.112 (min: 251.552 / max: 268.736) us
cupy.sum (shape = ( 128, 131072), axis=1): 15.486 us +/- 0.493 (min: 14.912 / max: 20.079) us 163.952 us +/- 1.113 (min: 153.408 / max: 167.776) us
cupy.sum (shape = ( 256, 65536), axis=1): 15.452 us +/- 0.521 (min: 14.856 / max: 23.175) us 157.041 us +/- 1.026 (min: 146.016 / max: 161.760) us
cupy.sum (shape = ( 512, 32768), axis=1): 15.510 us +/- 0.465 (min: 15.009 / max: 20.597) us 156.881 us +/- 1.409 (min: 144.384 / max: 171.776) us
cupy.sum (shape = ( 1024, 16384), axis=1): 15.353 us +/- 0.347 (min: 14.678 / max: 17.878) us 146.936 us +/- 0.745 (min: 136.352 / max: 149.088) us
cupy.sum (shape = ( 2048, 8192), axis=1): 15.585 us +/- 0.900 (min: 14.983 / max: 24.859) us 150.195 us +/- 1.087 (min: 139.200 / max: 159.040) us
cupy.sum (shape = ( 4096, 4096), axis=1): 15.416 us +/- 0.391 (min: 14.964 / max: 19.412) us 171.991 us +/- 0.819 (min: 160.864 / max: 175.840) us
cupy.sum (shape = ( 8192, 2048), axis=1): 15.537 us +/- 0.666 (min: 14.912 / max: 22.735) us 204.149 us +/- 0.813 (min: 193.024 / max: 211.072) us
cupy.sum (shape = ( 16384, 1024), axis=1): 15.567 us +/- 0.648 (min: 14.950 / max: 21.197) us 283.519 us +/- 0.790 (min: 272.960 / max: 288.704) us
cupy.sum (shape = ( 32768, 512), axis=1): 15.739 us +/- 0.819 (min: 14.927 / max: 24.207) us 450.278 us +/- 0.899 (min: 439.168 / max: 458.464) us
cupy.sum (shape = ( 65536, 256), axis=1): 15.554 us +/- 0.536 (min: 14.976 / max: 21.361) us 425.004 us +/- 0.738 (min: 413.824 / max: 430.400) us
cupy.sum (shape = ( 131072, 128), axis=1): 15.450 us +/- 0.741 (min: 14.815 / max: 23.849) us 404.190 us +/- 0.833 (min: 392.960 / max: 411.936) us
cupy.sum (shape = ( 262144, 64), axis=1): 15.367 us +/- 0.971 (min: 14.726 / max: 27.690) us 390.224 us +/- 1.087 (min: 379.296 / max: 400.640) us
cupy.sum (shape = ( 524288, 32), axis=1): 15.601 us +/- 0.731 (min: 14.994 / max: 24.911) us 387.660 us +/- 0.767 (min: 376.832 / max: 396.288) us
cupy.sum (shape = ( 1048576, 16), axis=1): 15.540 us +/- 0.514 (min: 14.901 / max: 20.031) us 368.645 us +/- 0.689 (min: 358.048 / max: 372.768) us
cupy.sum (shape = ( 2097152, 8), axis=1): 15.433 us +/- 0.406 (min: 14.849 / max: 18.176) us 328.896 us +/- 0.617 (min: 318.400 / max: 331.328) us
cupy.sum (shape = ( 4194304, 4), axis=1): 15.590 us +/- 0.781 (min: 14.976 / max: 26.174) us 302.788 us +/- 0.883 (min: 291.744 / max: 312.896) us
cupy.sum (shape = ( 8388608, 2), axis=1): 15.390 us +/- 0.531 (min: 14.741 / max: 20.199) us 288.805 us +/- 0.846 (min: 277.888 / max: 298.144) us
cupy.sum (shape = (16777216, 1), axis=1): 15.459 us +/- 0.744 (min: 14.897 / max: 28.141) us 187.830 us +/- 0.952 (min: 177.056 / max: 199.776) us
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment