Skip to content

Instantly share code, notes, and snippets.

@bshambaugh
Last active September 20, 2017 19:40
Show Gist options
  • Save bshambaugh/20c5a53c82c8e63ed623334f4dac73e9 to your computer and use it in GitHub Desktop.
Save bshambaugh/20c5a53c82c8e63ed623334f4dac73e9 to your computer and use it in GitHub Desktop.
GPU: GTX 1070; CPU: i7-7700HQ; 16GB DDR4 RAM - September 16, 2017
TAP version 13
ok 1 sgemm: 128x128 . 128x128
# 7.567 GFlops/sec ±5.50% n = 61 µ = 1ms
ok 2 sgemm: 128x256 . 256x128
# 4.195 GFlops/sec ±129.98% n = 38 µ = 2ms
ok 3 sgemm: 256x256 . 256x256
# 25.376 GFlops/sec ±5.82% n = 55 µ = 1ms
ok 4 sgemm: 512x256 . 256x512
# 42.685 GFlops/sec ±6.32% n = 53 µ = 3ms
ok 5 sgemm: 256x512 . 512x256
# 32.952 GFlops/sec ±5.77% n = 57 µ = 2ms
ok 6 sgemm: 512x512 . 512x512
# 57.069 GFlops/sec ±5.69% n = 53 µ = 5ms
ok 7 sgemm: 513x513 . 513x513
# 24.541 GFlops/sec ±4.47% n = 54 µ = 11ms
ok 8 sgemm: 1024x512 . 512x1024
# 54.417 GFlops/sec ±94.06% n = 29 µ = 20ms
ok 9 sgemm: 512x1024 . 1024x512
# 73.082 GFlops/sec ±5.28% n = 52 µ = 7ms
ok 10 sgemm: 1024x1024 . 1024x1024
# 128.455 GFlops/sec ±5.67% n = 47 µ = 17ms
ok 11 sgemm: 2048x2048 . 2048x2048
# 146.784 GFlops/sec ±3.11% n = 24 µ = 117ms
1..11
# tests 11
# pass 11
# ok
@waylonflinn
Copy link

Awesome! That 1070 is fast! 🥇

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment