Skip to content

Instantly share code, notes, and snippets.

@kadirmalak
Created March 19, 2019 11:59
Show Gist options
  • Save kadirmalak/954ff179d7a96d266eaea49f0855ae5e to your computer and use it in GitHub Desktop.
Save kadirmalak/954ff179d7a96d266eaea49f0855ae5e to your computer and use it in GitHub Desktop.
export OMP_NUM_THREADS=6; export OMP_PROC_BIND="CLOSE"; export GOMP_CPU_AFFINITY="0 1 2 3 4 5"
> export OMP_NUM_THREADS=6
> export OMP_PROC_BIND="CLOSE"
> export GOMP_CPU_AFFINITY="0 1 2 3 4 5"
> ./runtests --gtest_filter="PlaygroundTests.FastScalar"
Note: Google Test filter = PlaygroundTests.FastScalar
[==========] Running 1 test from 1 test case.
[----------] Global test environment set-up.
[----------] 1 test from PlaygroundTests
[ RUN ] PlaygroundTests.FastScalar
Scalar Addition - x.add(3.14159265359)
TestName OpNum Warmup NumIter DataType Inplace Shape Strides Axis Orders avg (us) median (us) min (us) max (us) stdev (us)
sAdd 0 10 1000 FLOAT false [16] [1]/[1] N/A c/c 0 0 0 6 0.19
sAdd 0 10 1000 FLOAT true [16] [1]/[1] N/A c/c 0 0 0 5 0.16
sAdd 0 10 1000 FLOAT false [64] [1]/[1] N/A c/c 0 0 0 4 0.13
sAdd 0 10 1000 FLOAT true [64] [1]/[1] N/A c/c 0 0 0 5 0.16
sAdd 0 10 1000 FLOAT false [256] [1]/[1] N/A c/c 0 0 0 4 0.13
sAdd 0 10 1000 FLOAT true [256] [1]/[1] N/A c/c 0 0 0 4 0.13
sAdd 0 10 1000 FLOAT false [1024] [1]/[1] N/A c/c 0 0 0 1 0.03
sAdd 0 10 1000 FLOAT true [1024] [1]/[1] N/A c/c 0 0 0 4 0.13
sAdd 0 10 1000 FLOAT false [4096] [1]/[1] N/A c/c 4 2 2 2291 72.35
sAdd 0 10 1000 FLOAT true [4096] [1]/[1] N/A c/c 2 2 2 8 0.45
sAdd 0 10 1000 FLOAT false [16384] [1]/[1] N/A c/c 3 3 3 228 7.11
sAdd 0 10 1000 FLOAT true [16384] [1]/[1] N/A c/c 3 3 3 7 0.18
sAdd 0 10 1000 FLOAT false [65536] [1]/[1] N/A c/c 4 5 3 17 0.93
sAdd 0 10 1000 FLOAT true [65536] [1]/[1] N/A c/c 3 3 3 7 0.42
sAdd 0 10 1000 FLOAT false [262144] [1]/[1] N/A c/c 7 7 7 15 0.41
sAdd 0 10 1000 FLOAT true [262144] [1]/[1] N/A c/c 7 7 6 14 0.34
sAdd 0 10 1000 FLOAT false [1048576] [1]/[1] N/A c/c 29 30 29 41 0.89
sAdd 0 10 1000 FLOAT true [1048576] [1]/[1] N/A c/c 22 22 22 32 0.79
sAdd 0 10 1000 FLOAT false [4194304] [1]/[1] N/A c/c 3126 3126 2947 3403 15.99
sAdd 0 10 1000 FLOAT true [4194304] [1]/[1] N/A c/c 234 231 220 275 9.31
sAdd 0 10 1000 FLOAT false [16777216] [1]/[1] N/A c/c 12380 12379 12357 12682 14.79
sAdd 0 10 1000 FLOAT true [16777216] [1]/[1] N/A c/c 8502 8503 8444 8582 18.84
sAdd 0 10 1000 FLOAT false [67108864] [1]/[1] N/A c/c 49702 49701 49650 49817 20.19
sAdd 0 10 1000 FLOAT true [67108864] [1]/[1] N/A c/c 33957 33953 33871 34133 35.72
sAdd 0 10 1000 FLOAT false [268435456] [1]/[1] N/A c/c 198437 198436 198289 198681 47.15
sAdd 0 10 1000 FLOAT true [268435456] [1]/[1] N/A c/c 135656 135645 135524 135965 59.14
sAdd 0 10 1000 FLOAT false [1073741824] [1]/[1] N/A c/c 794002 794000 793571 795473 172.00
sAdd 0 10 1000 FLOAT true [1073741824] [1]/[1] N/A c/c 542388 542367 542051 543108 159.63
Scalar Subtraction - x.sub(3.14159265359)
TestName OpNum Warmup NumIter DataType Inplace Shape Strides Axis Orders avg (us) median (us) min (us) max (us) stdev (us)
sSub 1 10 1000 FLOAT false [16] [1]/[1] N/A c/c 0 0 0 9 0.29
sSub 1 10 1000 FLOAT true [16] [1]/[1] N/A c/c 0 0 0 8 0.25
sSub 1 10 1000 FLOAT false [64] [1]/[1] N/A c/c 0 0 0 7 0.22
sSub 1 10 1000 FLOAT true [64] [1]/[1] N/A c/c 0 0 0 5 0.16
sSub 1 10 1000 FLOAT false [256] [1]/[1] N/A c/c 0 0 0 6 0.19
sSub 1 10 1000 FLOAT true [256] [1]/[1] N/A c/c 0 0 0 8 0.25
sSub 1 10 1000 FLOAT false [1024] [1]/[1] N/A c/c 0 0 0 0 0.00
sSub 1 10 1000 FLOAT true [1024] [1]/[1] N/A c/c 0 0 0 4 0.13
sSub 1 10 1000 FLOAT false [4096] [1]/[1] N/A c/c 2 2 2 16 0.46
sSub 1 10 1000 FLOAT true [4096] [1]/[1] N/A c/c 1 2 1 6 0.30
sSub 1 10 1000 FLOAT false [16384] [1]/[1] N/A c/c 2 3 2 14 0.61
sSub 1 10 1000 FLOAT true [16384] [1]/[1] N/A c/c 3 3 3 9 0.30
sSub 1 10 1000 FLOAT false [65536] [1]/[1] N/A c/c 5 6 4 17 1.33
sSub 1 10 1000 FLOAT true [65536] [1]/[1] N/A c/c 4 4 4 8 0.21
sSub 1 10 1000 FLOAT false [262144] [1]/[1] N/A c/c 10 10 10 18 0.46
sSub 1 10 1000 FLOAT true [262144] [1]/[1] N/A c/c 10 10 10 17 0.37
sSub 1 10 1000 FLOAT false [1048576] [1]/[1] N/A c/c 35 35 34 42 0.92
sSub 1 10 1000 FLOAT true [1048576] [1]/[1] N/A c/c 34 34 33 42 0.80
sSub 1 10 1000 FLOAT false [4194304] [1]/[1] N/A c/c 3099 3098 3081 3353 11.27
sSub 1 10 1000 FLOAT true [4194304] [1]/[1] N/A c/c 165 160 155 268 17.04
sSub 1 10 1000 FLOAT false [16777216] [1]/[1] N/A c/c 12389 12388 12361 12705 14.34
sSub 1 10 1000 FLOAT true [16777216] [1]/[1] N/A c/c 8498 8498 8438 8577 19.00
sSub 1 10 1000 FLOAT false [67108864] [1]/[1] N/A c/c 49647 49646 49597 49757 20.47
sSub 1 10 1000 FLOAT true [67108864] [1]/[1] N/A c/c 33900 33896 33805 34102 33.93
sSub 1 10 1000 FLOAT false [268435456] [1]/[1] N/A c/c 198424 198421 198291 198693 52.77
sSub 1 10 1000 FLOAT true [268435456] [1]/[1] N/A c/c 135547 135540 135364 136058 74.17
sSub 1 10 1000 FLOAT false [1073741824] [1]/[1] N/A c/c 793128 793118 792668 793852 159.70
sSub 1 10 1000 FLOAT true [1073741824] [1]/[1] N/A c/c 541863 541848 541411 542707 177.09
Scalar Multiplication - x.mul(3.14159265359)
TestName OpNum Warmup NumIter DataType Inplace Shape Strides Axis Orders avg (us) median (us) min (us) max (us) stdev (us)
sMul 2 10 1000 FLOAT false [16] [1]/[1] N/A c/c 0 0 0 8 0.26
sMul 2 10 1000 FLOAT true [16] [1]/[1] N/A c/c 0 0 0 8 0.25
sMul 2 10 1000 FLOAT false [64] [1]/[1] N/A c/c 0 0 0 5 0.16
sMul 2 10 1000 FLOAT true [64] [1]/[1] N/A c/c 0 0 0 6 0.19
sMul 2 10 1000 FLOAT false [256] [1]/[1] N/A c/c 0 0 0 5 0.16
sMul 2 10 1000 FLOAT true [256] [1]/[1] N/A c/c 0 0 0 1 0.03
sMul 2 10 1000 FLOAT false [1024] [1]/[1] N/A c/c 0 0 0 5 0.16
sMul 2 10 1000 FLOAT true [1024] [1]/[1] N/A c/c 0 0 0 4 0.13
sMul 2 10 1000 FLOAT false [4096] [1]/[1] N/A c/c 2 2 2 11 0.35
sMul 2 10 1000 FLOAT true [4096] [1]/[1] N/A c/c 2 2 2 6 0.16
sMul 2 10 1000 FLOAT false [16384] [1]/[1] N/A c/c 2 2 2 13 0.73
sMul 2 10 1000 FLOAT true [16384] [1]/[1] N/A c/c 3 3 3 8 0.23
sMul 2 10 1000 FLOAT false [65536] [1]/[1] N/A c/c 5 6 4 17 1.02
sMul 2 10 1000 FLOAT true [65536] [1]/[1] N/A c/c 3 4 3 11 0.56
sMul 2 10 1000 FLOAT false [262144] [1]/[1] N/A c/c 9 9 8 16 0.41
sMul 2 10 1000 FLOAT true [262144] [1]/[1] N/A c/c 7 7 7 15 0.58
sMul 2 10 1000 FLOAT false [1048576] [1]/[1] N/A c/c 34 34 33 44 1.31
sMul 2 10 1000 FLOAT true [1048576] [1]/[1] N/A c/c 22 22 22 29 0.63
sMul 2 10 1000 FLOAT false [4194304] [1]/[1] N/A c/c 3102 3101 3082 3189 7.70
sMul 2 10 1000 FLOAT true [4194304] [1]/[1] N/A c/c 129 128 122 164 5.60
sMul 2 10 1000 FLOAT false [16777216] [1]/[1] N/A c/c 12369 12368 12337 12745 18.62
sMul 2 10 1000 FLOAT true [16777216] [1]/[1] N/A c/c 8499 8499 8453 8725 18.73
sMul 2 10 1000 FLOAT false [67108864] [1]/[1] N/A c/c 49635 49635 49575 49749 21.65
sMul 2 10 1000 FLOAT true [67108864] [1]/[1] N/A c/c 33940 33936 33851 34089 32.57
sMul 2 10 1000 FLOAT false [268435456] [1]/[1] N/A c/c 198404 198403 198267 198591 45.87
sMul 2 10 1000 FLOAT true [268435456] [1]/[1] N/A c/c 135654 135647 135529 135884 53.57
sMul 2 10 1000 FLOAT false [1073741824] [1]/[1] N/A c/c 794082 794072 793560 795066 149.69
sMul 2 10 1000 FLOAT true [1073741824] [1]/[1] N/A c/c 542314 542296 541939 542950 165.17
Scalar Division - x.div(3.14159265359)
TestName OpNum Warmup NumIter DataType Inplace Shape Strides Axis Orders avg (us) median (us) min (us) max (us) stdev (us)
sDiv 3 10 1000 FLOAT false [16] [1]/[1] N/A c/c 0 0 0 1 0.03
sDiv 3 10 1000 FLOAT true [16] [1]/[1] N/A c/c 0 0 0 12 0.38
sDiv 3 10 1000 FLOAT false [64] [1]/[1] N/A c/c 0 0 0 5 0.16
sDiv 3 10 1000 FLOAT true [64] [1]/[1] N/A c/c 0 0 0 0 0.00
sDiv 3 10 1000 FLOAT false [256] [1]/[1] N/A c/c 0 0 0 5 0.16
sDiv 3 10 1000 FLOAT true [256] [1]/[1] N/A c/c 0 0 0 5 0.16
sDiv 3 10 1000 FLOAT false [1024] [1]/[1] N/A c/c 0 0 0 0 0.00
sDiv 3 10 1000 FLOAT true [1024] [1]/[1] N/A c/c 0 0 0 4 0.13
sDiv 3 10 1000 FLOAT false [4096] [1]/[1] N/A c/c 1 2 1 13 0.41
sDiv 3 10 1000 FLOAT true [4096] [1]/[1] N/A c/c 2 2 2 7 0.21
sDiv 3 10 1000 FLOAT false [16384] [1]/[1] N/A c/c 3 3 3 11 0.45
sDiv 3 10 1000 FLOAT true [16384] [1]/[1] N/A c/c 3 3 3 8 0.19
sDiv 3 10 1000 FLOAT false [65536] [1]/[1] N/A c/c 5 6 4 16 1.04
sDiv 3 10 1000 FLOAT true [65536] [1]/[1] N/A c/c 3 3 3 7 0.50
sDiv 3 10 1000 FLOAT false [262144] [1]/[1] N/A c/c 8 9 8 17 0.47
sDiv 3 10 1000 FLOAT true [262144] [1]/[1] N/A c/c 7 7 7 14 0.34
sDiv 3 10 1000 FLOAT false [1048576] [1]/[1] N/A c/c 33 33 33 41 1.10
sDiv 3 10 1000 FLOAT true [1048576] [1]/[1] N/A c/c 22 22 22 29 0.61
sDiv 3 10 1000 FLOAT false [4194304] [1]/[1] N/A c/c 3109 3108 3093 3146 7.11
sDiv 3 10 1000 FLOAT true [4194304] [1]/[1] N/A c/c 137 133 128 221 12.75
sDiv 3 10 1000 FLOAT false [16777216] [1]/[1] N/A c/c 12389 12388 12365 12584 11.53
sDiv 3 10 1000 FLOAT true [16777216] [1]/[1] N/A c/c 8503 8503 8457 8699 18.62
sDiv 3 10 1000 FLOAT false [67108864] [1]/[1] N/A c/c 49657 49656 49601 49752 20.30
sDiv 3 10 1000 FLOAT true [67108864] [1]/[1] N/A c/c 33934 33931 33866 34125 32.36
sDiv 3 10 1000 FLOAT false [268435456] [1]/[1] N/A c/c 198555 198553 198430 198711 41.72
sDiv 3 10 1000 FLOAT true [268435456] [1]/[1] N/A c/c 135606 135597 135480 135881 61.33
sDiv 3 10 1000 FLOAT false [1073741824] [1]/[1] N/A c/c 793924 793922 793544 794391 138.94
sDiv 3 10 1000 FLOAT true [1073741824] [1]/[1] N/A c/c 542327 542318 541987 542847 150.03
Scalar Power - x.pow(3.14159265359)
TestName OpNum Warmup NumIter DataType Inplace Shape Strides Axis Orders avg (us) median (us) min (us) max (us) stdev (us)
sPow 31 10 1000 FLOAT false [16] [1]/[1] N/A c/c 0 0 0 1 0.03
sPow 31 10 1000 FLOAT true [16] [1]/[1] N/A c/c 0 0 0 9 0.28
sPow 31 10 1000 FLOAT false [64] [1]/[1] N/A c/c 1 1 1 5 0.13
sPow 31 10 1000 FLOAT true [64] [1]/[1] N/A c/c 1 1 1 5 0.13
sPow 31 10 1000 FLOAT false [256] [1]/[1] N/A c/c 2 2 2 6 0.13
sPow 31 10 1000 FLOAT true [256] [1]/[1] N/A c/c 2 2 2 6 0.18
sPow 31 10 1000 FLOAT false [1024] [1]/[1] N/A c/c 6 6 6 11 0.23
sPow 31 10 1000 FLOAT true [1024] [1]/[1] N/A c/c 6 6 6 10 0.21
sPow 31 10 1000 FLOAT false [4096] [1]/[1] N/A c/c 7 7 7 20 0.57
sPow 31 10 1000 FLOAT true [4096] [1]/[1] N/A c/c 7 7 7 11 0.39
sPow 31 10 1000 FLOAT false [16384] [1]/[1] N/A c/c 17 17 17 39 3.10
sPow 31 10 1000 FLOAT true [16384] [1]/[1] N/A c/c 17 17 17 25 0.56
sPow 31 10 1000 FLOAT false [65536] [1]/[1] N/A c/c 61 61 61 69 1.01
sPow 31 10 1000 FLOAT true [65536] [1]/[1] N/A c/c 61 61 61 68 0.96
sPow 31 10 1000 FLOAT false [262144] [1]/[1] N/A c/c 239 239 236 247 1.53
sPow 31 10 1000 FLOAT true [262144] [1]/[1] N/A c/c 237 237 236 246 1.52
sPow 31 10 1000 FLOAT false [1048576] [1]/[1] N/A c/c 946 946 943 957 1.65
sPow 31 10 1000 FLOAT true [1048576] [1]/[1] N/A c/c 944 944 939 1044 3.57
sPow 31 10 1000 FLOAT false [4194304] [1]/[1] N/A c/c 3800 3797 3781 4519 25.39
sPow 31 10 1000 FLOAT true [4194304] [1]/[1] N/A c/c 3766 3764 3758 4417 22.43
sPow 31 10 1000 FLOAT false [16777216] [1]/[1] N/A c/c 15194 15192 15149 16031 35.73
sPow 31 10 1000 FLOAT true [16777216] [1]/[1] N/A c/c 15192 15192 15159 15896 31.31
sPow 31 10 1000 FLOAT false [67108864] [1]/[1] N/A c/c 60404 60402 60379 60796 25.41
sPow 31 10 1000 FLOAT true [67108864] [1]/[1] N/A c/c 60961 60735 60683 65050 946.87
sPow 31 10 1000 FLOAT false [268435456] [1]/[1] N/A c/c 241609 241594 241508 243712 103.62
sPow 31 10 1000 FLOAT true [268435456] [1]/[1] N/A c/c 242968 242962 242861 243705 51.78
sPow 31 10 1000 FLOAT false [1073741824] [1]/[1] N/A c/c 966731 966243 965923 972188 1267.25
sPow 31 10 1000 FLOAT true [1073741824] [1]/[1] N/A c/c 971808 971836 971330 972683 214.29
[ OK ] PlaygroundTests.FastScalar (9813258 ms)
[----------] 1 test from PlaygroundTests (9813258 ms total)
[----------] Global test environment tear-down
[==========] 1 test from 1 test case ran. (9813258 ms total)
[ PASSED ] 1 test.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment