[libc-commits] [libc] [libc] Polish GPU benchmarking (PR #153900)
Leandro Lacerda via libc-commits
libc-commits at lists.llvm.org
Fri Aug 15 16:20:11 PDT 2025
leandrolcampos wrote:
Here's what I get on my *NVIDIA GeForce RTX 4070 Laptop GPU*
```bash
[1/4] Running hermetic test libc.benchmarks.gpu.src.ctype.isalnum_benchmark
Running Suite: LlvmLibcIsAlNumGpuBenchmark
Benchmark | Cycles (Mean) | Stddev | Min | Max | Iterations | Threads |
------------------------------------------------------------------------------------------------------
IsAlnum | 53 | 0 | 53 | 53 | 11904 | 64 |
IsAlnumSingleThread | 53 | 0 | 53 | 53 | 186 | 1 |
IsAlnumSingleWave | 53 | 0 | 53 | 53 | 5952 | 32 |
IsAlnumCapital | 53 | 0 | 53 | 53 | 11904 | 64 |
IsAlnumNotAlnum | 43 | 0 | 43 | 43 | 11904 | 64 |
[2/4] Running hermetic test libc.benchmarks.gpu.src.ctype.isalpha_benchmark
Running Suite: LlvmLibcIsAlphaGpuBenchmark
Benchmark | Cycles (Mean) | Stddev | Min | Max | Iterations | Threads |
------------------------------------------------------------------------------------------------------
IsAlpha | 53 | 0 | 53 | 53 | 186 | 1 |
[3/4] Running hermetic test libc.benchmarks.gpu.src.math.sin_benchmark
Running Suite: LlvmLibcSinGpuBenchmark
Benchmark | Cycles (Mean) | Stddev | Min | Max | Iterations | Threads |
------------------------------------------------------------------------------------------------------
Sin_1 | 3120 | 153 | 2934 | 3610 | 2735008 | 32 |
Sin_128 | 364 | 6 | 346 | 382 | 12032 | 32 |
Sin_1024 | 354 | 2 | 348 | 359 | 5952 | 32 |
Sin_4096 | 361 | 1 | 358 | 363 | 4160 | 32 |
SinTwoPi_1 | 2217 | 11 | 2204 | 2494 | 24032 | 32 |
SinTwoPi_128 | 262 | 2 | 258 | 268 | 24032 | 32 |
SinTwoPi_1024 | 271 | 1 | 271 | 276 | 24032 | 32 |
SinTwoPi_4096 | 282 | 0 | 281 | 282 | 1984 | 32 |
SinTwoPow30_1 | 3133 | 11 | 3116 | 3177 | 1984 | 32 |
SinTwoPow30_128 | 348 | 2 | 345 | 352 | 1344 | 32 |
SinTwoPow30_1024 | 359 | 1 | 358 | 361 | 1984 | 32 |
SinTwoPow30_4096 | 368 | 0 | 367 | 368 | 576 | 32 |
SinVeryLarge_1 | 2855 | 15 | 2821 | 3082 | 8480 | 32 |
SinVeryLarge_128 | 316 | 2 | 310 | 324 | 17024 | 32 |
SinVeryLarge_1024 | 317 | 1 | 316 | 320 | 1344 | 32 |
SinVeryLarge_4096 | 325 | 0 | 325 | 326 | 896 | 32 |
NvSin_1 | 2521 | 69 | 2264 | 2876 | 5952 | 32 |
NvSin_128 | 1860 | 2 | 1857 | 1865 | 576 | 32 |
NvSin_1024 | 2069 | 1 | 2068 | 2071 | 352 | 32 |
NvSin_4096 | 2087 | 0 | 2087 | 2088 | 352 | 32 |
NvSinTwoPi_1 | 1106 | 1 | 1104 | 1108 | 1344 | 32 |
NvSinTwoPi_128 | 925 | 0 | 925 | 925 | 352 | 32 |
NvSinTwoPi_1024 | 1136 | 0 | 1136 | 1136 | 352 | 32 |
NvSinTwoPi_4096 | 1155 | 0 | 1155 | 1155 | 352 | 32 |
NvSinTwoPow30_1 | 1106 | 1 | 1104 | 1107 | 1984 | 32 |
NvSinTwoPow30_128 | 925 | 0 | 925 | 927 | 576 | 32 |
NvSinTwoPow30_1024 | 1136 | 0 | 1136 | 1136 | 352 | 32 |
NvSinTwoPow30_4096 | 1155 | 0 | 1155 | 1155 | 352 | 32 |
NvSinVeryLarge_1 | 2496 | 23 | 2251 | 2840 | 12032 | 32 |
NvSinVeryLarge_128 | 1827 | 1 | 1826 | 1829 | 576 | 32 |
NvSinVeryLarge_1024 | 2034 | 0 | 2034 | 2035 | 352 | 32 |
NvSinVeryLarge_4096 | 2052 | 0 | 2052 | 2052 | 352 | 32 |
Sinf_1 | 2200 | 170 | 1521 | 2397 | 507776 | 32 |
Sinf_128 | 240 | 5 | 218 | 255 | 47616 | 32 |
Sinf_1024 | 240 | 2 | 235 | 248 | 5952 | 32 |
Sinf_4096 | 259 | 1 | 257 | 261 | 896 | 32 |
SinfTwoPi_1 | 1441 | 11 | 1425 | 1760 | 33856 | 32 |
SinfTwoPi_128 | 147 | 1 | 146 | 150 | 12032 | 32 |
SinfTwoPi_1024 | 147 | 0 | 145 | 148 | 4160 | 32 |
SinfTwoPi_4096 | 165 | 0 | 165 | 167 | 4160 | 32 |
SinfTwoPow30_1 | 1087 | 10 | 1078 | 1160 | 1984 | 32 |
SinfTwoPow30_128 | 102 | 1 | 101 | 104 | 576 | 32 |
SinfTwoPow30_1024 | 102 | 0 | 100 | 103 | 2880 | 32 |
SinfTwoPow30_4096 | 121 | 0 | 121 | 122 | 4160 | 32 |
SinfVeryLarge_1 | 1924 | 14 | 1867 | 2277 | 24032 | 32 |
SinfVeryLarge_128 | 205 | 1 | 205 | 207 | 896 | 32 |
SinfVeryLarge_1024 | 206 | 0 | 204 | 206 | 5952 | 32 |
SinfVeryLarge_4096 | 225 | 0 | 224 | 226 | 8480 | 32 |
NvSinf_1 | 1023 | 6 | 1018 | 1036 | 2880 | 32 |
NvSinf_128 | 786 | 0 | 786 | 788 | 576 | 32 |
NvSinf_1024 | 975 | 1 | 970 | 975 | 2880 | 32 |
NvSinf_4096 | 1010 | 0 | 1010 | 1010 | 352 | 32 |
NvSinfTwoPi_1 | 162 | 3 | 162 | 509 | 362464 | 32 |
NvSinfTwoPi_128 | 141 | 0 | 141 | 143 | 2880 | 32 |
NvSinfTwoPi_1024 | 331 | 0 | 330 | 331 | 896 | 32 |
NvSinfTwoPi_4096 | 366 | 0 | 366 | 366 | 352 | 32 |
NvSinfTwoPow30_1 | 1025 | 10 | 1018 | 1280 | 47616 | 32 |
NvSinfTwoPow30_128 | 776 | 0 | 776 | 777 | 352 | 32 |
NvSinfTwoPow30_1024 | 966 | 0 | 966 | 967 | 352 | 32 |
NvSinfTwoPow30_4096 | 1000 | 0 | 1000 | 1000 | 352 | 32 |
NvSinfVeryLarge_1 | 1002 | 2 | 1000 | 1023 | 4160 | 32 |
NvSinfVeryLarge_128 | 757 | 0 | 757 | 758 | 576 | 32 |
NvSinfVeryLarge_1024 | 948 | 0 | 948 | 948 | 352 | 32 |
NvSinfVeryLarge_4096 | 981 | 0 | 981 | 981 | 352 | 32 |
[4/4] Running hermetic test libc.benchmarks.gpu.src.math.atan2_benchmark
Running Suite: LlvmLibcAtan2GpuBenchmark
Benchmark | Cycles (Mean) | Stddev | Min | Max | Iterations | Threads |
------------------------------------------------------------------------------------------------------
Atan2_1 | 4081 | 954 | 1892 | 5264 | 24032 | 32 |
Atan2_128 | 2505 | 37 | 2346 | 2601 | 33856 | 32 |
Atan2_1024 | 2747 | 13 | 2708 | 2783 | 2880 | 32 |
Atan2_4096 | 2749 | 6 | 2737 | 2759 | 576 | 32 |
Atan2TwoPi_1 | 2737 | 16 | 2728 | 3156 | 24032 | 32 |
Atan2TwoPi_128 | 1068 | 3 | 1064 | 1109 | 8480 | 32 |
Atan2TwoPi_1024 | 1301 | 1 | 1300 | 1303 | 576 | 32 |
Atan2TwoPi_4096 | 1302 | 0 | 1302 | 1303 | 576 | 32 |
Atan2TwoPow30_1 | 2733 | 16 | 2722 | 3157 | 24032 | 32 |
Atan2TwoPow30_128 | 1072 | 3 | 1065 | 1108 | 5952 | 32 |
Atan2TwoPow30_1024 | 1302 | 1 | 1301 | 1304 | 896 | 32 |
Atan2TwoPow30_4096 | 1303 | 0 | 1302 | 1303 | 352 | 32 |
Atan2Large_1 | 3569 | 382 | 1125 | 3878 | 131648 | 32 |
Atan2Large_128 | 1811 | 17 | 1766 | 1839 | 1984 | 32 |
Atan2Large_1024 | 2050 | 4 | 2043 | 2059 | 576 | 32 |
Atan2Large_4096 | 2049 | 2 | 2046 | 2053 | 352 | 32 |
NvAtan2_1 | 2909 | 38 | 2866 | 3341 | 17024 | 32 |
NvAtan2_128 | 2836 | 3 | 2830 | 2853 | 2880 | 32 |
NvAtan2_1024 | 3075 | 1 | 3074 | 3076 | 352 | 32 |
NvAtan2_4096 | 3075 | 1 | 3074 | 3076 | 352 | 32 |
NvAtan2TwoPi_1 | 2032 | 13 | 2032 | 2383 | 24032 | 32 |
NvAtan2TwoPi_128 | 1979 | 3 | 1979 | 1999 | 1984 | 32 |
NvAtan2TwoPi_1024 | 2218 | 0 | 2217 | 2218 | 352 | 32 |
NvAtan2TwoPi_4096 | 2218 | 0 | 2218 | 2218 | 352 | 32 |
NvAtan2TwoPow30_1 | 2032 | 8 | 2032 | 2184 | 12032 | 32 |
NvAtan2TwoPow30_128 | 1979 | 2 | 1979 | 1998 | 1984 | 32 |
NvAtan2TwoPow30_1024 | 2218 | 0 | 2217 | 2218 | 352 | 32 |
NvAtan2TwoPow30_4096 | 2218 | 0 | 2218 | 2219 | 576 | 32 |
NvAtan2Large_1 | 2032 | 12 | 2032 | 2354 | 24032 | 32 |
NvAtan2Large_128 | 1979 | 2 | 1979 | 1997 | 1984 | 32 |
NvAtan2Large_1024 | 2218 | 0 | 2217 | 2218 | 352 | 32 |
NvAtan2Large_4096 | 2219 | 0 | 2218 | 2219 | 352 | 32 |
```
https://github.com/llvm/llvm-project/pull/153900
More information about the libc-commits
mailing list