Created
October 31, 2024 18:10
-
-
Save gnattu/f2e77e40b2c38513401799a27b0979c1 to your computer and use it in GitHub Desktop.
A18 Pro SME Test
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| -- SME features -- | |
| FEAT_SME = true | |
| FEAT_SME2 = true | |
| SME_F32F32 = true | |
| SME_BI32I32 = true | |
| SME_B16F32 = true | |
| SME_F16F32 = true | |
| SME_I8I32 = true | |
| SME_I16I32 = true | |
| FEAT_SME_F64F64 = true | |
| FEAT_SME_I16I64 = true | |
| Vector length = 512 bits | |
| -- Peak outer product accumulating to ZA storage -- | |
| type | ZA tiles | GFLOPS | |
| ------ | ----------- | --------- | |
| f32 | 4 (full) | 2020.6770714631093 | |
| f32 | 3 | 1479.875713563119 | |
| f32 | 2 | 986.2385720929755 | |
| f32 | 1 | 493.1428344395738 | |
| ------ | ----------- | --------- | |
| f64 | 8 (full) | 492.6355246702803 | |
| f64 | 7 | 493.3370590559147 | |
| f64 | 6 | 493.58159159176046 | |
| f64 | 5 | 493.01545391605447 | |
| f64 | 4 | 493.364593119522 | |
| f64 | 3 | 369.8177980901915 | |
| f64 | 2 | 246.57052663515506 | |
| f64 | 1 | 123.00738605537532 | |
| ------ | ----------- | --------- | |
| f16f32 | 4 (full) | 3945.498144715227 | |
| f16f32 | 3 | 2959.117081146947 | |
| f16f32 | 2 | 1973.8621425669255 | |
| f16f32 | 1 | 986.0796892500823 | |
| ------ | ----------- | --------- | |
| i16i32 | 4 (full) | 3940.619728738648 | |
| i16i32 | 3 | 3946.997553201997 | |
| i16i32 | 2 | 3945.478344072721 | |
| i16i32 | 1 | 1971.38697373329 | |
| ------ | ----------- | --------- | |
| i8i32 | 4 (full) | 15785.353643828112 | |
| i8i32 | 3 | 15782.409583621618 | |
| i8i32 | 2 | 15794.852019666441 | |
| i8i32 | 1 | 7892.768720834235 | |
| -- Peak fused multiply-add rate accumulating to ZA storage (SME2) -- | |
| type | VGx | Z registers | ZA slices | GFLOPS | |
| ------ | ----- | ------------|------------| --------- | |
| f32 | 4 | 64, 64 | 64 (full) | 246.52987592163385 | |
| f32 | 4 | 60, 60 | 60 | 246.49289238707664 | |
| f32 | 4 | 56, 56 | 56 | 246.64735260988377 | |
| f32 | 4 | 52, 52 | 52 | 246.71805567525936 | |
| f32 | 4 | 48, 48 | 48 | 246.78427001818642 | |
| f32 | 4 | 44, 44 | 44 | 246.63221693380234 | |
| f32 | 4 | 40, 40 | 40 | 246.85899650371482 | |
| f32 | 4 | 36, 36 | 36 | 246.7620730095331 | |
| f32 | 4 | 32, 32 | 32 | 246.94471747966446 | |
| f32 | 4 | 28, 28 | 28 | 246.76533608006656 | |
| f32 | 4 | 24, 24 | 24 | 246.52257676709027 | |
| f32 | 4 | 20, 20 | 20 | 246.7862358832971 | |
| f32 | 4 | 16, 16 | 16 | 246.89155336739344 | |
| f32 | 4 | 12, 12 | 12 | 246.91219104830782 | |
| f32 | 4 | 8, 8 | 8 | 246.89958974643366 | |
| f32 | 4 | 4, 4 | 4 | 122.99246391542839 | |
| ------ | ----- | ------------|------------| --------- | |
| f32 | 2 | 64, 64 | 64 (full) | 252.46959395759 | |
| f32 | 2 | 62, 62 | 62 | 253.7441866778383 | |
| f32 | 2 | 60, 60 | 60 | 253.63436646748232 | |
| f32 | 2 | 58, 58 | 58 | 253.63237085460574 | |
| f32 | 2 | 56, 56 | 56 | 253.7001329099886 | |
| f32 | 2 | 54, 54 | 54 | 253.54355435442952 | |
| f32 | 2 | 52, 52 | 52 | 253.65238203511456 | |
| f32 | 2 | 50, 50 | 50 | 253.79418982743275 | |
| f32 | 2 | 48, 48 | 48 | 253.6772846966111 | |
| f32 | 2 | 46, 46 | 46 | 253.76630485142735 | |
| f32 | 2 | 44, 44 | 44 | 252.72672935175055 | |
| f32 | 2 | 42, 42 | 42 | 253.9464199506656 | |
| f32 | 2 | 40, 40 | 40 | 253.0740629628339 | |
| f32 | 2 | 38, 38 | 38 | 253.6457667908648 | |
| f32 | 2 | 36, 36 | 36 | 253.64204491192572 | |
| f32 | 2 | 34, 34 | 34 | 253.7408603410841 | |
| f32 | 2 | 32, 32 | 32 | 253.49721690752764 | |
| f32 | 2 | 30, 30 | 30 | 253.76970267310696 | |
| f32 | 2 | 28, 28 | 28 | 253.61816313648674 | |
| f32 | 2 | 26, 26 | 26 | 253.69746754461562 | |
| f32 | 2 | 24, 24 | 24 | 253.82468689782942 | |
| f32 | 2 | 22, 22 | 22 | 253.74581445558186 | |
| f32 | 2 | 20, 20 | 20 | 253.56149906350626 | |
| f32 | 2 | 18, 18 | 18 | 253.73406123350884 | |
| f32 | 2 | 16, 16 | 16 | 253.816228622985 | |
| f32 | 2 | 14, 14 | 14 | 253.7380456672071 | |
| f32 | 2 | 12, 12 | 12 | 253.77223170314105 | |
| f32 | 2 | 10, 10 | 10 | 253.27209140292183 | |
| f32 | 2 | 8, 8 | 8 | 253.74096700741995 | |
| f32 | 2 | 6, 6 | 6 | 190.17181558012908 | |
| f32 | 2 | 4, 4 | 4 | 126.54374596042456 | |
| f32 | 2 | 2, 2 | 2 | 61.777726077543264 | |
| ------ | ----- | ------------|------------| --------- | |
| f64 | 4 | 64, 64 | 64 (full) | 123.35463780762429 | |
| f64 | 4 | 60, 60 | 60 | 123.28482127412944 | |
| f64 | 4 | 56, 56 | 56 | 123.34000187939328 | |
| f64 | 4 | 52, 52 | 52 | 123.33304585764122 | |
| f64 | 4 | 48, 48 | 48 | 123.3344216311309 | |
| f64 | 4 | 44, 44 | 44 | 123.42134847465502 | |
| f64 | 4 | 40, 40 | 40 | 123.37502989097453 | |
| f64 | 4 | 36, 36 | 36 | 122.98645707724606 | |
| f64 | 4 | 32, 32 | 32 | 123.44028290256198 | |
| f64 | 4 | 28, 28 | 28 | 123.39106452368851 | |
| f64 | 4 | 24, 24 | 24 | 123.43807201612424 | |
| f64 | 4 | 20, 20 | 20 | 123.45643289998198 | |
| f64 | 4 | 16, 16 | 16 | 123.26099975022795 | |
| f64 | 4 | 12, 12 | 12 | 123.43648485344647 | |
| f64 | 4 | 8, 8 | 8 | 123.37101585352252 | |
| f64 | 4 | 4, 4 | 4 | 61.46397075190854 | |
| ------ | ----- | ------------|------------| --------- | |
| f64 | 2 | 64, 64 | 64 (full) | 125.11255853960411 | |
| f64 | 2 | 62, 62 | 62 | 125.3112913433041 | |
| f64 | 2 | 60, 60 | 60 | 126.7996720379316 | |
| f64 | 2 | 58, 58 | 58 | 126.9368817298433 | |
| f64 | 2 | 56, 56 | 56 | 126.84295074399571 | |
| f64 | 2 | 54, 54 | 54 | 126.85462023723944 | |
| f64 | 2 | 52, 52 | 52 | 126.82056078524644 | |
| f64 | 2 | 50, 50 | 50 | 126.82266281162126 | |
| f64 | 2 | 48, 48 | 48 | 126.839550239837 | |
| f64 | 2 | 46, 46 | 46 | 126.30930427131061 | |
| f64 | 2 | 44, 44 | 44 | 126.82445261917715 | |
| f64 | 2 | 42, 42 | 42 | 126.85754702787662 | |
| f64 | 2 | 40, 40 | 40 | 126.82951174347741 | |
| f64 | 2 | 38, 38 | 38 | 126.89681042320717 | |
| f64 | 2 | 36, 36 | 36 | 126.83689412462681 | |
| f64 | 2 | 34, 34 | 34 | 126.80221641789738 | |
| f64 | 2 | 32, 32 | 32 | 126.85564317142824 | |
| f64 | 2 | 30, 30 | 30 | 126.76300434922547 | |
| f64 | 2 | 28, 28 | 28 | 126.90127678065556 | |
| f64 | 2 | 26, 26 | 26 | 126.86997885754498 | |
| f64 | 2 | 24, 24 | 24 | 126.68495735723302 | |
| f64 | 2 | 22, 22 | 22 | 126.8803002668444 | |
| f64 | 2 | 20, 20 | 20 | 126.9006384498021 | |
| f64 | 2 | 18, 18 | 18 | 126.80581776462064 | |
| f64 | 2 | 16, 16 | 16 | 126.60867636499349 | |
| f64 | 2 | 14, 14 | 14 | 126.80764582171636 | |
| f64 | 2 | 12, 12 | 12 | 126.8124722127124 | |
| f64 | 2 | 10, 10 | 10 | 126.91970214040082 | |
| f64 | 2 | 8, 8 | 8 | 126.82836966728321 | |
| f64 | 2 | 6, 6 | 6 | 95.11122554400598 | |
| f64 | 2 | 4, 4 | 4 | 63.21968449189134 | |
| f64 | 2 | 2, 2 | 2 | 30.84616079766244 | |
| ------ | ----- | ------------|------------| --------- | |
| f16f32 | 4 | 32, 32 | 64 (full) | 493.2870033747888 | |
| f16f32 | 4 | 28, 28 | 56 | 493.46538242998497 | |
| f16f32 | 4 | 24, 24 | 48 | 493.1326099114331 | |
| f16f32 | 4 | 20, 20 | 40 | 493.62338849317604 | |
| f16f32 | 4 | 16, 16 | 32 | 493.7257700401928 | |
| f16f32 | 4 | 12, 12 | 24 | 493.7095679667141 | |
| f16f32 | 4 | 8, 8 | 16 | 493.6246163346055 | |
| f16f32 | 4 | 4, 4 | 8 | 246.7976367738042 | |
| ------ | ----- | ------------|------------| --------- | |
| f16f32 | 2 | 32, 32 | 64 (full) | 503.1475898573258 | |
| f16f32 | 2 | 30, 30 | 60 | 507.439975235872 | |
| f16f32 | 2 | 28, 28 | 56 | 507.36773260134964 | |
| f16f32 | 2 | 26, 26 | 52 | 507.4472225536503 | |
| f16f32 | 2 | 24, 24 | 48 | 507.4647452826799 | |
| f16f32 | 2 | 22, 22 | 44 | 507.6903239167751 | |
| f16f32 | 2 | 20, 20 | 40 | 506.5846105136095 | |
| f16f32 | 2 | 18, 18 | 36 | 507.6248379246869 | |
| f16f32 | 2 | 16, 16 | 32 | 507.6735627460256 | |
| f16f32 | 2 | 14, 14 | 28 | 444.26273305336247 | |
| f16f32 | 2 | 12, 12 | 24 | 374.1541339647184 | |
| f16f32 | 2 | 10, 10 | 20 | 308.0882188278218 | |
| f16f32 | 2 | 8, 8 | 16 | 246.87042316338892 | |
| f16f32 | 2 | 6, 6 | 12 | 246.97174215571278 | |
| f16f32 | 2 | 4, 4 | 8 | 246.8847079756539 | |
| f16f32 | 2 | 2, 2 | 4 | 123.45971705231977 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment