diff options
| author | Konstantin <const@const.me> | 2023-01-24 14:59:42 +0100 |
|---|---|---|
| committer | Konstantin <const@const.me> | 2023-01-24 14:59:42 +0100 |
| commit | b13a68d3a458e86cb1791d47c332985b026d2eef (patch) | |
| tree | 1ec2d9d759c000c874919501d0e3950c5fd73489 /ComputeShaders | |
| parent | a6c544b2a577ae0597e78a17ad36057a522bc1e4 (diff) | |
Comments
Diffstat (limited to 'ComputeShaders')
| -rw-r--r-- | ComputeShaders/mulMatTiled.hlsl | 5 |
1 files changed, 1 insertions, 4 deletions
diff --git a/ComputeShaders/mulMatTiled.hlsl b/ComputeShaders/mulMatTiled.hlsl index 4835cb9..bfc51e8 100644 --- a/ComputeShaders/mulMatTiled.hlsl +++ b/ComputeShaders/mulMatTiled.hlsl @@ -4,13 +4,10 @@ #ifndef TILE_SIZE static const uint TILE_SIZE = 32; #endif - #ifndef THREADS_Y -// Performance measures on Ryzen 7 5700G iGPU, the time is just for this shader: -// 1 (32 threads per group) - 17.1 seconds, 2 - 9.02424 seconds, 4 - 6.95762 seconds, 6 - 6.79011 seconds, 8 - 6.67279 seconds, 10 - 6.9456 seconds, 16 - 7.20502 seconds -// On nVidia, 8 is also the fastest option. static const uint THREADS_Y = 8; #endif +// The above values have a following constraint: TILE_SIZE = THREADS_Y * N * 4 where N is an integer #ifndef STREAM_SECOND_MATRIX // Funfact: enabling this on 1080Ti ruins the performance, by a factor of 3.5 |