We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent d767dbe commit f7db90aCopy full SHA for f7db90a
ggml/src/ggml-cuda/fattn-wmma-f16.cuh
@@ -25,7 +25,7 @@ static bool ggml_cuda_should_use_wmma_fattn(const int cc) {
25
#if defined(GGML_USE_HIP) && !defined(GGML_HIP_ROCWMMA_FATTN)
26
return false;
27
#else
28
- if ((GGML_CUDA_CC_IS_NVIDIA(cc) && ggml_cuda_highest_compiled_arch(cc) >= GGML_CUDA_CC_VOLTA && cc == GGML_CUDA_CC_VOLTA) ||
+ if ((GGML_CUDA_CC_IS_NVIDIA(cc) && ggml_cuda_highest_compiled_arch(cc) == GGML_CUDA_CC_VOLTA) ||
29
GGML_CUDA_CC_IS_RDNA3(cc) || GGML_CUDA_CC_IS_MTHREADS(cc)) {
30
return true;
31
} else if (GGML_CUDA_CC_IS_CDNA(cc)){
0 commit comments