Skip to content

Add option to disable MMA support on Turing #15360

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 4 commits into from
Closed

Add option to disable MMA support on Turing #15360

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
cb85412
Add option to disable MMA on Turing
pt13762104 Aug 16, 2025
36201c6
Revert wrong CMakeLists
pt13762104 Aug 16, 2025
6d7ef15
Recover FP16 performance and improve general performance
pt13762104 Aug 16, 2025
cf492dd
Fix comments
pt13762104 Aug 19, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions ggml/src/ggml-cuda/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -83,6 +83,10 @@ if (CUDAToolkit_FOUND)
add_compile_definitions(GGML_CUDA_FORCE_CUBLAS)
endif()

if (GGML_CUDA_DISABLE_TURING_MMA)
add_compile_definitions(GGML_CUDA_DISABLE_TURING_MMA)
endif()

if (GGML_CUDA_NO_VMM)
add_compile_definitions(GGML_CUDA_NO_VMM)
endif()
Expand Down
22 changes: 14 additions & 8 deletions ggml/src/ggml-cuda/common.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -224,9 +224,9 @@ typedef float2 dfloat2;
#define FAST_FP16_AVAILABLE
#endif // defined(FP16_AVAILABLE) && __CUDA_ARCH__ != 610

#if (!defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA) || defined(GGML_USE_MUSA)
#if ((!defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA) || defined(GGML_USE_MUSA)) && !defined(GGML_CUDA_DISABLE_TURING_MMA)
#define FP16_MMA_AVAILABLE
#endif // (!defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA) || defined(GGML_USE_MUSA)
#endif // ((!defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA) || defined(GGML_USE_MUSA)) && !defined(GGML_CUDA_DISABLE_TURING_MMA)

#if defined(GGML_HIP_ROCWMMA_FATTN) && (defined(CDNA) || defined(RDNA3) || (defined(GGML_HIP_ROCWMMA_FATTN_GFX12) && defined(RDNA4)))
#define FP16_MMA_AVAILABLE
Expand All @@ -236,9 +236,9 @@ typedef float2 dfloat2;
#define AMD_MFMA_AVAILABLE
#endif // defined(GGML_USE_HIP) && defined(CDNA) && !defined(GGML_HIP_NO_MMQ_MFMA)

#if !defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_TURING
#if !defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_TURING && !defined(GGML_CUDA_DISABLE_TURING_MMA)
#define TURING_MMA_AVAILABLE
#endif // !defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_TURING
#endif // !defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_TURING && !defined(GGML_CUDA_DISABLE_TURING_MMA)

#if !defined(GGML_USE_HIP) && __CUDA_ARCH__ >= GGML_CUDA_CC_AMPERE
#define AMPERE_MMA_AVAILABLE
Expand All @@ -252,6 +252,12 @@ typedef float2 dfloat2;
#define FLASH_ATTN_AVAILABLE
#endif // !defined(GGML_CUDA_NO_FA) && !(defined(GGML_USE_MUSA) && __MUSA_ARCH__ < 220)

#if defined(GGML_CUDA_DISABLE_TURING_MMA)
#define MMA_DISABLE (GGML_CUDA_CC_IS_NVIDIA(cc) && cc == 750)
#else
#define MMA_DISABLE 0
#endif

static bool fp16_available(const int cc) {
return ggml_cuda_highest_compiled_arch(cc) >= GGML_CUDA_CC_PASCAL;
}
Expand All @@ -262,7 +268,7 @@ static bool fast_fp16_available(const int cc) {

// To be used for feature selection of external libraries, e.g. cuBLAS.
static bool fast_fp16_hardware_available(const int cc) {
return (GGML_CUDA_CC_IS_NVIDIA(cc) && cc >= GGML_CUDA_CC_PASCAL && cc != 610) || GGML_CUDA_CC_IS_AMD(cc) ||
return (GGML_CUDA_CC_IS_NVIDIA(cc) && cc >= GGML_CUDA_CC_PASCAL && cc != 610 && !MMA_DISABLE) || GGML_CUDA_CC_IS_AMD(cc) ||
(GGML_CUDA_CC_IS_MTHREADS(cc) && cc >= GGML_CUDA_CC_QY2);
}

Expand All @@ -271,7 +277,7 @@ static bool fp16_mma_available(const int cc) {
#if defined(GGML_USE_HIP) && !defined(GGML_HIP_ROCWMMA_FATTN)
return false;
#else
if ((GGML_CUDA_CC_IS_NVIDIA(cc) && ggml_cuda_highest_compiled_arch(cc) >= GGML_CUDA_CC_VOLTA) ||
if ((GGML_CUDA_CC_IS_NVIDIA(cc) && ggml_cuda_highest_compiled_arch(cc) >= GGML_CUDA_CC_VOLTA && !MMA_DISABLE) ||
GGML_CUDA_CC_IS_CDNA(cc) || GGML_CUDA_CC_IS_RDNA3(cc) ||
GGML_CUDA_CC_IS_MTHREADS(cc)) {
return true;
Expand All @@ -289,7 +295,7 @@ static bool fp16_mma_available(const int cc) {

// To be used for feature selection of external libraries, e.g. cuBLAS.
static bool fp16_mma_hardware_available(const int cc) {
return (GGML_CUDA_CC_IS_NVIDIA(cc) && cc >= GGML_CUDA_CC_VOLTA) ||
return (GGML_CUDA_CC_IS_NVIDIA(cc) && cc >= GGML_CUDA_CC_VOLTA && !MMA_DISABLE) ||
GGML_CUDA_CC_IS_CDNA(cc) || GGML_CUDA_CC_IS_RDNA3(cc) || GGML_CUDA_CC_IS_RDNA4(cc) ||
(GGML_CUDA_CC_IS_MTHREADS(cc) && cc >= GGML_CUDA_CC_QY2);
}
Expand All @@ -312,7 +318,7 @@ static bool amd_mfma_available(const int cc) {

// Volta technically had FP16 tensor cores but they work very differently compared to Turing and later.
static bool turing_mma_available(const int cc) {
return GGML_CUDA_CC_IS_NVIDIA(cc) && ggml_cuda_highest_compiled_arch(cc) >= GGML_CUDA_CC_TURING;
return GGML_CUDA_CC_IS_NVIDIA(cc) && ggml_cuda_highest_compiled_arch(cc) >= GGML_CUDA_CC_TURING && !MMA_DISABLE;
}

static bool ampere_mma_available(const int cc) {
Expand Down