Skip to content

cuda : add half2 __shfl_xor() for ROCm 5.5 #7263

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

Merged
merged 1 commit into from
May 18, 2024
Merged

cuda : add half2 __shfl_xor() for ROCm 5.5 #7263

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
14 changes: 14 additions & 0 deletions ggml-cuda/common.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -315,6 +315,20 @@ static __device__ __forceinline__ int __dp4a(const int a, const int b, int c) {
#endif
return c;
}

#if defined(__HIP_PLATFORM_AMD__) && HIP_VERSION < 50600000
// __shfl_xor() for half2 was added in ROCm 5.6
static __device__ __forceinline__ half2 __shfl_xor(half2 var, int laneMask, int width) {
typedef union half2_b32 {
half2 val;
int b32;
} half2_b32_t;
half2_b32_t tmp;
tmp.val = var;
tmp.b32 = __shfl_xor(tmp.b32, laneMask, width);
return tmp.val;
}
#endif // defined(__HIP_PLATFORM_AMD__) && HIP_VERSION < 50600000
#endif // defined(GGML_USE_HIPBLAS)

#define FP16_AVAILABLE (defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) || __CUDA_ARCH__ >= CC_PASCAL
Expand Down
Loading