happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

HIP : fix RDNA4 build (#17792)

This commit is contained in:
Johannes Gäßler 2025-12-05 13:47:52 +01:00 committed by GitHub
parent 1be97831e4
commit 6016d0bd41
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
ggml/src/ggml-cuda/mma.cuh
View File

@ -560,7 +560,7 @@ namespace ggml_cuda_mma {
xi[0] = xs[0];
xi[1] = xs[1];
#endif // defined(RDNA4)
}else if constexpr (I == 16 && J == 8) {
} else if constexpr (I == 16 && J == 8) {
int64_t * xi = (int64_t *) t.x;
#if defined(RDNA4)
const int64_t * xs = (int64_t *) ((const int *) xs0 + (threadIdx.x % t.I) * stride + 4 * (threadIdx.x / t.I));
@ -577,14 +577,13 @@ namespace ggml_cuda_mma {
const int64_t * xs1 = xs + 2;
xi[2] = xs1[0];
xi[3] = xs1[1];
}else{
#endif // defined(RDNA4)
} else {
NO_DEVICE_CODE;
}
} else {
NO_DEVICE_CODE;
}
#endif // defined(RDNA4)
#else
#pragma unroll
for (int l = 0; l < t.ne; ++l) {