Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

CUDA: rename macros to avoid conflicts with WinAPI (#10736)

Browse Source 
* Renames NVIDIA GPU-architecture flags to avoid name clashes with WinAPI. (e.g. CC_PASCAL, GPU architecture or WinAPI pascal compiler flag?)

* Reverts erroneous rename in SYCL-code.

* Renames GGML_CUDA_MIN_CC_DP4A to GGML_CUDA_CC_DP4A.

* Renames the rest of the compute capability macros for consistency.
This commit is contained in:
Andreas Kieslinger
2024-12-10 18:23:24 +01:00
committed by GitHub
parent a86ad841f1
commit 750cb3e246
10 changed files with 69 additions and 71 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/convert.cu
+3 -3
View File
@@ -26,7 +26,7 @@ static __global__ void dequantize_block(const void * __restrict__ vx, dst_t * __
template <bool need_check>
static __global__ void dequantize_block_q8_0_f16(const void * __restrict__ vx, half * __restrict__ y, const int64_t k) {
#if __CUDA_ARCH__ >= CC_PASCAL
#if __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL
constexpr int nint = CUDA_Q8_0_NE_ALIGN/sizeof(int) + WARP_SIZE;
const int64_t i0 = CUDA_Q8_0_NE_ALIGN*blockIdx.x;
@@ -64,7 +64,7 @@ static __global__ void dequantize_block_q8_0_f16(const void * __restrict__ vx, h
GGML_UNUSED(y);
GGML_UNUSED(k);
NO_DEVICE_CODE;
#endif // __CUDA_ARCH__ >= CC_PASCAL
#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_PASCAL
}
template<typename dst_t>
@@ -599,7 +599,7 @@ to_fp16_cuda_t ggml_get_to_fp16_cuda(ggml_type type) {
case GGML_TYPE_Q5_1:
return dequantize_block_cuda<QK5_1, QR5_1, dequantize_q5_1>;
case GGML_TYPE_Q8_0:
if (ggml_cuda_info().devices[ggml_cuda_get_device()].cc >= CC_PASCAL) {
if (ggml_cuda_info().devices[ggml_cuda_get_device()].cc >= GGML_CUDA_CC_PASCAL) {
return dequantize_block_q8_0_f16_cuda;
}
return dequantize_block_cuda<QK8_0, QR8_0, dequantize_q8_0>;