CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-31 08:51:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

cuda : check if this fixes Pascal card regression (#3882)

Browse Source
This commit is contained in:
Georgi Gerganov
2023-11-02 08:35:10 +02:00
committed by GitHub
parent 183b3fac6c
commit 4d719a6d4e
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml-cuda.cu
View File

@@ -7420,7 +7420,7 @@ static void ggml_cuda_mul_mat(const ggml_tensor * src0, const ggml_tensor * src1
} else if (all_on_device && !use_tensor_cores && src0->type == GGML_TYPE_F16 && !ggml_is_contiguous(src0) && !ggml_is_transposed(src1) && src1->ne[1] == 1) { } else if (all_on_device && !use_tensor_cores && src0->type == GGML_TYPE_F16 && !ggml_is_contiguous(src0) && !ggml_is_transposed(src1) && src1->ne[1] == 1) {
// KQV single-batch // KQV single-batch
ggml_cuda_mul_mat_vec_nc(src0, src1, dst); ggml_cuda_mul_mat_vec_nc(src0, src1, dst);
} else if (all_on_device && src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_F32 && !ggml_is_transposed(src0) && !ggml_is_transposed(src1)) { } else if (all_on_device && use_tensor_cores && src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_F32 && !ggml_is_transposed(src0) && !ggml_is_transposed(src1)) {
// KQ + KQV multi-batch // KQ + KQV multi-batch
ggml_cuda_mul_mat_mat_batched_cublas(src0, src1, dst); ggml_cuda_mul_mat_mat_batched_cublas(src0, src1, dst);
} else if (src0->type == GGML_TYPE_F32) { } else if (src0->type == GGML_TYPE_F32) {