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: fix FP16 cuBLAS GEMM (#11396)

Browse Source
This commit is contained in:
Johannes Gäßler
2025-01-24 21:02:43 +01:00
committed by GitHub
parent 9fbadaef4f
commit c5d9effb49
1 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
ggml/src/ggml-cuda/ggml-cuda.cu
View File

@@ -1114,8 +1114,8 @@ static void ggml_cuda_op_mul_mat_cublas(
CUBLAS_CHECK(
cublasGemmEx(ctx.cublas_handle(id), CUBLAS_OP_T, CUBLAS_OP_N,
row_diff, src1_ncols, ne10,
&alpha, src0_ptr, CUDA_R_16F, ne00,
src1_ptr, CUDA_R_16F, ne10,
&alpha, src0_ptr, CUDA_R_16F, ne00,
src1_ptr, CUDA_R_16F, ne10,
&beta, dst_dd_i, CUDA_R_32F, ldc,
CUBLAS_COMPUTE_32F,
CUBLAS_GEMM_DEFAULT_TENSOR_OP));
@@ -1128,9 +1128,9 @@ static void ggml_cuda_op_mul_mat_cublas(
CUBLAS_CHECK(
cublasGemmEx(ctx.cublas_handle(id), CUBLAS_OP_T, CUBLAS_OP_N,
row_diff, src1_ncols, ne10,
&alpha_f16, src0_ptr, CUDA_R_16F, ne00,
src1_ptr, CUDA_R_16F, ne10,
&beta_f16, dst_dd_i, CUDA_R_16F, ldc,
&alpha_f16, src0_ptr, CUDA_R_16F, ne00,
src1_ptr, CUDA_R_16F, ne10,
&beta_f16, dst_f16.get(), CUDA_R_16F, ldc,
CUBLAS_COMPUTE_16F,
CUBLAS_GEMM_DEFAULT_TENSOR_OP));