haraldwolff/llama.cpp
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity

cuda : fix LLAMA_CUDA_F16 build (#6197)

Browse source
This commit is contained in:
slaren 2024-03-21 13:59:53 +01:00 committed by GitHub
parent cfd3be76e3
commit 03a8f8fafe
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
ggml-cuda.cu
View file

@ -9453,7 +9453,7 @@ static void ggml_cuda_op_dequantize_mul_mat_vec(
// on some GPUs it is faster to convert src1 to half and to use half precision intrinsics
#ifdef GGML_CUDA_F16
cuda_pool_alloc<half> src1_dfloat_a;
ggml_cuda_pool_alloc<half> src1_dfloat_a(ctx.pool());
half * src1_dfloat = nullptr; // dfloat == half
bool src1_convert_f16 =