diff options
| author | slaren <slarengh@gmail.com> | 2024-03-21 13:59:53 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-03-21 14:59:53 +0200 |
| commit | 03a8f8fafec2e21009be9fe7297d92e5d4538964 (patch) | |
| tree | 820954cad8602f4ee0bcdd80c2191c138afb80b9 /ggml-cuda.cu | |
| parent | cfd3be76e37dab92c846d75a2421178f20db4a11 (diff) | |
cuda : fix LLAMA_CUDA_F16 build (#6197)
Diffstat (limited to 'ggml-cuda.cu')
| -rw-r--r-- | ggml-cuda.cu | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/ggml-cuda.cu b/ggml-cuda.cu index 280839ea..04c6f5d0 100644 --- a/ggml-cuda.cu +++ b/ggml-cuda.cu @@ -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 = |