diff options
| -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 = |