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

refactor boolean logic

Browse source
This commit is contained in:
JohannesGaessler 2024-02-11 19:01:44 +01:00
parent 005de593ad
commit b1f6fab684
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

@ -9967,7 +9967,7 @@ static void ggml_cuda_mul_mat(const ggml_tensor * src0, const ggml_tensor * src1
#ifdef CUDA_USE_TENSOR_CORES
// when tensor cores are available, use them for large batch size
// ref: https://github.com/ggerganov/llama.cpp/pull/3776
use_mul_mat_q = use_mul_mat_q && !(fp16_performance_good && src1->ne[1] > MMQ_MAX_BATCH_SIZE);
use_mul_mat_q = use_mul_mat_q && (!fp16_performance_good || src1->ne[1] <= MMQ_MAX_BATCH_SIZE);
#endif // CUDA_USE_TENSOR_CORES
#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)