diff options
| -rw-r--r-- | ggml-cuda.cu | 4 | ||||
| -rw-r--r-- | ggml-opencl.cpp | 4 | ||||
| -rw-r--r-- | llama.cpp | 2 |
3 files changed, 9 insertions, 1 deletions
diff --git a/ggml-cuda.cu b/ggml-cuda.cu index 7edd1a9..fed2a7c 100644 --- a/ggml-cuda.cu +++ b/ggml-cuda.cu @@ -13,6 +13,10 @@ #include "ggml-cuda.h" #include "ggml.h" +#if defined(_MSC_VER) +#pragma warning(disable: 4244 4267) // possible loss of data +#endif + static_assert(sizeof(half) == sizeof(ggml_fp16_t), "wrong fp16 size"); #define CUDA_CHECK(err) \ diff --git a/ggml-opencl.cpp b/ggml-opencl.cpp index 1d4db96..95f4cec 100644 --- a/ggml-opencl.cpp +++ b/ggml-opencl.cpp @@ -15,6 +15,10 @@ #include "ggml.h" +#if defined(_MSC_VER) +#pragma warning(disable: 4244 4267) // possible loss of data +#endif + #define CL_DMMV_BLOCK_SIZE 32 #define MULTILINE_QUOTE(...) #__VA_ARGS__ @@ -1253,7 +1253,7 @@ static void llama_model_load_internal( vram_scratch = n_batch * MB; ggml_cuda_set_scratch_size(vram_scratch); if (n_gpu_layers > 0) { - fprintf(stderr, "%s: allocating batch_size x 1 MB = %ld MB VRAM for the scratch buffer\n", + fprintf(stderr, "%s: allocating batch_size x 1 MB = %zd MB VRAM for the scratch buffer\n", __func__, vram_scratch / MB); } } |