diff options
| author | dylan <canardleteer@users.noreply.github.com> | 2023-07-07 11:25:25 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2023-07-07 21:25:25 +0300 |
| commit | 84525e7962bee0abef91108948bbf7f7bfdcf421 (patch) | |
| tree | e2e732cb057398249b7b98f46d57d3208a10a2e5 /Makefile | |
| parent | a7e20edf2266169ccd97a4eb949a593d628fbd64 (diff) | |
docker : add support for CUDA in docker (#1461)
Co-authored-by: canardleteer <eris.has.a.dad+github@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Diffstat (limited to 'Makefile')
| -rw-r--r-- | Makefile | 8 |
1 files changed, 7 insertions, 1 deletions
@@ -163,7 +163,12 @@ ifdef LLAMA_CUBLAS LDFLAGS += -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L/usr/local/cuda/lib64 -L/opt/cuda/lib64 -L$(CUDA_PATH)/targets/x86_64-linux/lib OBJS += ggml-cuda.o NVCC = nvcc - NVCCFLAGS = --forward-unknown-to-host-compiler -arch=native + NVCCFLAGS = --forward-unknown-to-host-compiler +ifdef CUDA_DOCKER_ARCH + NVCCFLAGS += -Wno-deprecated-gpu-targets -arch=$(CUDA_DOCKER_ARCH) +else + NVCCFLAGS += -arch=native +endif # CUDA_DOCKER_ARCH ifdef LLAMA_CUDA_FORCE_DMMV NVCCFLAGS += -DGGML_CUDA_FORCE_DMMV endif # LLAMA_CUDA_FORCE_DMMV @@ -187,6 +192,7 @@ ifdef LLAMA_CUDA_KQUANTS_ITER else NVCCFLAGS += -DK_QUANTS_PER_ITERATION=2 endif + ggml-cuda.o: ggml-cuda.cu ggml-cuda.h $(NVCC) $(NVCCFLAGS) $(CXXFLAGS) -Wno-pedantic -c $< -o $@ endif # LLAMA_CUBLAS |