Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-20 01:31:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

fixed cmake, 8bit MMV should be working now

Browse Source
This commit is contained in:
Concedo
2023-07-12 11:22:55 +08:00
parent 7516488550
commit 8f4ed0d18c
3 changed files with 37 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Makefile
+13 -5
View File
@@ -144,19 +144,27 @@ ifdef LLAMA_CUBLAS
CUBLASLD_FLAGS = -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L/usr/local/cuda/lib64 -L/opt/cuda/lib64 -L$(CUDA_PATH)/targets/x86_64-linux/lib
CUBLAS_OBJS = ggml-cuda.o ggml_v2-cuda.o ggml_v2-cuda-legacy.o
NVCC = nvcc
NVCCFLAGS = --forward-unknown-to-host-compiler -arch=native -DGGML_CUDA_FORCE_DMMV
NVCCFLAGS = --forward-unknown-to-host-compiler
ifdef CUDA_DOCKER_ARCH
NVCCFLAGS += -Wno-deprecated-gpu-targets -arch=$(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
ifdef LLAMA_CUDA_DMMV_X
NVCCFLAGS += -DGGML_CUDA_DMMV_X=$(LLAMA_CUDA_DMMV_X)
else
NVCCFLAGS += -DGGML_CUDA_DMMV_X=32
endif # LLAMA_CUDA_DMMV_X
ifdef LLAMA_CUDA_DMMV_Y
ifdef LLAMA_CUDA_MMV_Y
NVCCFLAGS += -DGGML_CUDA_MMV_Y=$(LLAMA_CUDA_MMV_Y)
NVCCFLAGS += -DGGML_CUDA_DMMV_Y=$(LLAMA_CUDA_DMMV_Y)
else ifdef LLAMA_CUDA_DMMV_Y
NVCCFLAGS += -DGGML_CUDA_MMV_Y=$(LLAMA_CUDA_DMMV_Y) # for backwards compatibility
else
NVCCFLAGS += -DGGML_CUDA_DMMV_Y=1
NVCCFLAGS += -DGGML_CUDA_MMV_Y=1
endif # LLAMA_CUDA_DMMV_Y
endif # LLAMA_CUDA_MMV_Y
ifdef LLAMA_CUDA_DMMV_F16
NVCCFLAGS += -DGGML_CUDA_DMMV_F16
endif # LLAMA_CUDA_DMMV_F16