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

Update build flags.

Browse source 
GGML_CUDA_DMMV_Y is now GGML_CUDA_MMV_Y
so update your build instructions.

GGML_CUDA_FORCE_DMMV is always enabled.

---------

Co-authored-by: YellowRoseCx <80486540+YellowRoseCx@users.noreply.github.com>
This commit is contained in:
Henri Vasserman 2023-07-13 13:44:02 +03:00
parent cd36b185ff
commit 2ec4466db5
No known key found for this signature in database
GPG key ID: 2995FC0F58B1A986
2 changed files with 8 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

6
CMakeLists.txt
View file

@ -365,11 +365,9 @@ if (LLAMA_HIPBLAS)
add_compile_definitions(GGML_USE_HIPBLAS GGML_USE_CUBLAS)
add_library(ggml-rocm OBJECT ggml-cuda.cu ggml-cuda.h)
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_DMMV_X=${LLAMA_CUDA_DMMV_X})
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_DMMV_Y=${LLAMA_CUDA_DMMV_Y})
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_MMV_Y=${LLAMA_CUDA_MMV_Y})
target_compile_definitions(ggml-rocm PRIVATE K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER})
if (LLAMA_CUDA_FORCE_DMMV)
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_FORCE_DMMV)
endif()
target_compile_definitions(ggml-rocm PRIVATE GGML_CUDA_FORCE_DMMV)
set_source_files_properties(ggml-cuda.cu PROPERTIES LANGUAGE CXX)
target_link_libraries(ggml-rocm PRIVATE hip::device PUBLIC hip::host roc::hipblas)

14
Makefile
View file

@ -226,20 +226,18 @@ ifdef LLAMA_HIPBLAS
CC := $(ROCM_PATH)/llvm/bin/clang
CXX := $(ROCM_PATH)/llvm/bin/clang++
GPU_TARGETS ?= gfx803 gfx900 gfx906 gfx908 gfx90a gfx1030 gfx1100
LLAMA_CUDA_DMMV_X ?= 32
LLAMA_CUDA_DMMV_Y ?= 1
LLAMA_CUDA_DMMV_X ?= 32
LLAMA_CUDA_MMV_Y ?= 1
LLAMA_CUDA_KQUANTS_ITER ?= 2
CFLAGS += -DGGML_USE_HIPBLAS -DGGML_USE_CUBLAS $(shell $(ROCM_PATH)/bin/hipconfig -C)
CXXFLAGS += -DGGML_USE_HIPBLAS -DGGML_USE_CUBLAS $(shell $(ROCM_PATH)/bin/hipconfig -C)
LDFLAGS += -L/opt/rocm/lib -Wl,-rpath=$(ROCM_PATH)/lib -lhipblas -lamdhip64
OBJS += ggml-cuda.o
ifdef LLAMA_CUDA_KQUANTS_ITER
CXXFLAGS += -DK_QUANTS_PER_ITERATION=$(LLAMA_CUDA_KQUANTS_ITER)
else
CXXFLAGS += -DK_QUANTS_PER_ITERATION=2
endif
ggml-cuda.o: CXXFLAGS += $(addprefix --offload-arch=,$(GPU_TARGETS))
ggml-cuda.o: CXXFLAGS += -DGGML_CUDA_DMMV_X=$(LLAMA_CUDA_DMMV_X)
ggml-cuda.o: CXXFLAGS += -DGGML_CUDA_DMMV_Y=$(LLAMA_CUDA_DMMV_Y)
ggml-cuda.o: CXXFLAGS += -DGGML_CUDA_MMV_Y=$(LLAMA_CUDA_MMV_Y)
ggml-cuda.o: CXXFLAGS += -DGGML_CUDA_FORCE_DMMV
ggml-cuda.o: CXXFLAGS += -DK_QUANTS_PER_ITERATION=$(LLAMA_CUDA_KQUANTS_ITER)
ggml-cuda.o: ggml-cuda.cu ggml-cuda.h
$(CXX) $(CXXFLAGS) -x hip -c -o $@ $<
endif # LLAMA_HIPBLAS