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

llamafile : tmp disable + build sgemm.o when needed (#6716)

Browse source 
* build : sgemm.o only when needed

ggml-ci

* llamafile : tmp disable due to MoE bug

ggml-ci
This commit is contained in:
Georgi Gerganov 2024-04-17 23:58:26 +03:00 • committed by GitHub
parent 8dd1ec8b3f
commit 3b8f1ec4b1
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 41 additions and 25 deletions
Download patch file Download diff file Expand all files Collapse all files

28
Makefile
View file

@ -219,13 +219,6 @@ ifdef LLAMA_DISABLE_LOGS
MK_CPPFLAGS += -DLOG_DISABLE_LOGS
endif # LLAMA_DISABLE_LOGS
# disable ggml.c's use of sgemm.cpp
ifdef LLAMA_NO_LLAMAFILE
MK_CPPFLAGS += -DGGML_USE_LLAMAFILE=0
else
MK_CPPFLAGS += -DGGML_USE_LLAMAFILE=1
endif
# warnings
WARN_FLAGS = -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function
MK_CFLAGS += $(WARN_FLAGS) -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes -Werror=implicit-int \
@ -391,6 +384,15 @@ ifdef LLAMA_OPENBLAS
MK_LDFLAGS += $(shell pkg-config --libs openblas)
endif # LLAMA_OPENBLAS
# TODO: temporary disable until MoE is fixed
# https://github.com/ggerganov/llama.cpp/pull/6716
LLAMA_NO_LLAMAFILE := 1
ifndef LLAMA_NO_LLAMAFILE
MK_CPPFLAGS += -DGGML_USE_LLAMAFILE
OBJS += sgemm.o
endif
ifdef LLAMA_BLIS
MK_CPPFLAGS += -DGGML_USE_OPENBLAS -I/usr/local/include/blis -I/usr/include/blis
MK_LDFLAGS += -lblis -L/usr/local/lib
@ -487,11 +489,9 @@ ggml-cuda/%.o: ggml-cuda/%.cu ggml-cuda/%.cuh ggml.h ggml-common.h ggml-cuda/com
ggml-cuda.o: ggml-cuda.cu ggml-cuda.h ggml.h ggml-backend.h ggml-backend-impl.h ggml-common.h $(wildcard ggml-cuda/*.cuh)
$(NVCC_COMPILE)
endif # LLAMA_CUDA
ifdef LLAMA_CLBLAST
MK_CPPFLAGS += -DGGML_USE_CLBLAST $(shell pkg-config --cflags-only-I clblast OpenCL)
MK_CFLAGS += $(shell pkg-config --cflags-only-other clblast OpenCL)
MK_CXXFLAGS += $(shell pkg-config --cflags-only-other clblast OpenCL)
@ -610,6 +610,11 @@ ggml-mpi.o: ggml-mpi.c ggml-mpi.h
$(CC) $(CFLAGS) -c $< -o $@
endif # LLAMA_MPI
ifndef LLAMA_NO_LLAMAFILE
sgemm.o: sgemm.cpp sgemm.h ggml.h
$(CXX) $(CXXFLAGS) -c $< -o $@
endif
GF_CC := $(CC)
include scripts/get-flags.mk
@ -683,16 +688,13 @@ ggml-backend.o: ggml-backend.c ggml.h ggml-backend.h
ggml-quants.o: ggml-quants.c ggml.h ggml-quants.h ggml-common.h
$(CC) $(CFLAGS) -c $< -o $@
sgemm.o: sgemm.cpp sgemm.h ggml.h
$(CXX) $(CXXFLAGS) -c $< -o $@
unicode.o: unicode.cpp unicode.h
$(CXX) $(CXXFLAGS) -c $< -o $@
unicode-data.o: unicode-data.cpp unicode-data.h
$(CXX) $(CXXFLAGS) -c $< -o $@
OBJS += ggml-alloc.o ggml-backend.o ggml-quants.o unicode.o unicode-data.o sgemm.o
OBJS += ggml-alloc.o ggml-backend.o ggml-quants.o unicode.o unicode-data.o
llama.o: llama.cpp unicode.h ggml.h ggml-alloc.h ggml-backend.h ggml-cuda.h ggml-metal.h llama.h
$(CXX) $(CXXFLAGS) -c $< -o $@