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

llama : offload to RPC in addition to other backends (#7640)

Browse source 
* llama : offload to RPC in addition to other backends

* - fix copy_tensor being called on the src buffer instead of the dst buffer

- always initialize views in the view_src buffer

- add RPC backend to Makefile build

- add endpoint to all RPC object names

* add rpc-server to Makefile

* Update llama.cpp

Co-authored-by: slaren <slarengh@gmail.com>

---------

Co-authored-by: slaren <slarengh@gmail.com>
This commit is contained in:
Radoslav Gerganov 2024-06-03 20:03:26 +03:00 • committed by GitHub
parent a5735e4426
commit bde7cd3cd9
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
6 changed files with 86 additions and 53 deletions
Download patch file Download diff file Expand all files Collapse all files

29
Makefile
View file

@ -69,6 +69,10 @@ ifeq ($(UNAME_S),Darwin)
endif
endif
ifdef LLAMA_RPC
BUILD_TARGETS += rpc-server
endif
default: $(BUILD_TARGETS)
test: $(TEST_TARGETS)
@ -429,6 +433,11 @@ ifdef LLAMA_BLIS
MK_LDFLAGS += -lblis -L/usr/local/lib
endif # LLAMA_BLIS
ifdef LLAMA_RPC
MK_CPPFLAGS += -DGGML_USE_RPC
OBJS += ggml-rpc.o
endif # LLAMA_RPC
ifdef LLAMA_CUBLAS
# LLAMA_CUBLAS is deprecated and will be removed in the future
LLAMA_CUDA := 1
@ -654,11 +663,26 @@ ggml-metal-embed.o: ggml-metal.metal ggml-common.h
endif
endif # LLAMA_METAL
OBJS += ggml-alloc.o ggml-backend.o ggml-quants.o unicode.o unicode-data.o
COMMON_H_DEPS = common/common.h common/sampling.h common/log.h llama.h
COMMON_DEPS = common.o sampling.o grammar-parser.o build-info.o json-schema-to-grammar.o
ifndef LLAMA_NO_LLAMAFILE
sgemm.o: sgemm.cpp sgemm.h ggml.h
$(CXX) $(CXXFLAGS) -c $< -o $@
endif
ifdef LLAMA_RPC
ggml-rpc.o: ggml-rpc.cpp ggml-rpc.h
$(CXX) $(CXXFLAGS) -c $< -o $@
rpc-server.o: examples/rpc/rpc-server.cpp ggml-rpc.h
$(CXX) $(CXXFLAGS) -c $< -o $@
rpc-server: rpc-server.o ggml.o llama.o $(COMMON_DEPS) $(OBJS)
$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
endif # LLAMA_RPC
GF_CC := $(CC)
include scripts/get-flags.mk
@ -738,14 +762,9 @@ unicode.o: unicode.cpp unicode.h
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
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 $@
COMMON_H_DEPS = common/common.h common/sampling.h common/log.h llama.h
COMMON_DEPS = common.o sampling.o grammar-parser.o build-info.o json-schema-to-grammar.o
common.o: common/common.cpp $(COMMON_H_DEPS)
$(CXX) $(CXXFLAGS) -c $< -o $@