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

Fix hordeconfig max context setting, and add Makefile flags for cuda F16/KQuants per iter. (#252)

Browse source 
* Fix hordeconfig maxcontext setting.

* cuda: Bring DMMV_F16 and KQUANTS_ITER Makefile flags over from llama.
This commit is contained in:
Ycros 2023-06-22 01:01:46 +10:00 • committed by GitHub
parent dfdd20240c
commit b1f00fa9cc
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 9 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

8
Makefile
View file

@ -149,6 +149,14 @@ ifdef LLAMA_CUDA_DMMV_Y
else else
NVCCFLAGS += -DGGML_CUDA_DMMV_Y=1 NVCCFLAGS += -DGGML_CUDA_DMMV_Y=1
endif # LLAMA_CUDA_DMMV_Y endif # LLAMA_CUDA_DMMV_Y
ifdef LLAMA_CUDA_DMMV_F16
NVCCFLAGS += -DGGML_CUDA_DMMV_F16
endif # LLAMA_CUDA_DMMV_F16
ifdef LLAMA_CUDA_KQUANTS_ITER
NVCCFLAGS += -DK_QUANTS_PER_ITERATION=$(LLAMA_CUDA_KQUANTS_ITER)
else
NVCCFLAGS += -DK_QUANTS_PER_ITERATION=2
endif
ggml-cuda.o: ggml-cuda.cu ggml-cuda.h ggml-cuda.o: ggml-cuda.cu ggml-cuda.h
$(NVCC) $(NVCCFLAGS) $(CXXFLAGS) $(CUBLAS_CXXFLAGS) -Wno-pedantic -c $< -o $@ $(NVCC) $(NVCCFLAGS) $(CXXFLAGS) $(CUBLAS_CXXFLAGS) -Wno-pedantic -c $< -o $@
ggml_v2-cuda.o: otherarch/ggml_v2-cuda.cu otherarch/ggml_v2-cuda.h ggml_v2-cuda.o: otherarch/ggml_v2-cuda.cu otherarch/ggml_v2-cuda.h

2
koboldcpp.py
View file

@ -724,7 +724,7 @@ def main(args):
sys.exit(2) sys.exit(2)
if args.hordeconfig and args.hordeconfig[0]!="": if args.hordeconfig and args.hordeconfig[0]!="":
global friendlymodelname, maxhordelen, showdebug global friendlymodelname, maxhordelen, maxhordectx, showdebug
friendlymodelname = "koboldcpp/"+args.hordeconfig[0] friendlymodelname = "koboldcpp/"+args.hordeconfig[0]
if len(args.hordeconfig) > 1: if len(args.hordeconfig) > 1:
maxhordelen = int(args.hordeconfig[1]) maxhordelen = int(args.hordeconfig[1])