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

ggml : remove K_QUANTS_PER_ITERATION (#8306)

Browse source 
ggml-ci
This commit is contained in:
Georgi Gerganov 2024-07-10 15:23:12 +03:00 • committed by GitHub
parent 91deef4606
commit 117f7adbd9
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
13 changed files with 110 additions and 282 deletions
Download patch file Download diff file Expand all files Collapse all files

8
Makefile
View file

@ -640,12 +640,6 @@ ifdef GGML_CUDA_DMMV_F16
MK_NVCCFLAGS += -DGGML_CUDA_F16
endif # GGML_CUDA_DMMV_F16
ifdef GGML_CUDA_KQUANTS_ITER
MK_NVCCFLAGS += -DK_QUANTS_PER_ITERATION=$(GGML_CUDA_KQUANTS_ITER)
else
MK_NVCCFLAGS += -DK_QUANTS_PER_ITERATION=2
endif
ifdef GGML_CUDA_PEER_MAX_BATCH_SIZE
MK_NVCCFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=$(GGML_CUDA_PEER_MAX_BATCH_SIZE)
else
@ -734,7 +728,6 @@ ifdef GGML_HIPBLAS
GGML_CUDA_DMMV_X ?= 32
GGML_CUDA_MMV_Y ?= 1
GGML_CUDA_KQUANTS_ITER ?= 2
MK_CPPFLAGS += -DGGML_USE_HIPBLAS -DGGML_USE_CUDA
@ -751,7 +744,6 @@ endif # GGML_HIP_UMA
HIPFLAGS += $(addprefix --offload-arch=,$(AMDGPU_TARGETS))
HIPFLAGS += -DGGML_CUDA_DMMV_X=$(GGML_CUDA_DMMV_X)
HIPFLAGS += -DGGML_CUDA_MMV_Y=$(GGML_CUDA_MMV_Y)
HIPFLAGS += -DK_QUANTS_PER_ITERATION=$(GGML_CUDA_KQUANTS_ITER)
ifdef GGML_CUDA_FORCE_DMMV
HIPFLAGS += -DGGML_CUDA_FORCE_DMMV