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

CUDA: enable peer access between devices (#2470)

Browse source
This commit is contained in:
Johannes Gäßler 2023-09-17 16:37:53 +02:00 • committed by GitHub
parent 8b428c9bc8
commit 111163e246
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 63 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

5
Makefile
View file

@ -368,6 +368,11 @@ ifdef LLAMA_CUDA_KQUANTS_ITER
else
NVCCFLAGS += -DK_QUANTS_PER_ITERATION=2
endif
ifdef LLAMA_CUDA_PEER_MAX_BATCH_SIZE
NVCCFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=$(LLAMA_CUDA_PEER_MAX_BATCH_SIZE)
else
NVCCFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=128
endif # LLAMA_CUDA_PEER_MAX_BATCH_SIZE
#ifdef LLAMA_CUDA_CUBLAS
# NVCCFLAGS += -DGGML_CUDA_CUBLAS
#endif # LLAMA_CUDA_CUBLAS