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

Multi GPU support, CUDA refactor, CUDA scratch buffer (#1703)

Browse source 
* CUDA multi GPU + scratch

ggml_cuda_compute_forward

Tensor parallelism

ggml_cuda_add

ggml_cuda_rms_norm

ggml_cuda_silu

CUDA scratch buffer

--main-gpu CLI option
This commit is contained in:
Johannes Gäßler 2023-06-06 21:33:23 +02:00 • committed by GitHub
parent 44f906e853
commit 17366df842
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
12 changed files with 1221 additions and 544 deletions
Download patch file Download diff file Expand all files Collapse all files

2
examples/main/README.md
View file

@ -286,5 +286,7 @@ These options provide extra functionality and customization when running the LLa
- `--verbose-prompt`: Print the prompt before generating text.
- `--mtest`: Test the model's functionality by running a series of tests to ensure it's working properly.
- `-ngl N, --n-gpu-layers N`: When compiled with appropriate support (currently CLBlast or cuBLAS), this option allows offloading some layers to the GPU for computation. Generally results in increased performance.
- `-mg i, --main-gpu i`: When using multiple GPUs this option controls which GPU is used for small tensors for which the overhead of splitting the computation across all GPUs is not worthwhile. The GPU in question will use slightly more VRAM to store a scratch buffer for temporary results. By default GPU 0 is used. Requires cuBLAS.
- `-ts SPLIT, --tensor-split SPLIT`: When using multiple GPUs this option controls how large tensors should be split across all GPUs. `SPLIT` is a comma-separated list of non-negative values that assigns the proportion of data that each GPU should get in order. For example, "3,2" will assign 60% of the data to GPU 0 and 40% to GPU 1. By default the data is split in proportion to VRAM but this may not be optimal for performance. Requires cuBLAS.
- `--lora FNAME`: Apply a LoRA (Low-Rank Adaptation) adapter to the model (implies --no-mmap). This allows you to adapt the pretrained model to specific tasks or domains.
- `--lora-base FNAME`: Optional model to use as a base for the layers modified by the LoRA adapter. This flag is used in conjunction with the `--lora` flag, and specifies the base model for the adaptation.