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

cleaning up the documentation

Browse source
This commit is contained in:
matteo serva 2024-07-31 16:43:53 +02:00
parent c1d255e91b
commit 390f54eb74
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
docs/build.md
View file

@ -180,7 +180,7 @@ For Jetson user, if you have Jetson Orin, you can try this: [Offical Support](ht
The environment variable [`CUDA_VISIBLE_DEVICES`](https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars) can be used to specify which GPU(s) will be used. The environment variable [`CUDA_VISIBLE_DEVICES`](https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars) can be used to specify which GPU(s) will be used.
The environment variable `GGML_CUDA_ENABLE_UNIFIED_MEMORY=1` can be used to enable unified memory in linux. This allows using system RAM when the GPU VRAM is exhausted. It is useful when the model barely fits in VRAM and inference is causing OOM errors. Should be enabled with `-ngl 99` to avoid sharing memory bandwidth with the CPU. In windows this setting is available in the nvidia control panel as `System Memory Fallback`. The environment variable `GGML_CUDA_ENABLE_UNIFIED_MEMORY=1` can be used to enable unified memory in linux. This allows using system RAM when the GPU VRAM is exhausted. Should be enabled with `-ngl 99` to avoid sharing memory bandwidth with the CPU. In windows this setting is available in the nvidia control panel as `System Memory Fallback`.
The following compilation options are also available to tweak performance: The following compilation options are also available to tweak performance: