For those who run multiple llama.cpp instances sharing Tesla P40

Posted by muxxington@reddit | LocalLLaMA | View on Reddit | 0 comments

gppm now supports power and performance state management with multiple llama.cpp instances. [https://github.com/crashr/gppm](https://github.com/crashr/gppm)