Smart-Launcher (llama.cpp,ik_lama.cpp) Auto-tuning optimization script

[raketenkater]

HI i created a auto-tuning script for llama.cpp,ik_llama.cpp that gets you the max tokens per seconds on weird setups like mine 3090ti + 4070 + 3060.

https://github.com/raketenkater/llm-server