ggrun

llama.cpp launcher that places big MoE models across mismatched multi-GPU rigs using exact VRAM math.

Desktopfreeglobal

ggrun is a free, open-source launcher for llama.cpp and ik_llama.cpp that calculates exact VRAM, RAM, and per-GPU bandwidth to place large mixture-of-experts models across a mismatched multi-GPU rig, rather than requiring identical GPUs. It's aimed at local LLM hobbyists trying to run big models on hardware they've assembled over time rather than a uniform GPU setup.

Website

Full match profile

Behind the summary, Matchbox keeps a richer profile of ggrun - the signals our matcher actually reads to decide when to surface it. It stays private; claim the listing to see and control it.

Problem & pain-point mapping
Who we surface it to (audience fit)
What it's a strong alternative to
Trust & credibility signals

Claim ggrun to unlock →

Try Matchbox with your own problem

Describe what is not working - we’ll show you whether ggrun (or something else) actually fits.

Try a problem