37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.
C avx2ccpu-inferenceggufinference-enginelinuxllmlocal-llmmixture-of-expertsmoenvmequantizationqwen3simdspeculative-decodingvulkan
| Name | Last commit message | Last commit date |
|---|---|---|
| .. | ||
| conf.c 1.7 KB | ||
| conf.h 1.0 KB | ||
| geo_data.h 2.7 MB | ||
| geo.c 20.0 KB | ||
| geo.h 4.3 KB | ||
| https_get.c 3.6 KB | ||
| https_get.h 1.3 KB | ||
| locate.c 7.6 KB | ||
| locate.h 1.3 KB | ||
| run.c 3.4 KB | ||
| run.h 1.2 KB | ||
| template.c 15.5 KB | ||
| template.h 2.6 KB | ||
| tz.c 7.6 KB | ||
| tz.h 1.1 KB | ||