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
Release 2026-10-11 (2): the binary package builds on Ubuntu 22.04
camauri8fcfc8556bda
diff --git a/.github/workflows/release.yml b/.github/workflows/release.ymlindex 424f648..94ff3fd 100644--- a/.github/workflows/release.yml+++ b/.github/workflows/release.yml@@ -30,7 +30,7 @@ jobs:linux-x86_64:runs-on: ubuntu-22.04steps:- - uses: actions/checkout@v4+ - uses: actions/checkout@v7- name: The shaders' compiler and the Vulkan headersrun: |@@ -101,7 +101,7 @@ jobs:- name: Keep the package (run by hand)if: github.event_name == 'workflow_dispatch'- uses: actions/upload-artifact@v4+ uses: actions/upload-artifact@v7with:name: ${{ env.PACKAGE }}path: |diff --git a/tests/test_github_parse.c b/tests/test_github_parse.cindex 286bc07..ea22a7c 100644--- a/tests/test_github_parse.c+++ b/tests/test_github_parse.c@@ -10,6 +10,7 @@* directory of its own under ~/tmp.* The sources belong to the program, so they are compiled in here.*/+#define _GNU_SOURCE /* memmem: Ubuntu 22.04's glibc (2.35) declares it only so */#include <stdio.h>#include <stdlib.h>#include <string.h>