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
/* SPDX-License-Identifier: GPL-3.0-or-later
Copyright (C) 2026 Maurizio Cammalleri */
/* dlfcn.h for Windows (src/common/win/): a DLL loaded and its functions
found, through LoadLibrary and GetProcAddress (win_posix.c). */
#ifndef JANAS_WIN_DLFCN_H
#define JANAS_WIN_DLFCN_H
#define RTLD_LAZY 1
#define RTLD_NOW 2
#define RTLD_LOCAL 0
#define RTLD_GLOBAL 0x100
void *dlopen(const char *file, int mode);
void *dlsym(void *handle, const char *name);
int dlclose(void *handle);
char *dlerror(void);
#endif