Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦
C
/* CPU MXFP4 reference for tests/test_mxfp4_cuda.cu.
*
* quant.h is C: it declares `static _Thread_local QScratch g_qscratch`, which
* nvcc's C++ front end rejects. Compiling the reference here as C and exposing
* one symbol keeps the test comparing against the SAME code the engine runs,
* rather than a C++-friendly re-implementation that could drift from it. */
#include "../quant.h"
void mxfp4_ref(float *y, const float *x, const unsigned char *q4,
const unsigned char *e8s, int S, int I, int O) {
matmul_mxfp4(y, x, q4, e8s, S, I, O);
}