llama.cpp
[SYCL] Use native subgroup size for K-quant DMMV kernels on Intel
#21700
Merged

Commits
  • SYCL: use native subgroup size for K-quant DMMV
    PMZFX committed 11 days ago
Loading