Skip to content

CPU kernel parity: decode and tile kernels at llama.cpp speed or better on zen2, zen4, Intel Granite Rapids and M4 Pro - #3924

Merged
borisbat merged 123 commits into
masterfrom
bbatkin/kernel-parity
Sep 2, 2026
Merged

CPU kernel parity: decode and tile kernels at llama.cpp speed or better on zen2, zen4, Intel Granite Rapids and M4 Pro#3924
borisbat merged 123 commits into
masterfrom
bbatkin/kernel-parity

Commits

Commits on Sep 2, 2026