mirror of
https://github.com/ikawrakow/ik_llama.cpp.git
synced 2026-02-25 07:34:10 +00:00
This is a 1st working version on Zen4. We get PP-512(LLaMA-3.1-8B) = 226 t/s, so 16% slower than iq4_nl_x4.
This is a 1st working version on Zen4. We get PP-512(LLaMA-3.1-8B) = 226 t/s, so 16% slower than iq4_nl_x4.