From 6d0462d4a39085a9f9da04e0a5fc7cc9d4578818 Mon Sep 17 00:00:00 2001 From: Kawrakow Date: Mon, 2 Dec 2024 07:25:39 +0100 Subject: IQ4_NL_X4 (#118) * Adding iq4_nl_x4 Looks very promising - I get PP-512(LLaMA-3.1-8B) = 230 t/s on the Ryzen-7950X! This is faster than any other quant and ~40% faster than iq4_nl. * iq4_nl_x4: getting amazing This Zen4 variant gets us to PP-512(LLaMA-3.1-8B) = 263 t/s! * iq4_nl_x4: AVX2 Here we gain only 25% compared to iq4_nl * iq4_nl_x4: NEON On M2-Max we get PP-512(LLaMA-3.1-8B) = 109.7 t/s, up from 82.4 t/s for iq4_nl. * iq4_nl_x4: minor NEON improvement and cleanup This gets us to 110.3 t/s. In comparison, IQ4_NL_4_4 in mainline llama.cpp achieves 92.3 t/s. * iq4_nl_x4: NEON specialization for matrix x vector --------- Co-authored-by: Iwan Kawrakow --- include/llama.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/llama.h') diff --git a/include/llama.h b/include/llama.h index 965e5f50..f94dcb1a 100644 --- a/include/llama.h +++ b/include/llama.h @@ -179,6 +179,8 @@ extern "C" { LLAMA_FTYPE_MOSTLY_IQ3_KL = 146, // except 1d tensors LLAMA_FTYPE_MOSTLY_IQ2_KS = 147, // except 1d tensors LLAMA_FTYPE_MOSTLY_IQ4_KSS = 148, // except 1d tensors + // + LLAMA_FTYPE_MOSTLY_IQ4_NL_X4 = 225, // except 1d tensors LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file }; -- cgit v1.2.3