ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-03-05 03:20:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

Q6_K_R4 (#130)

Browse Source 
* Adding q6_k_r4

* q6_k_r4: 1st functional AVX2 version

* q6_k_r4: AVX2 and simple Zen4

"Simple" as in processing 4 instead of 8 rows at once.
On Zen4 we get PP-512(LLaMA-3.1-8B) = 238.3 t/s vs
195.2 t/s for Q6_K. TG-128 @ 1 thread is 7.94 t/s
vs 5.38 t/s for Q6_K.

* q6_k_r4: 1st NEON version

PP-512(LLaMA-3.1-8B) = 78 t/s vs 57.6 t/s for q6_K.
TG-128 is slightly lower rthan q6_K for low number of threads,
becomes very slightly better at 8 threads.

* q6_k_r4: slightly faster NEON

PP-512(LLaMA-3.1-8B) = 83.25 t/s

* q6_k_r4: slightly faster Zen4

238.3 t/s -> 243.2 t/s

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2024-12-10 12:26:40 +01:00
committed by GitHub
parent 3ec193b485
commit 361174ee6a
10 changed files with 381 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
include/llama.h
View File

@@ -184,6 +184,7 @@ extern "C" {
LLAMA_FTYPE_MOSTLY_Q8_0_R4 = 207, // except 1d tensors
LLAMA_FTYPE_MOSTLY_Q5_0_R4 = 208, // except 1d tensors
LLAMA_FTYPE_MOSTLY_Q4_K_R4 = 214, // except 1d tensors
LLAMA_FTYPE_MOSTLY_Q6_K_R4 = 218, // except 1d tensors
LLAMA_FTYPE_MOSTLY_IQ4_NL_R4 = 225, // except 1d tensors
LLAMA_FTYPE_MOSTLY_IQ4_XS_R4 = 230, // except 1d tensors
LLAMA_FTYPE_MOSTLY_Q6_0_R4 = 235, // except 1d tensors