Iwan Kawrakow cd4266eb58 Experimenting with dequant + f16 GEMM on NEON ...

iq2_kt: PP512 = 79 t/s from 42 t/s
iq3_kt: PP512 = 81 t/s from 35 t/s

Also, found the reason why the f16 implementation for iq4_kt was
not working: it overflows. It works after mltiplying with the row scale
before doing the multiply-adds.

2025-05-31 16:10:33 +03:00

..

cmake

Merge mainline llama.cpp (#3)

2024-07-27 07:55:01 +02:00

include

Trellis quants with CPU inference (#441)

2025-05-23 09:17:52 +03:00

src

Experimenting with dequant + f16 GEMM on NEON

2025-05-31 16:10:33 +03:00

.gitignore

Merge mainline llama.cpp (#3)

2024-07-27 07:55:01 +02:00

CMakeLists.txt

Option to enable disable the IQK CPU FA kernels (#429)

2025-05-17 11:21:58 +03:00