Iwan Kawrakow 2986d3c21f Fused ffn_up*unary_op(ffn_gate) for MMVQ (no bias) ...

We see nearly 2% TG speedup for Ling-mini-2.0 and
about 1% for DeepSeek-Lite.

2025-10-24 11:27:18 +03:00

..

cmake

Merge mainline llama.cpp (#3)

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

include

Fused mul + multi_add op (#858)

2025-10-24 07:40:35 +03:00

src

Fused ffn_up*unary_op(ffn_gate) for MMVQ (no bias)

2025-10-24 11:27:18 +03:00

.gitignore

Merge mainline llama.cpp (#3)

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

CMakeLists.txt

Set default value of GGML_SCHED_MAX_COPIES to 1 (#751)

2025-09-02 07:04:39 +02:00