KawrakowandGitHub 7642ac3eca Fix massive inefficiency in CUDA Q->f32/f16 and f32/f16->Q copies (#2279) ...

* CUDA indexer topk: this is better for PP

* Don't overstep

* Cleanup

* Allow Q8_0 cache in the CUDA DSA implementation

* DS4: do not cast caches to f32

* Fix massive inefficiency in CUDA Q->f32/f16 and f32/f16->Q copies

* Re-enable -ictk | --indexer-cache-type-k

2026-08-08 17:26:59 +03:00

..

cmake

Merge mainline llama.cpp (#3)

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

include

DS4: faster long-context TG (#2201)

2026-07-30 13:13:42 +03:00

src

Fix massive inefficiency in CUDA Q->f32/f16 and f32/f16->Q copies (#2279)

2026-08-08 17:26:59 +03:00

.gitignore

Merge mainline llama.cpp (#3)

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

CMakeLists.txt

Chunked experts (CPU) (#2202)

2026-07-30 13:16:02 +03:00