llama.cpp

Files

T

Titaniumtown 6509138622 sycl: fuse mul_mat(gate) + mul_mat(up) + GLU for q4_K dense FFN (#26779 )

Measured on Arc Pro B70 (Battlemage, Level Zero), llama-bench -r 20, two
interleaved rounds, tg128:

    qwen2.5-3B-Instruct Q4_K_M    154.18 -> 158.53 t/s   +2.8%
    gemma-2-2b-it Q4_K_M          162.45 -> 165.62 t/s   +2.0%

llama-batched-bench on qwen2.5-3B, S_TG by batch size:

      B=1   142.72 -> 147.57 t/s    +3.4%
      B=2   243.72 -> 268.26 t/s   +10.1%
      B=4   359.58 -> 398.02 t/s   +10.7%
      B=8   449.75 -> 505.63 t/s   +12.4%

2026-08-14 02:26:23 -04:00

snapdragon

hexagon: MUL_MAT and MUL_MAT_ID rework : 32x32 tiled weight repack, kernel-params, cached graphs (#24954 )

2026-06-24 12:14:25 -07:00

VirtGPU

ggml-virtgpu: Fix some build commands (#20341 )

2026-03-12 15:47:45 +08:00

BLIS.md

make : deprecate (#10514 )

2024-12-02 21:22:53 +02:00

CANN.md

CANN: update docker images to 8.5.0 and improve CANN.md (#20801 )