ggml-org/llama.cpp b11172

on GitHub

metal : optimize sparse FA + clean-up (#29377)

• metal : cache sparse FA indices in shared memory

Assisted-by: pi:llama.cpp/DeepSeek-V4-Flash-Vision-Exp

• metal : simplify shared memory size calculation

Assisted-by: pi:llama.cpp/DeepSeek-V4-Flash-Vision-Exp

• pi : update general

• metal : unroll sparse index load

Assisted-by: pi:llama.cpp/DeepSeek-V4-Flash-Vision-Exp