llama.cpp b11554：Vulkan 后端优化 mul_mat_id 重复处理
原标题：b11554
AI 摘要
llama.cpp 发布 b11554 版本，主要更新为 Vulkan 后端优化：在预处理阶段处理 mul_mat_id 的重复行 ID，而非循环处理，通过始终启用「hoist row ids」优化并生成紧凑的 tile 描述列表，在需要时发射多个 tile，从而收紧 workgroup 数量上界并让尾部提前退出。该版本同时提供 macOS、Linux、Windows、Android 等多平台多后端（CUDA、Vulkan、ROCm、SYCL、OpenVINO 等）的预编译二进制包。
正文节选
<details open> vulkan: handle mul_mat_id duplicates in the prepass rather than looping (#29998) * vulkan: compute every row of an expert in mul_mm_id when ids repeat * vulkan: handle mul_mat_id duplicates in the prepass rather than looping Extend the "hoist row ids" optimization to always be enabled and to emit a compact list of tile descriptions that need to run, and to emit multiple tiles when needed. Then we can launch a tighter upper bound on the number of workgroups and the tail can tri