llama.cpp b10675 发布：Vulkan 后端优化 MoE 推理

AI 摘要

llama.cpp 发布 b10675 版本，主要更新为 Vulkan 后端新增对 shader 中行 ID 和专家计数的提升支持，优化了 MoE 模型的推理性能。该版本提供了多平台预编译二进制，包括 macOS、Linux、Windows、Android 等。

正文节选

<details open> Vulkan: add hoisting support for row IDs and expert count in shaders (#26686) * vulkan: add hoisting support for row IDs and expert count in shaders * use hoisted row ids in coopmat2 * vulkan: address review feedback on count_experts - use vk_op_count_experts_push_constants instead of a raw uint vector - apply the fastdiv trick to the ne00 div/mod in count_experts - compute the per-expert offsets with subgroupExclusiveAdd when the device supports it, keeping the serial path

发布时间：2026-08-28 23:48

抓取时间：2026-08-29 00:07

来源机构：ggml-org