llama.cpp b10730 发布：优化 Qwen4Exp 索引头求和，提示处理提速

AI 摘要

llama.cpp 发布 b10730 版本，针对 Qwen4Exp 模型优化了 indexer 头的求和操作，通过切片方式替代转置和 sum_rows，减少了内存拷贝和计算量，在 RTX PRO 6000 上提示处理速度从 2170 提升到 2366 t/s，生成速度不受影响。该版本还提供了多平台预编译二进制文件。

正文节选

<details open> qwen4exp: sum the indexer heads by slices (#28023) * qwen4exp: sum the indexer heads by slices The head reduction went through a transpose and a sum_rows over ne[1], which left sum_rows with ne0 = 4, one block per row for a four element reduction, and the transpose copied the whole block by token surface twice on the way in. The heads are adjacent on ne[1], so each one is a strided view and the sum is a short chain of adds. RTX PRO 6000, Qwen3.8-Flash-Next UD-Q4_K_XL, fa on,

发布时间：2026-09-01 13:08

抓取时间：2026-09-01 13:44

来源机构：ggml-org