English(EN) b12293: cpu: accumulate flash attention V in F32 (#29810)

llama.cpp 更新修复了长上下文溢出问题

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-11 10:55

llama.cpp 项目发布了更新 b12293，该更新解决了其 CPU 内核在处理长上下文时出现的溢出问题。更新将内核修改为以 F32 格式累加 V 缓存，从而防止了之前导致 NaN 值的溢出。此更改确保了在扩展上下文长度时更稳定的性能，尽管在某些 ARM 架构上可能会导致速度略有下降。 AI

影响提高了 llama.cpp 推理引擎的稳定性和上下文处理能力。

排序理由这是针对特定开源项目的软件更新，并非前沿发布、重大行业举措或研究论文。

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 3 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 这是针对特定开源项目的软件更新，并非前沿发布、重大行业举措或研究论文。
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: infra; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

llama.cpp — Releases TIER_1 English(EN) · ServeurpersoCom · 2026-11-11 10:55

b12293: cpu: 在 F32 中累积 flash attention V (#29810)

<ul> <li>cpu: accumulate flash attention V in F32 in the one-chunk kernel</li> </ul> <p>The one-chunk kernel accumulated an F16 V cache in an F16 running sum<br /> that is only normalized at the end, which overflows past 65504 on long<br /> contexts: inf, then NaN on every follow…