https://beanhsiang.github.io/tags/Performance/ Recent content in Performance on Innovation with tech Hugo zh-Hans Tue, 06 Oct 2026 12:00:00 +0800 https://beanhsiang.github.io/post/2026-10-06-how-to-make-ai-responses-faster-on-microsoft-foundry-lessons-from-2040-measurements/ Tue, 06 Oct 2026 12:00:00 +0800 https://beanhsiang.github.io/post/2026-10-06-how-to-make-ai-responses-faster-on-microsoft-foundry-lessons-from-2040-measurements/ <p> <a data-fancybox="gallery" href="https://beanhsiang.github.io/uploads/20261007113827-foundry-faster-responses-title.png"> <img class="mx-auto" alt="How to make AI responses faster on Microsoft Foundry" src="https://beanhsiang.github.io/uploads/20261007113827-foundry-faster-responses-title.png" /> </a> </p> <p><em>题图来自 Microsoft Foundry 博客的 <a href="https://techcommunity.microsoft.com/blog/azure-ai-foundry-blog/how-to-make-ai-responses-faster-on-microsoft-foundry-lessons-from-2040-measureme/4560073?WT.mc_id=AI-MVP-5003172">How to make AI responses faster on Microsoft Foundry: lessons from 2,040 measurements</a>，作者是 Yassine El Ghali。</em></p> <p>模型换得更快，应用不一定就更快。这篇文章的作者在 Microsoft Foundry 上把文本、图片、文件、工具和 MCP 几类工作负载都测了一遍，想知道哪些改动真能降低延迟，又不牺牲正确性。我读下来最大的感受是，延迟的来源比我平时想的杂得多：输出长度、重复的 prompt 内容、图片和文档处理、工具选择、连接建立、串行等待，还有多出来的模型请求，都会占时间。</p>