<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Floating-Point on 兜忆轩 | Cedric's Compiler Notes</title><link>https://douyixuan.github.io/tags/floating-point/</link><description>Recent content in Floating-Point on 兜忆轩 | Cedric's Compiler Notes</description><image><title>兜忆轩 | Cedric's Compiler Notes</title><url>https://douyixuan.github.io/images/papermod-cover.png</url><link>https://douyixuan.github.io/images/papermod-cover.png</link></image><generator>Hugo -- 0.146.0</generator><language>en-us</language><lastBuildDate>Fri, 23 Oct 2026 12:24:00 +0800</lastBuildDate><atom:link href="https://douyixuan.github.io/tags/floating-point/index.xml" rel="self" type="application/rss+xml"/><item><title>038 · 不知道 exp 怎么算，也能证明浮点 kernel 等价吗？</title><link>https://douyixuan.github.io/posts/ai-silicon-papers/2026-10-23/</link><pubDate>Fri, 23 Oct 2026 12:24:00 +0800</pubDate><guid>https://douyixuan.github.io/posts/ai-silicon-papers/2026-10-23/</guid><description>这篇论文把超越函数保留为不透明符号，证明生成 kernel 与形式规格对同一符号的使用一致。关键问题是：这种对任意函数都成立的定理，为什么仍依赖真实实现采用同一个函数？</description></item></channel></rss>