<?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>Formal-Equivalence on 兜忆轩 | Cedric's Compiler Notes</title><link>https://douyixuan.github.io/tags/formal-equivalence/</link><description>Recent content in Formal-Equivalence 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>Tue, 29 Sep 2026 12:06:00 +0800</lastBuildDate><atom:link href="https://douyixuan.github.io/tags/formal-equivalence/index.xml" rel="self" type="application/rss+xml"/><item><title>014 · 通过验证的经验，就会变成可迁移知识吗？</title><link>https://douyixuan.github.io/posts/ai-silicon-papers/2026-09-29/</link><pubDate>Tue, 29 Sep 2026 12:06:00 +0800</pubDate><guid>https://douyixuan.github.io/posts/ai-silicon-papers/2026-09-29/</guid><description>ChipMEM 只把通过 EDA 验证的轨迹提炼为程序性记忆，并用工具调用统计指导重试；实验显示它能改善部分 RTL 优化和未见测试平台任务，但迁移证据仍受样本量与单次评测限制。</description></item><item><title>011 · 原 RTL 既是题目又是答案，评测到了什么？</title><link>https://douyixuan.github.io/posts/ai-silicon-papers/2026-09-26/</link><pubDate>Sat, 26 Sep 2026 12:26:00 +0800</pubDate><guid>https://douyixuan.github.io/posts/ai-silicon-papers/2026-09-26/</guid><description>RTL-BenchLS 用原 RTL 与形式等价构造自监督任务，将可验证评测扩展到 10,028 个设计；代价是把正确性锚定为复现参考实现的行为，而不是满足独立规格。</description></item><item><title>008 · 过了编译，为何长 RTL 仍几乎全错？</title><link>https://douyixuan.github.io/posts/ai-silicon-papers/2026-09-21/</link><pubDate>Mon, 21 Sep 2026 12:22:00 +0800</pubDate><guid>https://douyixuan.github.io/posts/ai-silicon-papers/2026-09-21/</guid><description>ChipVerilog 把 RTL 生成评测推进到长模块和跨模块层级：强模型大多能生成可编译代码，但功能正确率随规模与层级迅速下降；这也暴露了规格来源和验证口径的边界。</description></item></channel></rss>