[2603.20075] Agentic Harness for Real-World Compilers

[2603.20075] Agentic Harness for Real-World Compilers

arXiv - AI 3 min read

About this article

Abstract page for arXiv paper 2603.20075: Agentic Harness for Real-World Compilers

Computer Science > Software Engineering arXiv:2603.20075 (cs) [Submitted on 20 Mar 2026] Title:Agentic Harness for Real-World Compilers Authors:Yingwei Zheng, Cong Li, Shaohua Li, Yuqun Zhang, Zhendong Su View a PDF of the paper titled Agentic Harness for Real-World Compilers, by Yingwei Zheng and 4 other authors View PDF HTML (experimental) Abstract:Compilers are critical to modern computing, yet fixing compiler bugs is difficult. While recent large language model (LLM) advancements enable automated bug repair, compiler bugs pose unique challenges due to their complexity, deep cross-domain expertise requirements, and sparse, non-descriptive bug reports, necessitating compiler-specific tools. To bridge the gap, we introduce llvm-autofix, the first agentic harness designed to assist LLM agents in understanding and fixing compiler bugs. Our focus is on LLVM, one of the most widely used compiler infrastructures. Central to llvm-autofix are agent-friendly LLVM tools, a benchmark llvm-bench of reproducible LLVM bugs, and a tailored minimal agent llvm-autofix-mini for fixing LLVM bugs. Our evaluation demonstrates a performance decline of 60% in frontier models when tackling compiler bugs compared with common software bugs. Our minimal agent llvm-autofix-mini also outperforms the state-of-the-art by approximately 22%. This emphasizes the necessity for specialized harnesses like ours to close the loop between LLMs and compiler engineering. We believe this work establishes a founda...

Originally published on March 23, 2026. Curated by AI News.

Related Articles

Llms

Nicolas Carlini (67.2k citations on Google Scholar) says Claude is a better security researcher than him, made $3.7 million from exploiting smart contracts, and found vulnerabilities in Linux and Ghost

Link: https://m.youtube.com/watch?v=1sd26pWhfmg The Linux exploit is especially interesting because it was introduced in 2003 and was nev...

Reddit - Artificial Intelligence · 1 min ·
Llms

[P] I built an autonomous ML agent that runs experiments on tabular data indefinitely - inspired by Karpathy's AutoResearch

Inspired by Andrej Karpathy's AutoResearch, I built a system where Claude Code acts as an autonomous ML researcher on tabular binary clas...

Reddit - Machine Learning · 1 min ·
Llms

[R] BraiNN: An Experimental Neural Architecture with Working Memory, Relational Reasoning, and Adaptive Learning

BraiNN An Experimental Neural Architecture with Working Memory, Relational Reasoning, and Adaptive Learning BraiNN is a compact research‑...

Reddit - Machine Learning · 1 min ·
Llms

We hit 150 stars on our AI setup tool!

yo folks, we just hit 150 stars on our open source tool that auto makes AI context files. got 90 PRs merged and 20 issues that ppl are pi...

Reddit - Artificial Intelligence · 1 min ·
More in Llms: This Week Guide Trending

No comments

No comments yet. Be the first to comment!

Stay updated with AI News

Get the latest news, tools, and insights delivered to your inbox.

Daily or weekly digest • Unsubscribe anytime