Demystifying Deep Learning Compiler Front End Bugs: An LLM-Aided Empirical Study

A new empirical study of 123 frontend bugs in TorchDynamo, the default deep learning compiler frontend for PyTorch 2, has produced a taxonomy of 7 root cause categories and 15 subcategories, and used an LLM-aided methodology to uncover 23 previously unknown bugs (15 confirmed) in recent releases. The study, submitted to arXiv on 28 Jul 2026, is the first systematic analysis of frontend defects in a deep learning compiler, providing actionable insights for DLC development and testing.

Computer Science Programming Languages Submitted on 28 Jul 2026 Title:Demystifying Deep Learning Compiler Frontend Bugs: An LLM-Aided Empirical Study View PDF /pdf/2607.25651 HTML experimental https://arxiv.org/html/2607.25651v1 Abstract:Deep learning compilers DLCs are designed to translate deep learning programs into optimized, hardware-specific code. Typically, DLC frontends translate programs into graph-based intermediate representations IRs to enable optimizations. Defects introduced during this stage termed \emph{fBug}s are severe yet understudied, as prior work predominantly focuses on low-level APIs and operators or treats DLCs as monolithic entities. To bridge this gap, we conduct the first systematic empirical study of \emph{fBug}s in TorchDynamo, the default DLC frontend for PyTorch 2, the most popular DL framework. Leveraging a domain-knowledge-enhanced LLM-aided methodology, we analyze 123 \emph{fBug}s and construct a taxonomy comprising 7 root cause categories and 15 subcategories. Our findings provide actionable insights for DLC development and testing. Furthermore, we leverage the LLM to generate targeted, root cause-aware test cases to detect new bugs. We uncovered 23 previously unknown \emph{fBug}s in recent releases 15 confirmed across eight sub categories, demonstrating the efficacy of our methodology in testing and hardening DLC frontends. References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .