[2604.08720] Demystifying the Silence of Correctness Bugs in PyTorch Compiler

Computer Science > Software Engineering arXiv:2604.08720 (cs) [Submitted on 9 Apr 2026] Title:Demystifying the Silence of Correctness Bugs in PyTorch Compiler Authors:Meiziniu Li, Dongze Li, Jianmeng Liu, Shing-Chi Cheung View a PDF of the paper titled Demystifying the Silence of Correctness Bugs in PyTorch Compiler, by Meiziniu Li and 3 other authors View PDF HTML (experimental) Abstract:Performance optimization of AI infrastructure is key to the fast adoption of large language models (LLMs). The PyTorch compiler (this http URL), a core optimization tool for deep learning (DL) models (including LLMs), has received due attention. However, this http URL is prone to correctness bugs, which cause incorrect outputs of compiled DL models without triggering exceptions, crashes, or warnings. These bugs pose a serious threat to the reliability of downstream LLM applications. Data from the PyTorch community shows that 19.2% of high-priority issues are incorrect outputs of compiled DL models induced by this http URL bugs, the second-most-common bug category (only behind program crashes at 19.57%). However, no systematic study has been conducted to specifically characterize and thereby detect these bugs. In this paper, we present the first empirical study of the correctness bugs in this http URL, examine their characteristics, and assess the effectiveness of existing fuzzers in detecting them. Based on our findings, we propose a proof-of-concept testing technique named AlignGuard, ta...