一句话：对深度学习编译器缺陷进行大规模实证研究，揭示其与传统编译器缺陷的异同。
深度学习（DL）编译器被越来越多地用于生成优化代码，以提升 DL 模型在特定硬件上的运行时性能。然而，与传统编译器一样，DL 编译器也可能生成不正确的代码，导致模型行为异常，在关键任务系统中可能造成灾难性后果。另一方面，DL 编译器处理的 DL 模型与传统命令式程序有本质区别：DL 模型中的程序逻辑是隐式的。因此，传统编译器的缺陷特征需要在 DL 编译器背景下重新审视。本文对 DL 编译器缺陷进行了全面研究，旨在揭示其特性，并探讨这些特性与传统编译器缺陷的差异。
There are increasing uses of deep learning (DL) compilers to generate optimized code, boosting the runtime performance of DL models on specific hardware. Like their traditional counterparts, DL compilers can generate incorrect code, resulting in unexpected model behaviors that may cause catastrophic consequences in mission-critical systems. On the other hand, the DL models processed by DL compilers differ fundamentally from imperative programs in that the program logic in DL models is implicit. As such, various characteristics of the bugs arising from traditional compilers need to be revisited in the context of DL compilers.
DOI 原文 ·
@inproceedings{paperbot6347,
title = {A comprehensive study of deep learning compiler bugs},
author = {Qingchao Shen and Haoyang Ma and Junjie Chen and Yongqiang Tian and Shing-Chi Cheung and Xiang Chen},
booktitle = {FSE},
year = {2021},
doi = {10.1145/3468264.3468591}
}