paperbot · PL 论文追踪

全部论文 待读 收藏

P2 Simulee

Simulee
ICSE Proceedings of the ACM/IEEE 42nd International Conference on Software Engineering2020匹配度 3/10 · 论文涉及程序分析(解释 LLVM 字节码)和 CUDA 并发缺陷检测,与 PL/程序分析有弱关联;但未涉及 LLM/Agent、机器学习系统、浮点程序分析或贝叶斯程序分析,与用户当前核心方向(PL↔AI 交叉、浮点/贝叶斯与 LLM)匹配度较低。引用 27 inbox
Mingyuan Wu, Yicheng Ouyang, Husheng Zhou, Lingming Zhang, Cong Liu, Yuqun Zhang
crossref: 2026-09-12

AI 速览

一句话:提出首个轻量级 CUDA 同步缺陷检测框架 Simulee,通过解释 LLVM 字节码收集访存信息来自动检测同步缺陷。

CUDA 已成为通用 GPU 计算的主流平台与编程模型,但高效检测 CUDA 同步缺陷仍是难题。本文提出首个轻量级检测框架 Simulee:通过解释 CUDA 程序对应的 LLVM 字节码来建模执行,并收集内存访问信息,以自动发现一般性 CUDA 同步缺陷。作者基于 GitHub 上 7 个流行 CUDA 项目构建基准并开展实验。结果表明,Simulee 检测出初步研究中人工识别的 24 个缺陷中的 21 个,并在各项目中发现 24 个此前未知缺陷,其中 10 个已被开发者确认;其性能显著优于现有最先进的 CUDA 同步缺陷检测方法。

问题
如何有效且高效地检测 CUDA 同步缺陷,是该领域尚未解决的开放问题。
方法
提出 Simulee,解释 CUDA 程序的 LLVM 字节码以建模执行,收集内存访问信息并自动检测一般性同步缺陷。
评估
在 GitHub 上 7 个流行 CUDA 项目构成的基准上实验,检测人工识别缺陷和未知缺陷,并与现有先进方法比较。
原文摘要(Abstract)

While CUDA has become a mainstream parallel computing platform and programming model for general-purpose GPU computing, how to effectively and efficiently detect CUDA synchronization bugs remains a challenging open problem. In this paper, we propose the first lightweight CUDA synchronization bug detection framework, namely Simulee, to model CUDA program execution by interpreting the corresponding LLVM bytecode and collecting the memory-access information for automatically detecting general CUDA synchronization bugs. To evaluate the effectiveness and efficiency of Simulee, we construct a benchmark with 7 popular CUDA-related projects from GitHub, upon which we conduct an extensive set of experiments. The experimental results suggest that Simulee can detect 21 out of the 24 manually identified bugs in our preliminary study and also 24 previously unknown bugs among all projects, 10 of which have already been confirmed by the developers. Furthermore, Simulee significantly outperforms state-of-the-art approaches for CUDA synchronization bug detection.

链接与引用

DOI 原文 · PDF(开放获取) ·

BibTeX
@inproceedings{paperbot5641,
  title = {Simulee},
  author = {Mingyuan Wu and Yicheng Ouyang and Husheng Zhou and Lingming Zhang and Cong Liu and Yuqun Zhang},
  booktitle = {ICSE},
  year = {2020},
  doi = {10.1145/3377811.3380358}
}