Modern C compilers play a critical role in software development, yet they still contain defects that can lead to incorrect binary programs. Optimization passes are especially risky because they apply aggressive code transformations. However, existing testing methods focus mainly on core correctness and give limited attention to optimization behavior. Thus, many optimization-introduced bugs remain hard to detect, and their broader impact is not well understood. My doctoral research will investigate how current compiler testing approaches perform in detecting optimization-related defects. I plan to conduct a large-scale study that examines bugs reported by existing testing frameworks together with those documented in official compiler repositories. This study will help identify recurring patterns in optimization failures and highlight weaknesses in current testing practice. Using the insights from this analysis, I will design an optimization-oriented fuzzing framework that guides test generation toward transformation contexts that frequently lead to incorrect binary program. I will also build an accompanying optimization-focused test suite to support compiler developers and researchers. The overall goal of my work is to provide a deeper understanding of how optimization interacts with testing and to improve the reliability of modern C compilers.

@inproceedings{paperbot11709,
  title = {Detection of Optimization-introduced Bugs in C Compilers},
  author = {Abdullah Al Noman},
  booktitle = {ICSE},
  year = {2026},
  doi = {10.1145/3774748.3787647}
}