SHF:Small: Collaborative research: Language-Integrated Verification for Deterministic Parallelism

SHF:Small：协作研究：确定性并行性的语言集成验证

• 批准号: 2127277
• 负责人: Ryan Newton
• 金额: $ 25万
• 依托单位: Purdue University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-03-15 至 2024-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2127277&HistoricalAwards=false
• 关键词: SHF; Small; Collaborative; research; Language

Formal verification promises to deliver software that works precisely according to a given mathematical specification. The last decade has seen several impressive formally verified software artifacts that are provably free of large classes of defects. However, these artifacts required Herculean feats of engineering, using specialized proof assistants that are separated, by a wide chasm, from the legacy languages and libraries that are used to engineer efficient parallel software. This project's novelties are in new techniques to enable the development of efficient parallel computing systems with formal assurances about correctness and reliability. This project's impacts will be to make formal verification a part of mainstream software development.This project will build on two approaches discovered by the investigators. The first is refinement reflection, which turns existing programming languages into theorem provers, where the proofs are merely programs in that same language. Second, the investigators have introduced lattice variables, scheduling algorithms and a new linear type system to simplify the construction of efficient parallel software with well-defined behavior. This project will combine the above ideas into a framework wherein engineers can develop and verify parallel software, simply by writing programs in their host language, thereby economically integrating formal proofs within existing software development cycles.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

形式验证承诺提供能够根据给定的数学规范精确运行的软件。在过去的十年中，出现了一些令人印象深刻的经过正式验证的软件工件，这些软件工件已被证明不存在大量缺陷。然而，这些工件需要艰巨的工程壮举，需要使用专门的证明助手，这些助手与用于设计高效并行软件的遗留语言和库之间存在巨大的鸿沟。该项目的新颖之处在于新技术能够开发高效的并行计算系统，并正式保证正确性和可靠性。该项目的影响将使形式验证成为主流软件开发的一部分。该项目将建立在研究人员发现的两种方法的基础上。第一个是细化反射，它将现有的编程语言转变为定理证明器，其中证明仅仅是同一语言的程序。其次，研究人员引入了晶格变量、调度算法和新的线性类型系统，以简化具有明确定义行为的高效并行软件的构建。该项目将上述想法结合到一个框架中，工程师只需用其宿主语言编写程序即可开发和验证并行软件，从而经济地将形式证明集成到现有软件开发周期中。该奖项反映了 NSF 的法定使命，并被认为是值得的通过使用基金会的智力优势和更广泛的影响审查标准进行评估来获得支持。