CSR: Small: Safe Parallelization in a Dynamic Language

CSR：小：动态语言中的安全并行化

• 批准号: 1319617
• 负责人: Chen Ding
• 金额: $ 25.34万
• 依托单位: University of Rochester
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2013
• 资助国家: 美国
• 起止时间: 2013-10-01 至 2016-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1319617&HistoricalAwards=false
• 关键词: CSR; Small; Safe; Parallelization; Dynamic

Ubiquity of multicore processors has created a demand to support parallel processing in all programming languages. However, the dynamic nature of scripting languages makes supporting correct parallel programming in these languages especially difficult. Recent research has enabled safe parallel programming through speculative program parallelization. This project extends these techniques to support parallel programming in the Ruby language. The work focuses on three areas: the interpreter, the core language libraries, and the garbage collector. Through experimentation on Ruby, the research helps to understand the general synergy and limitations when a dynamic language meets dynamic parallelization.The resulting parallel Ruby implementation is a general-purpose parallel language that can be used by non-programming experts. Because of its ease of use, it is expected to be very useful for rapid prototyping of parallel programming applications and teaching parallel programming across all scientific disciplines. Transfer of this technology is facilitated by open-source distribution of the code on the Web. Outreach activities associated with the project include work with pre-college students, recruitment of students from under-represented groups, and collaboration with industry. Education is integrated with the research through involvement of students in the language implementation, and by experimentation with the use of the parallel Ruby language in several courses.

多核处理器的普遍存在产生了在所有编程语言中支持并行处理的需求。然而，脚本语言的动态特性使得在这些语言中支持正确的并行编程变得特别困难。最近的研究通过推测性程序并行化实现了安全的并行编程。该项目扩展了这些技术，以支持Ruby语言中的并行编程。这项工作主要集中在三个方面：解释器、核心语言库和垃圾收集器。通过在Ruby上的实验，研究有助于了解动态语言在遇到动态并行化时的一般协同作用和局限性，最终得到的并行Ruby实现是一种可供非编程专家使用的通用并行语言。由于它的易用性，它有望在并行编程应用程序的快速原型制作和跨所有科学学科的并行编程教学中非常有用。这项技术的转让是通过在Web上公开分发代码来促进的。与该项目相关的外联活动包括与大学预科学生合作，从代表性不足的群体中招募学生，以及与行业合作。教育与研究相结合，通过让学生参与语言实现，并在几门课程中试验使用并行Ruby语言。

项目成果

Timescale functions for parallel memory allocation

• DOI: 10.1145/3315573.3329987
• 发表时间: 2019-06
• 期刊: Proceedings of the 2019 ACM SIGPLAN International Symposium on Memory Management
• 作者: Pengcheng Li;Hao Luo;C. Ding

Beating OPT with Statistical Clairvoyance and Variable Size Caching

• DOI: 10.1145/3297858.3304067
• 发表时间: 2019-04
• 期刊: Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems
• 作者: Pengcheng Li;Colin Pronovost;William D. Wilson;B. Tait;Jie Zhou;C. Ding;J. Criswell

Codestitcher: inter-procedural basic block layout optimization

Codestitcher：程序间基本块布局优化

• DOI: 10.1145/3302516.3307358
• 发表时间: 2019
• 期刊: Proceedings of the 28th International Conference on Compiler Construction {CC} 2019
• 作者: Lavaee, Rahman;Criswell, John;Ding, Chen
• 通讯作者: Ding, Chen