CCRI: Medium: Cilk Infrastructure for Next-Generation Parallel-Programming Research

该项目正在开发一个新的开源软件平台和生态系统，称为OpenCilk，它基于屡获殊荣的Cilk并行计算技术。 随着摩尔定律的终结，应用程序的性能将不再提高--科学、国防和社会将无法享受到性能提高的成果--除非软件能够被设计得运行得更快，特别是通过使用并行计算：在多个处理器上运行应用程序。 OpenCilk提供了一个易于修改的软件平台，使并行编程研究人员能够探索在现代并行硬件上进行计算的新方法。 由于并行编程研究提供了允许领域专家对计算密集型应用程序进行编程的技术-包括新兴的国家优先事项，如人工智能，机器学习和计算工程和科学-该项目促进了并行计算和计算密集型应用领域的下一代研究。 该项目还支持Cilk Hub，一个社区驱动的组织，负责维护和增强OpenCilk，以及开发并行计算的教育材料。OpenCilk是一个完整的，强大的，开源的，易于访问的任务并行编程系统。 它提供了一个技术转移管道，使并行计算研究社区创建的研究贡献，以推进应用程序开发人员实践的并行编程。 随着时间的推移，该项目旨在整合来自语言，编译器，运行时系统，生产力工具，库和应用程序基准测试领域的成熟并行计算研究成果，然后将其集成到OpenCilk中。 Cilk Hub组织开发和发布OpenCilk软件，将并行计算技术上游到主流代码库，并提供教学材料和辅助文档。 它为研究人员和开发人员提供研讨会和教程，评估社区对OpenCilk技术的满意度，并管理社区提出的OpenCilk增强功能的集成。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This project is developing a new open-source software platform and ecosystem, called OpenCilk, which is based on the award-winning Cilk parallel-computing technology. As Moore's Law comes to an end, application performance will cease to improve - denying science, the national defense, and society the fruits of increasing performance - unless software can be engineered to run fast, especially through the use of parallel computing: the running of applications on multiple processors. OpenCilk provides an easy-to-modify software platform to enable parallel-programming researchers to explore new ways to compute on modern parallel hardware. Since parallel-programming research provides the technology that allows domain experts to program computationally intensive applications - including emerging national priorities such as artificial intelligence, machine learning, and computational engineering and science - this project fosters next-generation research in both parallel computing and computationally intensive application domains. This project also supports Cilk Hub, a community-driven organization charged with maintaining and enhancing OpenCilk, as well as developing educational materials for parallel computing.OpenCilk is a complete, robust, open-source, and easily accessible task-parallel programming system. It provides a technology-transfer conduit to enable research contributions created by the parallel-computing research community to advance parallel programming as practiced by application developers. Over time, this project aims to consolidate proven parallel-computing research results from the areas of languages, compilers, runtime systems, productivity tools, libraries, and application benchmarks, and then integrate them into OpenCilk. The Cilk Hub organization develops and releases OpenCilk software, upstreams parallel-computing technology to mainstream codebases, and provides teaching materials and ancillary documentation. It offers workshops and tutorials for researchers and developers, evaluates community satisfaction with the OpenCilk technology, and manages the integration of community-proposed enhancements to OpenCilk.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.