千人铸剑：IDEAS生产力项目如何推进软件生产力与可持续性
计算机与社会
2024-02-20 v2
摘要
计算与数据驱动的科学与工程正在彻底改变从所有计算尺度上科学与社会的发展。例如，美国能源部（U.S. DOE）百亿亿次计算项目（Exascale Computing Project）的团队通过利用前所未有的百亿亿次计算能力，开拓建模、仿真与分析的新前沿——构建支持下一代应用并应对计算机架构颠覆性变化的先进软件生态系统。然而，人们对科学软件开发者的生产力、软件的可持续性及其产生结果的可信度日益担忧。IDEAS项目的成员充当催化剂以应对这些挑战，通过培育软件社区、孵化与策管方法论及资源，以及传播知识来提升开发者生产力与软件可持续性。本文讨论这些协同活动如何推进科学发现——通过为从笔记本电脑到集群再到百亿亿次及更大规模的可复现、可持续科学构建更坚实基础，从而缓解技术风险。
引用
@article{arxiv.2311.02010,
title = {A cast of thousands: How the IDEAS Productivity project has advanced software productivity and sustainability},
author = {Lois Curfman McInnes and Michael Heroux and David E. Bernholdt and Anshu Dubey and Elsa Gonsiorowski and Rinku Gupta and Osni Marques and J. David Moulton and Hai Ah Nam and Boyana Norris and Elaine M. Raybourn and Jim Willenbring and Ann Almgren and Ross Bartlett and Kita Cranfill and Stephen Fickas and Don Frederick and William Godoy and Patricia Grubel and Rebecca Hartman-Baker and Axel Huebl and Rose Lynch and Addi Malviya Thakur and Reed Milewicz and Mark C. Miller and Miranda Mundt and Erik Palmer and Suzanne Parete-Koon and Megan Phinney and Katherine Riley and David M. Rogers and Ben Sims and Deborah Stevens and Gregory R. Watson},
journal= {arXiv preprint arXiv:2311.02010},
year = {2024}
}
备注
12 pages, 1 figure