Scalable Delivery of Scalable Libraries and Tools: How ECP Delivered a Software Ecosystem for Exascale and Beyond (2311.06995v1)

Abstract: The Exascale Computing Project (ECP) was one of the largest open-source scientific software development projects ever. It supported approximately 1,000 staff from US Department of Energy laboratories, and university and industry partners. About 250 staff contributed to 70 scientific libraries and tools to support applications on multiple exascale computing systems that were also under development. Funded as a construction project, ECP adopted an earned-value management system, based on milestones. and a key performance parameter system based, in part, on integrations. With accelerated delivery schedules and significant project risk, we also emphasized software quality using community policies, automated testing, and continuous integration. Software Development Kit teams provided cross-team collaboration. Products were delivered via E4S, a curated portfolio of libraries and tools. In this paper, we discuss the organizational and management elements that enabled the efficient and effective delivery of ECP libraries and tools, lessons learned and next steps.