Merging binaries, code, and personalities: experiences from open-source development

ORAL  · Invited

Abstract

Open source software is as much or more of a sociological challenge than it is a technical one. While the benefits of open-source development are substantial: transparency, shared resources, reduced cost per person, and easier cross-code comparison, it also requires significant effort. Tasks such as code reviews, helping users, and fixing issues add burden and require thoughtful responses. Additionally, as high performance computer systems increase both in hardware complexity (e.g. CPU+GPU, FPGA, etc.) and scale (core counts, number of GPUs per node), the technical overhead of developing and maintaining large, efficient codes necessitates large collaborative efforts. I will discuss these challenges from my perspective with suggestions based on things learned in over a decade of experience in development.

*This material is based upon work supported by the National Science Foundation under Grants No. PHY-2309211, No. PHY-2309231, and No. OAC-2209656 at Caltech and by No. PHY-2407742, No. PHY-2207342, and No. OAC-2209655 at Cornell. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. This work was supported by the Sherman Fairchild Foundation at Caltech and Cornell.

Presenters

  • Nils Deppe

    • Cornell University

Authors

  • Nils Deppe

    • Cornell University