| layout |
|---|
default |
In the realm of High-Performance Computing (HPC), the sustainability of research software is a pivotal concern that intersects with the capabilities and efficiency of supercomputing resources worldwide. But what does "sustainability" entail within the HPC context? How can we measure and enhance the longevity, efficiency, and adaptability of HPC software? What are sustainable project pathways toward self-sustainment within foundations for projects of all sizes? This half-day workshop is dedicated to exploring the concept of sustainability in HPC software projects, diving into metrics and methodologies for assessment and improvement as well as how projects can join foundations over time. The workshop will begin with a discussion on existing metrics that have been defined by groups such as CHAOSS (https://chaoss.community), OpenSSF (https://openssf.org), and others. It will then extend into the discussion on adoption/ development of new metrics and on processes by which new metrics can be developed that address the unique needs of HPC software, including performance optimization, scalability, portability, and the integration of next-generation computing technologies.
The workshop will facilitate a deep dive into the sustainability of HPC software, exploring how sustainability can be quantified and improved in the context of supercomputing. This involves examining factors such as community health, software engineering practices tailored to HPC, and the funding landscape for HPC projects. By taking a holistic, cultural approach to this dimension, we aim to elucidate the sustainability needs of software and applications, leading to more comprehensive strategies for fostering resilient, efficient, and future-proof software ecosystems. We anticipate the workshop outcomes will have the following impact:
- Development of HPC Skills: Share best practices and knowledge in software sustainability.
- Diversity: Emphasize inclusive participation to improve the health and sustainability of software projects.
- Education and Training: Provide learning opportunities about software stewardship, funding opportunities, community engagement, and sustainability.
- Promotion of early career talents: Encourage researchers and developers new to HPC and the eScience community to participate and contribute.
Through collaboration with the eScience community, CORSA seeks to address the sustainability challenges unique to HPC software, including but not limited to, managing the lifecycle of software on HPC platforms, ensuring compatibility with diverse and evolving hardware architectures, optimizing software for energy efficiency and computational performance, and engaging in self-sustainment with through pathways to various foundations and funding organizations. With workshops at venues such as eScience, We aim to bring international focus to these challenges, building on experiences from the DOE Exascale Computing Project and other pivotal HPC initiatives around the globe.
We expect to hold future workshops that will enable the community to apply lessons learned from this proposed eScience workshop, and continue to dialog about explore avenues for improving the sustainability and stewardship of HPC software.
This workshop caters to a broad audience engaged in scientific computing and research software, including researchers, academics, software developers and maintainers, open-source foundation representatives, policy makers, funding agency representatives, students, young professionals, project managers, and team leaders. It focuses on software sustainability, software stewardship, foundation pathways, and best practices.
Elaine M. Raybourn, Sandia National Laboratories Daniel S. Katz, University of Illinois Urbana-Champaign Greg Watson, Oak Ridge National Laboratory