Webinar

The HPSF: Building the Future of High-Performance Software, Together

Series: HPC Best Practices Webinars

The High Performance Software Foundation (HPSF), a project within the Linux Foundation, champions, promotes, and advances core portable software for HPC. It achieves this by increasing adoption, lowering barriers to contribution, and supporting open-source HPC projects. Announced at SC23 and officially launched in May 2024, the foundation has rapidly grown into a global endeavor with a robust membership. Today, many critical software projects across the HPC stack, ranging in size and scope, have joined the HPSF.

This presentation will introduce the foundation and its operational model. We will detail the support the HPSF offers to hosted projects, including tools and cross-cutting initiatives for event organization, CI/CD, continuous benchmarking, outreach, training, packaging, and distribution.

Following this general overview, the presentation will focus on two projects of varying scales that joined the foundation with distinct needs: Environment Modules and WarpX. We will explore how their integration into the foundation was managed and the concrete benefits each project has realized.

Presenters

Presenter Bios

Julien Bigot is a research scientist at CEA, France, where he leads the “Science of Computing” team at Maison de la Simulation. His main research interest is programming models for HPC. His research focuses on separation of concerns between performance optimization and semantic expression. He focuses specifically on GPU programming with the Kokkos framework, which he co-leads; data handling, including I/O and in situ processing, with the PDI library he created and the French NumPEx/Exa-DoST project that he co-leads; and community building through his involvement in HPSF, where he was the first elected general member representative.

Xavier Delaruelle began his career as an HPC systems administrator and is now the manager of the HPC Operations team at CEA. In this role, he coordinates CEA’s technical activities for the Alice Recoque infrastructure, which will become EuroHPC’s second exascale supercomputer. He is also the lead developer and project manager of the open-source Environment Modules project.

Axel Huebl is a computational laser-plasma physicist working on exascale simulations. As a scientist at Berkeley Lab, he leads the software architecture of the Beam, Plasma & Accelerator Simulation Toolkit (BLAST). He co-first-authored the paper that won the 2022 ACM Gordon Bell Prize for running the BLAST code WarpX on Frontier, the first reported exascale machine. He is a strong advocate for open science and founded the open particle-mesh data project (openPMD) for self-describing, scalable I/O and data science.