Harshula Jayasuriya1, Aidan Heerdegen1, Paul Leopardi1, Tommy Gatti1, Scott Wales2, Rui Yang3
1Australian Earth System Simulator (ACCESS-NRI), Australia, 2Bureau of Meteorology, Australia, 3National Computational Infrastructure (NCI), Australia
Biography:
Harshula is a member of the Model Release team at ACCESS-NRI. He is Free/Open Source Software advocate and a Debian Developer with two decades of experience working with software build systems. Once upon a time, he was a Linux/Unix filesystem engineer working on NFS and UDF at Red Hat Inc and Silicon Graphics Inc.
Aidan is the model release team lead for ACCESS-NRI. He is passionate about making software work for people. In previous roles he provided computational support to climate researchers in two ARC Centres of Excellence, and prior to that was a physical chemist working in a software role supporting X-ray scattering research.
Paul Leopardi is a Research Software Engineer at ACCESS-NRI, currently working on supporting an Earth System Model. His qualifications include a PhD in computational mathematics and experience in teaching numerical linear algebra, optimization and combinatorics. His 40 year background in computing includes working with parallel Fortran codes, High Performance Computing and postprocessing of operational weather forecasting data.
Tommy is a Research Software Engineer for the Model Release Team at ACCESS-NRI. He's a multidisciplinary programmer who enjoys developing CI/CD infrastructure and technical pathfinding in different domains. In previous industry roles he has been a driver developer and member of a third level support team.
Scott is part of the Next Generation Modelling team at the Bureau of Meteorology where he develops and supports HPC software infrastructure for numerical weather prediction. His focus is on developing reproducible model environments across different platforms. Previous roles include providing scientific computing support to climate researchers at the ARC CoE for Climate Extremes.
Rui works as a Senior HPC Specialist at the National Computational Infrastructure (NCI), where he leads the Software and Data Modernisation team. His role involves building and deploying research software environments across diverse domains, including climate and weather, geophysics, and data science. Additionally, Rui specializes in performance profiling and tuning for academic and data science packages.
Abstract:
The seamless integration and deployment of deterministic software is critical to enhance collaboration and achieve organisational objectives. Maintaining complex, bespoke software for scientific research and adhering to the FAIR principles (Findability, Accessibility, Interoperability, and Reusability) is a difficult task. It can be made easier by using tools such as Spack, a build-from-source package manager. Spack enables reliably building software, replicating results and reproducing science.
Three organisations working on climate research software had converged on using Spack and started a regular meeting in 2023, during Australian business hours. During these meetings, we discussed the issues that we faced and how to solve them. For example, discussions covered the different versions of compilers, dependencies, Github/Gitlab based CI/CD and end-user needs of each organisation. The objectives of the organisations have been achieved in a timely manner due to the capability and flexibility of Spack.
There is a global Spack community. However, events and interactive discussions are typically scheduled during North American and European business hours.
This BoF is an opportunity to connect with like-minded professionals to build a regional community to accelerate scientific discovery and innovation. The session will facilitate an interactive discussion among users, developers, and administrators to explore the challenges and best practices of CI/CD workflows. There will be short presentations from each organisation demonstrating the utility of Spack. These will be followed by a directed discussion, open to all participants, on deterministic software build workflows adhering to FAIR principles.