New tool that improves research reproducibility by integrating High Performance Computing with Continuous Integration #RSAA25

Continuous Integration (CI) is essential in modern software development, enabling frequent code integration, automated testing, and rapid feedback. For research, CI enhances reproducibility and efficiency, but High Performance Computing (HPC) environments pose unique challenges, such as hardware diversity, job scheduling, resource management requirements, and strict administrative policies. We present ‘hpci’, a software tool developed to integrate mainstream CI platforms with High Performance Computing (HPC) environments. We demonstrate how this software, in conjunction with data version control (DVC) and other tools, enables researchers to validate reproducibility. Changes committed to workflows trigger HPC jobs to run the amended workflow on test data, with DVC verifying that identical results are produced. Jobs are monitored by ‘hpci’, so that test failures are easily visible in CI logs. Catching issues early in development accelerates debugging and optimisation, while minimising the risk of failures in large-scale production jobs. Automatic validation gives researchers confidence to adapt and extend workflows, which particularly benefits collaboration and migration between different HPC systems. As computational research continues to increase in scale, integrating CI with HPC is critical for maintaining reproducibility across platforms and labs, empowering researchers to focus on discoveries rather than debugging.