CI/CD runtime systému MERIC

Abstract

This bachelor’s thesis focuses on the design and implementation of a Continuous Integration and Continuous Delivery (CI/CD) pipeline for MERIC, a tool used to analyze the energy consumption of parallel applications in high-performance computing (HPC) environments. The main objective was to develop a solution enabling automated, reliable, and reproducible testing across different hardware platforms. The proposed pipeline is based on GitLab CI and adapted to HPC-specific conditions, where jobs are executed via the Slurm scheduler. The solution includes a set of scripts for building the tool, running test scenarios, and validating output data. The tests cover different parallel execution modes and energy monitoring approaches. Validation mechanisms ensure the correctness of measured values, including CPU and GPU frequencies. The pipeline was tested on multiple types of compute nodes. The results demonstrate that continuous integration principles can be effectively applied even in HPC environments, providing a generally applicable approach to automated testing under these specific conditions.