Contributor
Manas Pratim Biswas

Estimating the energy cost of scientific software


Mentors
Benedikt Hegner, Caterina Doglioni
Organization
CERN-HSF
Technologies
python, machine learning, c++, git, deep learning, jupyter notebook, Software Profiling, Software Testing
Topics
High Throughput Computing, Scientific Software Profiling, Software Metrics, Estimating Energy Efficiency, Green Software Foundation
The Large Hadron Collider (LHC) experiments generate massive datasets composed of billions of proton-proton collisions. The analysis of this data requires high-throughput scientific computing that relies on efficient software algorithms. In this project, I aim to investigate whether small efficiency improvements in the LHC software can have a large energetic impact, given the sheer volume of data involved. Additionally, I aim to explore the impact of different computing architectures and job submission systems on energy efficiency. To achieve these goals, I will use metrics from the Green Software Foundation and other resources to estimate energy efficiency. I will then evaluate whether to make small changes to the code to improve efficiency and evaluate the potential savings. I will also test the software on different platforms and job submission systems. My expected results include a summary of metrics for software energy consumption, visualisation of test results, and identification of possible improvements to software algorithms. The project will provide valuable insights into the energy efficiency of scientific software, with potential applications beyond the LHC experiments.