gsoc project discussion

[ayushk687]

MDAnalysis is a widely used Python library for analyzing molecular dynamics simulations. As datasets grow larger and workflows become more complex, ensuring high performance across the core library is critical.

Currently, MDAnalysis uses the Airspeed Velocity (ASV) benchmarking framework for automated nightly performance tracking. However, the existing benchmark coverage is limited and does not comprehensively represent the performance of core functionalities.

This project aims to significantly expand benchmark coverage across the MDAnalysis codebase, identify performance bottlenecks, and implement targeted optimizations. By systematically benchmarking critical modules (e.g., trajectory handling, atom selection, distance calculations), the project will provide actionable insights into performance regressions and improvement opportunities.

The work will also establish a sustainable benchmarking strategy, ensuring that future contributions are performance-aware and regressions are detected early.

Expected Outcome:
Comprehensive ASV benchmark suite covering major core functionalities
Identification of performance bottlenecks using benchmark results
Implementation of optimized code paths for critical operations
Integration of benchmarks into CI workflows for continuous monitoring
Documentation for writing and maintaining ASV benchmarks
Performance comparison reports (before vs after optimization)
Recommended Skills:
Strong Python programming skills
Understanding of performance optimization techniques (profiling, vectorization, memory management)
Familiarity with benchmarking tools (ASV preferred)
Knowledge of NumPy and scientific computing workflows
Experience with Git and open-source contribution workflows

[BradyAJohnston]

Closing as the deadline for pre-prposals has passed and this reads like AI spam.