Parallel Processing Improvements in Julia Jet Reconstruction

This project aims to improve the parallel scaling of JetReconstruction.jl, a Julia package for jet clustering in high-energy physics. While the package already shows strong serial performance, its multi-threaded behaviour has not yet been systematically measured and appears suboptimal. My plan is to first build a reproducible benchmark suite, including a FastJet C++ reference, to measure event-level parallel performance across algorithms, multiplicity ranges, and thread counts. I will then use profiling to identify the dominant bottlenecks, such as allocation overhead, garbage collection pressure, load imbalance, or memory-locality issues. Based on that evidence, I will implement targeted optimisations and validate every change against reference outputs to ensure correctness is preserved. In addition to performance improvements, this project will produce reusable benchmarking and validation infrastructure, along with practical documentation for users on how to run JetReconstruction.jl efficiently on multi-core machines. The overall goal is to make the package scale better in parallel while also giving maintainers a clear, reproducible framework for future performance work.