Leveraging high performance with Julia for bias reduction in galaxy observations
Reducing biases affecting galaxy observations, in particular their distance measurements, is crucial in astrophysics. This can, for example, help to estimate the value of the Hubble constant which controls the Universe expansion. We focus here on the performance optimization of a Fortran 90 code, minimizing biases in galaxy distance measurement catalogs thanks to a simulated annealing algorithm embedded within a Metropolis–Hastings sampling. We first rewrite this code in Julia to benefit from high-level programming without compromising on performance. We then introduce several numerical and algorithmic improvements, which enable us to strongly reduce the sequential computation time. Finally, we design and implement a multi-thread algorithm which manages to exploit the potential parallelism of the Metropolis–Hastings sampling, while preserving data dependencies. Performance tests show the efficiency and scalability of our parallel Julia code on up to 32 CPU cores, considering up to 400K galaxies.
Authors
- Mike Petrault (ORCID: https://orcid.org/0009-0002-0558-4976)
- Florent Bréhard (ORCID: https://orcid.org/0000-0003-3909-2099)
- Jenny G. Sorce (ORCID: https://orcid.org/0000-0002-2307-2432)
- Bilal El Safah
Institutions
- Centre National de la Recherche Scientifique (FR)
- Université de Lille (FR)
- Université Paris-Saclay (FR)
- Institut d'Astrophysique Spatiale (FR)
- Centre de Recherche en Informatique, Signal et Automatique de Lille (FR)
- École Centrale de Lille (FR)
Publication Details
- Journal
- The International Journal of High Performance Computing Applications
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1177/10943420261503626
- Primary Topic
- Parallel Computing and Optimization Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00