Hermes - Towards an Ultimate High-Performance Algorithm for Cosmic Statistics of Large Data Sets (Feng)

4 Jun 2024, 09:30
17m
Lecture Theatre, 12 min + 5 discussion

Lecture Theatre, 12 min + 5 discussion

Description

Motivated by demanding a fast algorithm of cosmic statistics to tackle extremely large data sets from ongoing and upcoming galaxy surveys, we make the first release of Hermes (HypER-speed MultirEsolution cosmic Statistics), an open-source parallel/GPU-accelerated Python package. The core algorithm implemented in the Hermes Python package is based on the MRA-CS (MultiResolution Analysis - Cosmic Statistics) scheme. This talk is to present a brief description of core algorithm of Hermes, demonstrate its performances and numerical tests,and give an outlook of its wide applications in extracting cosmological information in upcoming galaxy surveys.

Presentation materials