A New Parallel Intel Xeon Phi Hydrodynamics Code for Massively Parallel Supercomputers
- Autores: Kulikov I.1, Chernykh I.1, Tutukov A.2
-
Afiliações:
- Institute of Computational Mathematics and Mathematical Geophysics
- Institute of Astronomy
- Edição: Volume 39, Nº 9 (2018)
- Páginas: 1207-1216
- Seção: Part 1. Special issue “High Performance Data Intensive Computing” Editors: V. V. Voevodin, A. S. Simonov, and A. V. Lapin
- URL: https://journals.rcsi.science/1995-0802/article/view/203110
- DOI: https://doi.org/10.1134/S1995080218090135
- ID: 203110
Citar
Resumo
In this paper, a new hydrodynamics code called gooPhi to simulate astrophysical flows on modern Intel Xeon Phi processors with KNL architecture is presented. A new vector numerical method implemented in the form of a program code for massively parallel architectures is proposed. A detailed description is given and a parallel implementation of the code is made. A performance of 173 gigaflops and 48 speedup are obtained on a single Intel Xeon Phi processor. A 97 per cent scalability is reached with 16 processors.
Palavras-chave
Sobre autores
I. Kulikov
Institute of Computational Mathematics and Mathematical Geophysics
Autor responsável pela correspondência
Email: kulikov@ssd.sscc.ru
Rússia, pr. Akademika Lavrent’eva 6, Novosibirsk, 630090
I. Chernykh
Institute of Computational Mathematics and Mathematical Geophysics
Email: kulikov@ssd.sscc.ru
Rússia, pr. Akademika Lavrent’eva 6, Novosibirsk, 630090
A. Tutukov
Institute of Astronomy
Email: kulikov@ssd.sscc.ru
Rússia, ul. Pyatnitskaya 48, Moscow, 119017