DocumentCode
37095
Title
Titan Takes on the Universe: Large-Scale Cosmology Simulation of Universe Mines for Halos Where Galaxies are Born
Author
Jones, Katie Elyce
Author_Institution
Oak Ridge Nat. Lab., Oak Ridge, TN, USA
Volume
17
Issue
3
fYear
2015
fDate
May-June 2015
Firstpage
68
Lastpage
69
Abstract
A multi-institutional team led by researchers at the US Department of Energy´s (DoE) Argonne National Laboratory has achieved high scalability and sustained performance on three of the most powerful systems in the world: IBM Blue Gene/Q supercomputers Mira at Argonne and Sequoia at Lawrence Livermore National Laboratory, and the Cray XK7 system Titan at Oak Ridge National Laboratory. The team set benchmarks for the largest cosmological simulations to date on different high-performance computing architectures on the massively parallel, CPU-only system Sequoia and on the hybrid accelerated Titan. This article takes a closer look at a recent Titan run and the team that´s currently pushing the state of the art in precision cosmology with the goal of better understanding cosmic acceleration.
Keywords
astronomy computing; cosmic acceleration; cosmology; galaxies; parallel architectures; Argonne; Argonne National Laboratory; Cray XK7 system Titan; IBM Blue Gene/Q supercomputers Mira; Lawrence Livermore National Laboratory; Oak Ridge National Laboratory; US Department of Energy; cosmic acceleration; galaxies; high-performance computing architectures; hybrid accelerated Titan run; large-scale cosmological simulations; multiinstitutional team; parallel CPU-only system Sequoia; precision cosmology; Cosmology; High performance computing; Scientific computing; Supercomputers; HPC; High-performance computing; cosmology; scientific computing; supercomputers;
fLanguage
English
Journal_Title
Computing in Science & Engineering
Publisher
ieee
ISSN
1521-9615
Type
jour
DOI
10.1109/MCSE.2015.63
Filename
7091787
Link To Document