Thumbnail
Access Restriction
Subscribed

Author Morozov, Vitali ♦ Kumaran, Kalyan ♦ Finkel, Hal ♦ Pope, Adrian ♦ Peterka, Tom ♦ Lukić, Zarija ♦ Daniel, David ♦ Habib, Salman ♦ Heitmann, Katrin ♦ Vishwanath, Venkatram ♦ Frontiere, Nicholas ♦ Insley, Joe ♦ Fasel, Patricia
Source ACM Digital Library
Content type Text
Publisher Association for Computing Machinery (ACM)
File Format PDF
Language English
Abstract Supercomputing is evolving toward hybrid and accelerator-based architectures with millions of cores. The Hardware/Hybrid Accelerated Cosmology Code (HACC) framework exploits this diverse landscape at the largest scales of problem size, obtaining high scalability and sustained performance. Developed to satisfy the science requirements of cosmological surveys, HACC melds particle and grid methods using a novel algorithmic structure that flexibly maps across architectures, including CPU/GPU, multi/many-core, and Blue Gene systems. In this Research Highlight, we demonstrate the success of HACC on two very different machines, the CPU/GPU system Titan and the BG/Q systems Sequoia and Mira, attaining very high levels of scalable performance. We demonstrate strong and weak scaling on Titan, obtaining up to 99.2% parallel efficiency, evolving 1.1 trillion particles. On Sequoia, we reach 13.94 PFlops (69.2% of peak) and 90% parallel efficiency on 1,572,864 cores, with 3.6 trillion particles, the largest cosmological benchmark yet performed. HACC design concepts are applicable to several other supercomputer applications.
Description Affiliation: Los Alamos National Laboratory, Los Alamos, New Mexico (Daniel, David; Fasel, Patricia) || Argonne National Laboratory, Lemont, IL (Habib, Salman; Morozov, Vitali; Frontiere, Nicholas; Finkel, Hal; Pope, Adrian; Heitmann, Katrin; Kumaran, Kalyan; Vishwanath, Venkatram; Peterka, Tom; Insley, Joe) || Lawrence Berkeley National Laboratory, Berkeley, CA (Lukić, Zarija)
Age Range 18 to 22 years ♦ above 22 year
Educational Use Research
Education Level UG and PG
Learning Resource Type Article
Publisher Date 2005-08-01
Publisher Place New York
Journal Communications of the ACM (CACM)
Volume Number 60
Issue Number 1
Page Count 8
Starting Page 97
Ending Page 104


Open content in new tab

   Open content in new tab
Source: ACM Digital Library