BlueGene/L applications: Parallelism On a Massive Scale

Autor: Ümit V. Çatalyürek, Mehul Patel, Alan Gara, Robert K. Yates, Martin Schulz, José E. Moreira, Bor Chan, Kai Kadau, William Clarence McLendon, Franz Franchetti, Peter Williams, Andy Yoo, Keith Henderson, Bob Walkup, Bruce Hendrickson, Timothy C. Germann, George Almási, Christoph Überhuber, Erik W. Draeger, James C. Sexton, John A. Gunnels, Andrew W. Cook, Edmond Chow, Stefan Kral, Frederick H. Streitz, Vasily V. Bulatov, Jeffrey Greenough, Gyan Bhanot, Steve Louis, C. A. Rendleman, Manish Gupta, Charles J. Archer, Michael Welcome, Jürgen Lorenz, Francois Gygi, William H. Cabot, Bronis R. de Supinski, Alison Kubota, Peter S. Lomdahl, Brian J. Miller, Thomas E. Spelce, James N. Glosli
Rok vydání: 2008
Předmět:
Zdroj: The International Journal of High Performance Computing Applications. 22:33-51
ISSN: 1741-2846
1094-3420
Popis: BlueGene/L (BG/L), developed through a partnership between IBM and Lawrence Livermore National Laboratory (LLNL), is currently the world's largest system both in terms of scale, with 131,072 processors, and absolute performance, with a peak rate of 367 Tflop/s. BG/L has led the last four Top500 lists with a Linpack rate of 280.6 Tflop/s for the full machine installed at LLNL and is expected to remain the fastest computer in the next few editions. However, the real value of a machine such as BG/L derives from the scientific breakthroughs that real applications can produce by successfully using its unprecedented scale and computational power. In this paper, we describe our experiences with eight large scale applications on BG/ L from several application domains, ranging from molecular dynamics to dislocation dynamics and turbulence simulations to searches in semantic graphs. We also discuss the challenges we faced when scaling these codes and present several successful optimization techniques. All applications show excellent scaling behavior, even at very large processor counts, with one code even achieving a sustained performance of more than 100 Tflop/s, clearly demonstrating the real success of the BG/L design.
Databáze: OpenAIRE
Externí odkaz: