The BlueGene/L supercomputer and quantum ChromoDynamics

ABSTRACT

We describe our methods for performing quantum chromodynamics (QCD) simulations that sustain up to 20% of the peak performance on BlueGene supercomputers. We present our methods, scaling properties, and first cutting edge results relevant to QCD. We show how this enables unprecedented computational scale that brings lattice QCD to the next generation of calculations. We present our QCD simulation that achieved 12.2 Teraflops sustained performance with perfect speedup to 32K CPU cores. Among other things, these calculations are critical for cosmology, for the heavy ion experiments at RHIC-BNL, and for the upcoming experiments at CERN-Geneva. Furthermore, we demonstrate how QCD dramatically exposes memory and network latencies inherent in any computer system and propose that QCD should be used as a new, powerful HPC benchmark. Our sustained performance demonstrates the excellent properties of the BlueGene/L system.

References

1. IBM Journal of R&D 2005. Vol 49, Number 2/3, March/May 2005.
2. CreutzM. 1986. Quarks Gluons and lattices, Cambridge Monographs in Mathematical Physics.
3. Monvay I. and Munster G. 1993. Quantum Fields on a Lattice, Cambridge Monographs in Mathematical Physics.
4. The Columbia Physics System. http://www.epcc.ed.ac.uk/~ukqcd/cps/
5. Nielsen H. B., Ninomiya M. 1981. The no-go theorem for chiral lattice fermions. Nucl. Phys. B185, 20.
6. Wilson K. G. 1975. New Phenomena in Subnuclear Physics. ed. A Zichichi (Plenum Press, New York), Part A, 69.
7. Kogut J., Susskind L. 1975. Staggered lattice fermions. Phys. Rev. D11, 395.
8. Kaplan D. B. 1992. Five dimensional lattice fermions. Phys. Lett. B288, 342.
9. Vranas P. M. 1998. Chiral Symmetry Restoration in the Schwinger Model with Domain Wall Fermions. Phys. Rev. D57, 1415.
10. Salapura V., Walkup R. and Gara A. 2005. Exploiting Workload Parallelism for Performance and Power Optimization. IBM Research report RC23724, September.
11. Gonzalez R., Gordon B., and Horowitz M. 1997. Supply and threshold voltage scaling for low power CMOS. IEEE Journal of Solid State Circuits, 32(8):1210--1216, August.
12. Martin A., Nystroem M., and Penzes P. 2001. ET2: a metric for time and energy efficiency of computation. Power-Aware Computing. Kluwer Academic Publishers.

Index Terms

• Computer systems organization

  • Architectures

    • Distributed architectures

      • Grid computing

    • Parallel architectures

      • Multicore architectures

    • Serial architectures

      • Superscalar architectures

  • Dependable and fault-tolerant systems and networks

• Computing methodologies

  • Modeling and simulation

    • Simulation evaluation

    • Simulation types and techniques

      • Massively parallel and high-performance simulations

• General and reference

  • Cross-computing tools and techniques

    • Performance

• Hardware

  • Electronic design automation

    • Modeling and parameter extraction

• Networks

  • Network performance evaluation

• Software and its engineering

  • Software organization and properties

    • Software system structures

      • Distributed systems organizing principles

        • Grid computing