Experimental Evaluation of Molecular Dynamics Simulations on Multi-core Systems

  • Sadaf R. Alam
  • Pratul K. Agarwal
  • Scott S. Hampton
  • Hong Ong
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5374)

Abstract

Multi-core processors introduce many challenges both at the system and application levels that need to be addressed in order to attain the best performance. In this paper, we study the impact of the multi-core technologies in the context of two scalable, production-level molecular dynamics simulation frameworks. Experimental analysis and observations in this paper provide for a better understanding of the interactions between the application and the underlying system features such as memory bandwidth, architectural optimization, and communication library implementation. In particular, we observe that parallel efficiencies could be as low as 50% on quad-core systems while a set of dual-core processors connected with a high speed interconnect can easily outperform the same number of cores on a socket or in a package. This indicates that certain modifications to the software stack and application implementations are necessary in order to fully exploit the performance of multi-core based systems.

Keywords

Multicore Performance Molecular Dynamics Simulation HPC 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Pearlman, D.A., Case, D.A., Caldwell, J.W., Ross, W.S., Cheatham, T.E., Debolt, S., Ferguson, D., Seibel, G., Kollman, P.: AMBER, a package of computer programs for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to simulate the structural and energetic properties of molecules. Comput. Phys. Commun. 91, 1–41 (1995)CrossRefMATHGoogle Scholar
  2. 2.
    Plimpton, S.J.: Fast Parallel Algorithms for Short-Range Molecular Dynamics. J. Comp. Phys. 117, 1–19 (1995), http://www.cs.sandia.gov/~sjplimp/lammps.html CrossRefMATHGoogle Scholar
  3. 3.
    Agarwal, P.K., Alam, S.R.: Biomolecular simulations on petascale: promises and challenges. Journal of Physics: Conference Series (SciDAC 2006) 46, 327–333 (2006)MathSciNetGoogle Scholar
  4. 4.
    Alam, S.R., Agarwal, P.K., Giest, A., Vetter, J.S.: Performance Characterization of Bio-molecular Simulations using Molecular Dynamics. In: ACM Symposium on Principle and Practices of Parallel Programming (PPOPP) (2006)Google Scholar
  5. 5.
    AMD Opteron Rev. F, http://multicore.amd.com/
  6. 6.
    Ramanthan, R.M.: Intel Multi-core Processors: Making the move to Quad-core and Beyond, white paper, www.intel.com/technology/architecture/downloads/quad-core-06.pdf
  7. 7.
    Vetter, J.S., et al.: Early Evaluation of the Cray XT3. In: IPDPS (2006)Google Scholar
  8. 8.
    Alam, S.R., Barrett, R.F., Kuehn, J.A., Roth, P.C., Vetter, J.S.: Characterization of Scientific Workloads on Systems with Multi-Core Processors. In: IEEE International Symposium on Workload Characterization (2006)Google Scholar
  9. 9.
    HPC Challenge benchmarks, icl.cs.utk.edu/hpcc/
  10. 10.
  11. 11.
    MD benchmarks for AMBER, CHARMM and NAMD, amber.scripps.edu/amber8.bench2.html
  12. 12.
    Performance Application Programming Interface (PAPI), icl.cs.utk.edu/papi/

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Sadaf R. Alam
    • 1
  • Pratul K. Agarwal
    • 1
  • Scott S. Hampton
    • 1
  • Hong Ong
    • 1
  1. 1.Computer Science and Mathematics DivisionOak Ridge National LaboratoryOak RidgeUSA

Personalised recommendations