Vectorization on Monte Carlo particle transport: an architectural study using the LANL benchmark “GAMTEB”

ABSTRACT

Fully vectorized versions of the Los Alamos National Laboratory benchmark code Gamteb, a Monte Carlo photon transport algorithm, were developed for the Cyber 205/ETA-10 and Cray X-MP/Y-MP architectures. Single-processor performance measurements of the vector and scalar implementations were modeled in a modified Amdahl's Law that accounts for additional data motion in the vector code. The performance and implementation strategy of the vector codes are related to architectural features of each machine. Speedups between fifteen and eighteen for Cyber 205/ETA-10 architectures, and about nine for CRAY X-MP/Y-MP architectures are observed. The best single processor execution time for the problem was 0.33 seconds on the ETA-10G, and 0.42 seconds on the CRAY Y-MP.

REFERENCES

• 1: N. Metropolis, "The Beginning of the Monte Carlo Method," Los Alamos Science, 15, pps. 122-143 (1987).
• 2: S. Ulam, R. D. Riehtmeyer and J. Von Neumann, "Statistical Methods in Neutron Diffusion," LAMS-551, Los Alamos National Laboratory report (1947).
• 3: S. Ulam and N. Metropolis, "The Monte Carlo Method," J. Am. Stat. Assoc., 44, 335(1949).
• 4: J.H. Curtis, "Sampling Methods Applied to Differential and Difference Equations," Proceedings IBM Seminar on Scientific Computation, IBM Corporation, New York, NY, pps. 87-109 (November 1949).
• 5: J. M. Hammersley and D. C. Handscomb, Monte Carlo Methods, Meuthen, London (1964).
• 6: k. R. Howell, "Application of Monte Carlo to Heat Transfer Problems," Advances in Heat Transfer, 5, Academic Press, New York, NY (1968).
• 7: K. Binder, Applications of the Monte Carlo Method in Statistical Physics, Springer-Verlag, Berlin (i 984).
• 8: N. Metropolis, "Monte Carlo: In the Beginning and Some Great Expectations," Monte Carlo Methods and Applications in Neutronics, Photonics and Statistical Physics, Cadarache Castle, France (1985).
• 9: A. Haji-Sheik, "Monte Carlo Methods," Handbook of Numerical Heat Transfer, John Wiley & Sons, Inc., New York, NY, pps. 673-722 (1988).
• 10: T.B. Brown, Vectorized Monte Carlo, PhD Dissertation, Department of Nuclear Engineering, University of Michigan (1981).
• 11: W.R. Martin, P. F. Nowak and J. A. Rathkopf, "Monte Carlo Photon Tracing on a Vector Supercomputer," IBM Journal of Research and Development, 30, No. 2 (March 1986).
• 12: Sequent Computer Systems, "Parallel Ray Tracing Study," TN-85-09 (rvp), Rev. 1.0 (1985).
• 13: F.W. Bobrowicz, I. E. Lynch and K. J. Fisher, "Vectorized Monte Carlo Photon Transport," Parallel Computing, 1,(1984).
• 14: F, W. Bobrowicz, K. J. Fisher, and R. G. Brickner, "Vectorized Monte Carlo Neutron Transport," Los Alamos National Laboratory report LA-UR-84-1269 (1984).
• 15: J. Briesmeister, ed., "MCNP: A General Monte Carlo Code for Neutron and Photon Transport," Los Alamos National Laboratory report LA-7396-M, Rev.2, September 1986.
• 16: M. L. Simmons , H. J. Wasserman, Performance comparison of the Cray-2 and Cray X-MP/416 supercomputers, Proceedings of the 1988 ACM/IEEE conference on Supercomputing, p.288-295, November 12-17, 1988, Orlando, Florida, United States
• 17: D.V. Pryor and P. J. Bums, "A Parallel Monte Carlo Model for Radiative Heat Transfer," Presented at the 1986 SIAM National Meeting, Boston, MA (July 21-25, 1986).
• 18: P.J. Burns and D. V. Pryor, "Vector and Parallel Monte Carlo Radiation Heat Transfer," to appear in Numerical Heat Transfer.
• 19: J. Maltby, "Vectofization of MONT3D," GS511 Final Project Report, Department of Mechanical Engineering, Colorado State University, Fort Collins, CO (May 1987).
• 20: O. M. Lubeck, Supercomputer performance: the theory, practice, and results, Advances in computers, Academic Press Professional, Inc., San Diego, CA, 1988
• 21: D. V. Pryor , P. J. Burns, Vectorized Monte Carlo molecular aerodynamics simulation of the Reyleigh problem, Proceedings of the 1988 ACM/IEEE conference on Supercomputing, p.384-391, November 12-17, 1988, Orlando, Florida, United States
• 22: P.I. Bums and D. V. Pryor, "Vector and Parallel Considerations for the Rayleigh Problem in Molecular Gas Dynamics," .Proceedings, 7th International Conference on Finite Element Methods in Flow Problems, Huntsville, AL (April 3-7, 1989).
• 23: D.B. Heifetz, "Veetorizing and Maerotasking Monte Carlo Neutral Particle Algorithms," Princeton Plasma Physics Labo:ratory, PPPL-2427 (April 1987).
• 24: G. M. Amdahl, "Validity of the Single Processor Approach to Achieving Large-scale Computing Capabilities," Proceedings of the American Federation of Information Processing Societies, Vol. 30, Washington, DC, pps. 483-485 (1967).
• 25: Control Data Corporation, CDC CYBER 200 Model 205 Computer System, Hardware Reference Manual, Control Data Corporation Publications and Graphics Division, Sunnyvale, California (1981 ).
• 26: H.J. Wasserman, "Los Alamos National Laboratory Computer Benckmarking 1988," Los Alamos National Laboratory Report LA- 11465-MS (1988).
• 27: R.W. Hockney and C. R. Jesshope, Parallel Computers 2, IOP Publishing, Ltd, Philadelphia, PA, (1988).
• 28: Control Data Corporation, CDC CYBER 200 Fortran Version 2 Reference Manual, Control Data Corporation Publications and Graphics Division, Sunnyvale, California (1981).
• 29: Cray Research, Inc., CRAY X-MP Computer Systems Four Processor Mainframe Reference Manual, Publication HR-0097, Mendota Heights, MN (1986).
• 30: Cray Research, inc., UNICOS Libraries, Macros and Opdefs Reference Manual, Publication SR-2013, Cray Research, Inc., Mendota Heights, MN (t987).
• 31: R. Kosketa, "Measurement of the Scientific Workload at Los Alamos National Laboratory Using the CRAY X-MP Hardware Performance Monitor," manuscript in preparation.
• 32: O.M. Lubeck and H. J. Wasserman, "Measurement of Vectorization Level in Fortran Programs," manuscript in preparation.

INDEX TERMS

Primary Classification:
  G. Mathematics of Computing
  G.3 PROBABILITY AND STATISTICS
      Subjects: Probabilistic algorithms (including Monte Carlo)

Additional Classification:
  C. Computer Systems Organization
  C.1 PROCESSOR ARCHITECTURES
      C.1.2 Multiple Data Stream Architectures (Multiprocessors)
          Subjects: Array and vector processors

  G. Mathematics of Computing
  G.1 NUMERICAL ANALYSIS
      G.1.0 General
          Subjects: Parallel algorithms

  J. Computer Applications
  J.2 PHYSICAL SCIENCES AND ENGINEERING
      Subjects: Physics

General Terms:
Algorithms, Design, Measurement, Performance