Computational Chemistry

• Description
• Projects
• References
• Codes

Description

The focus of this effort is on the integration of multipole-like algorithms for electrostatic forces in molecular dynamics simulations. In the project, two multipole-like algorithms and a multigrid algorithm will be integrated into the NAMD molecular dynamics package from the University of Illinois. Our effort consists in integrating a data parallel, non-adaptive three--dimensional code based on Dr. Chris Anderson's hierarchical O(N) N-body algorithm derived from computational elements based on Poisson's formula. DR. Alan McKenney is integrating a three--dimensional nonadaptive code based on Dr. Vladimir Rokhlin and Dr. Leslie Greengard's multipole algorithm, while Dr. Achi Brandt is involved in the multigrid version of NAMD with fast algorithms for electrostatic foces. The project is lead by Dr. Ruth Pachter of the Materials Laboratory at Wright-Patterson Air Force Base,

Projects

• Convert the existing non-adaptive Connection Machine Fortran code based on Anderson's three-dimensional sequential code to High Performance Fortran and integrate the code into the NAMD code. Carry out a performance analysis and comparison with alternate methods for a given accuracy.
• Convert the existing non-adaptive Fortran-90 code based on Anderson's three-dimensional sequential code to message passing codes based on
  • MPI and
  • Active Messages
  and integrate the codes into the NAMD code. Optimize the codes and carry out a performance analysis and comparison with alternate methods for a given accuracy.
• Integrate the adaptive HPF code based on Anderson's method currently being developed into the NAMD code, optimize the code and carry out a performance analysis and comparison with alternate methods for a given accuracy.
• Produce adaptive codes based on Anderson's method in
  • MPI and
  • Active Messages ,
  optimize and integrate the codes into the NAMD code, and carry out a performance analysis and comparison with alternate methods for a given accuracy.
• Carry out an error analysis and develop aposteriori and/or apriori error estimators and measure error as a function of particle distribution.

References

• Y. Charlie Hu, S. Lennart Johnsson and Shang-Hua Teng. A Data-Parallel Adaptive N-body Method (abstract). Invited talk at Parallel N-body solver minisymposia of the 8th SIAM Conference on Parallel Processing for Scientific Computing, Minneapolis, MN, March 1997.
  
  
• Yu Hu and S. Lennart Johnsson. Implementing O(N) N-body Algorithms Efficiently in Data-Parallel Languages . Journal of Scientific Programming, Vol. 5, pp. 000-000, 1996.
  
  
• Yu Hu and S. Lennart Johnsson. A Data-Parallel Implementation of Hierarchical N-body Methods. International Journal of Supercomputer Applications and High Performance Computing, 10(1): 3-40, 1996.
  
  
• Yu Hu and S. Lennart Johnsson. A Data-Parallel Implementation of O(N) Hierarchical N-body Methods. Supercomputing '96, Philadelphia, PA, November, 1996.
  
  
• Christopher R. Anderson, An implementation of the fast multipole method without multipoles, SIAM J. Sci. Stat. Comp., Vol. 13 no. 4, pp. 923-947, July 1992.
  
  
• Leslie Greengard and Vladimir Rokhlin. A Fast Algorithm for Particle Simulation. Journal of Computational Physics , Vol. 73, pp. 325-348, 1987.
  
  
• L. F. Greengard and V. Rokhlin. Rapid Evaluation of potential fields in three dimensions. Technical Report YALEU/DCS/RR-515, Dept. of Computer Science, Yale Univ. February 1987.
  
  
• L. Greengard and V. Rokhlin. On the efficient implementation of the fast multipole method. Tech. Rep. YALEU/DCS/RR-602, Dept. of Computer Science, Yale University, New Haven, CT, Feb. 1988.

• A. Brandt. Multilevel Computations: Review and Recent Developments. In S. McCormick, editor, Proc. Third Copper Mountain Conference on Multigrid Method. Copper Mountain, Colorado, April, 1987.