Efficient and accurate FPGA-based simulator for Molecular Dynamics

A molecular dynamics (MD) system is defined by the position and momentum of particles and their interactions. Solving the dynamics numerically and evaluating the interaction is computationally expensive even for a small number of particles in the system. We are focusing on long-ranged interactions, since the calculation time is O(N2) for an N particle system. There are many existing algorithms aimed at reducing the calculation time of MD simulations. Among the existing algorithms, multigrid (MG) method [1] reduces O(N2) calculation time to O(N) time while still achieving reasonable accuracy. Another movement to achieve much faster calculation time is running MD simulation on special purpose processors and customized hardware with ASICs or an FPGAs. In this paper, we design and implement FPGA-based MD simulator with an efficient MG method.

[1]  John A. Board,et al.  A portable distributed implementation of the parallel multipole tree algorithm , 1995, Proceedings of the Fourth IEEE International Symposium on High Performance Distributed Computing.

[2]  Eunjung Cho,et al.  An FPGA Design to Achieve Fast and Accurate Results for Molecular Dynamics Simulations , 2007, ISPA.

[3]  C. Sagui,et al.  Multigrid methods for classical molecular dynamics simulations of biomolecules , 2001 .

[4]  T. Darden,et al.  Molecular dynamics simulations of biomolecules: long-range electrostatic effects. , 1999, Annual review of biophysics and biomolecular structure.

[5]  Paul Chow,et al.  Reconfigurable molecular dynamics simulator , 2004, 12th Annual IEEE Symposium on Field-Programmable Custom Computing Machines.

[6]  Takashi Amisaki,et al.  Development of MD Engine: High‐speed accelerator with parallel processor design for molecular dynamics simulations , 1999 .

[7]  Robert D. Skeel,et al.  Multiple grid methods for classical molecular dynamics , 2002, J. Comput. Chem..

[8]  Thierry Matthey,et al.  ProtoMol, an object-oriented framework for prototyping novel algorithms for molecular dynamics , 2004, TOMS.

[9]  Martin C. Herbordt,et al.  Accelerating molecular dynamics simulations with configurable circuits , 2005, International Conference on Field Programmable Logic and Applications, 2005..

[10]  T Fukushige,et al.  A high performance system for molecular dynamics simulation of biomolecules using a special-purpose computer. , 1996, Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing.

[11]  Nobuaki Miyakawa,et al.  Development of MD Engine: High-speed accelerator with parallel processor design for molecular dynamics simulations , 1999, J. Comput. Chem..

[12]  Makoto Taiji,et al.  Fast and accurate molecular dynamics simulation of a protein using a special-purpose computer , 1997, J. Comput. Chem..

[13]  C. Brooks.  Computer simulation of liquids , 1989 .

[14]  Sanjay Banerjee,et al.  Efficient charge assignment and back interpolation in multigrid methods for molecular dynamics , 2005, J. Comput. Chem..

[15]  TAKASHI AMISAKI,et al.  Error evaluation in the design of a special‐purpose processor that calculates nonbonded forces in molecular dynamics simulations , 1995, J. Comput. Chem..

[16]  J. Board,et al.  Ewald summation techniques in perspective: a survey , 1996 .

[17]  M.G.B. Drew,et al.  The art of molecular dynamics simulation , 1996 .

[18]  S. W. Leeuw,et al.  An Iterative PPPM Method for Simulating Coulombic Systems on Distributed Memory Parallel Computers , 1998 .