Title :
Parallelizing molecular dynamics programs for distributed-memory machines
Author :
Hwang, Yuan-Shin ; Das, Raja ; Saltz, Joel H. ; Hodoscek, M. ; Brooks, Bernard R.
Author_Institution :
Maryland Univ., College Park, MD, USA
Abstract :
Using a data-parallel strategy to parallelize molecular dynamics programs makes it possible to simulate very large molecules on large numbers of processors. Applying on appropriate combination of data partitioning and iteration distribution algorithms not only optimizes communication overheads but also achieves good load balance, and hence good performance
Keywords :
digital simulation; distributed memory systems; iterative methods; molecular dynamics method; parallel programming; physics computing; resource allocation; communication overhead optimization; data partitioning; data-parallel strategy; distributed-memory machines; iteration distribution algorithms; load balance; molecular dynamics program parallelization; performance; Algorithm design and analysis; Atomic measurements; Chaotic communication; Concurrent computing; Educational institutions; Libraries; Matrix decomposition; Multiprocessing systems; Partitioning algorithms; Sparse matrices;
Journal_Title :
Computational Science & Engineering, IEEE
