A NEW PARALLEL METHOD FOR MOLECULAR-DYNAMICS SIMULATION OF MACROMOLECULAR SYSTEMS

Short-range molecular dynamics simulations of molecular systems are co mmonly parallelized by replicated-data methods, in which each processo r stores a copy of all atom positions. This enables computation of bon ded 2-, 3-, and 4-body forces within the molecular topology to be part itioned among processors straightforwardly A drawback to such methods is that the interprocessor communication scales as N (the number of at oms) independent of P (the number of processors). Thus, their parallel efficiency falls off rapidly when large numbers of processors are use d. In this article a new parallel method for simulating macromolecular or small-molecule systems is presented, called force-decomposition. I ts memory and communication costs scale as N/root P, allowing larger p roblems to be run faster on greater numbers of processors. Like replic ated-data techniques, and in contrast to spatial-decomposition approac hes, the new method can be simply load balanced and performs well even for irregular simulation geometries. The implementation of the algori thm in a prototypical macromolecular simulation code ParBond is also d iscussed. On a 1024-processor Intel Paragon, ParBond runs a standard b enchmark simulation of solvated myoglobin with a parallel efficiency o f 61% and at 40 times the speed of a vectorized version of CHARMM runn ing on a single Gray Y-MP processor. (C) 1996 by John Wiley & Sons, In c.