EVALUATING POLARIZABLE POTENTIALS ON DISTRIBUTED-MEMORY PARALLEL COMPUTERS - PROGRAM-DEVELOPMENT AND APPLICATIONS

The efficient evaluation of polarizable molecular mechanics potentials on distributed memory parallel computers is discussed. The program ex ecutes at 7-10 Mflops/node on a 32-node CM-5 partition and is 19 times faster than comparable code running on a single-processor HP 9000/735 . On the parallel computer, matrix inversion becomes a practical alter native to the commonly used iterative method for the calculation of in duced dipole moments. The former method is useful in cases such as fre e-energy perturbation (FEP) simulations, which require highly accurate induced dipole moments. Matrix inversion is performed 110 times faste r on the CM-5 than on the HP. We show that the accuracy which is neede d for FEP calculations with polarization can be obtained by either mat rix inversion or by performing a large number of iteration cycles to s atisfy convergence tolerances that are less than 10(-6) D. (C) 1995 by John Wiley & Sons, Inc.