Strategies and performance norms for efficient utilization of vector pipeline computers as illustrated by the classical mechanical simulation of rotationally inelastic collisions (Q581007)

We discuss the vectorization of a classical mechanical trajectory code which simulates the interaction of a rotationally excited rigid rotator with a colliding atom. This code was optimized in FORTRAN 77 and can be run on either a VAX 11/780 minicomputer or a CRAY-1 vector pipeline supercomputer. The article presents a global strategy for vectorizing a classical mechanical trajectory code, a set of performance citeria for characterizing the vectorization of a computer code, and an analysis of the four major subroutines of the vectorized trajectory code in terms of this strategy and these criteria.