Solving problems on concurrent processors
Solving problems on concurrent processors
A fast method for solving a class of tridiagonal linear systems
Communications of the ACM
Hi-index | 0.00 |
In this paper, we present a parallelized implementation of deflection yoke simulation using the method of moment to analyze characteristic properties of magnetic fields generated by components of deflection yoke. Seeking an optimal configuration of deflection yoke simulation necessitates a reduction of the execution time of analysis. The paper also describes a revised parallel LU decomposition and discusses a parallelization scheme used to simulate trajectories of 24 different electronic beams through a Cathod Ray Tube. The empirical results show the effect of Amdahl's Law caused by unparallelized parts, giving 20 as the upper limit of speedup. The parallelized deflection yoke simulation program running on 18 Intel Paragon processors executes about twice as fast as the original serial program running on a Cray-YMP/4E.