Time-partitioning simulation models for calculation of parallel computersA technique allowing time-staggered solution of partial differential equations is presented in this report. Using this technique, called time-partitioning, simulation execution speedup is proportional to the number of processors used because all processors operate simultaneously, with each updating of the solution grid at a different time point. The technique is limited by neither the number of processors available nor by the dimension of the solution grid. Time-partitioning was used to obtain the flow pattern through a cascade of airfoils, modeled by the Euler partial differential equations. An execution speedup factor of 1.77 was achieved using a two processor Cray X-MP/24 computer.
Document ID
19880059734
Acquisition Source
Legacy CDMS
Document Type
Conference Paper
Authors
Milner, Edward J. (NASA Lewis Research Center Cleveland, OH, United States)
Blech, Richard A. (NASA Lewis Research Center Cleveland, OH, United States)
Chima, Rodrick V. (NASA Lewis Research Center Cleveland, OH, United States)