Abstract

The paper proposes a technique for pipelining a computational process in the spatial decomposition of the computational domain for constructing efficient parallel algorithms for numerical solution of hydrodynamic problems oriented to cluster computing systems. Methods for achieving high efficiency of a parallel application based on a finite-difference explicit or semi-explicit numerical scheme are shown. An expression is written to determine the minimum size of the calculated area, at which the efficiency of the parallel program approaches unity.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call