J. Parris et J. Padovan, HIERARCHICALLY PARTITIONED SOLUTION STRATEGY FOR CFD APPLICATIONS .1.THEORY, Computer methods in applied mechanics and engineering, 126(3-4), 1995, pp. 197-222
This two-part paper presents the results of a benchmarked analytical-n
umerical investigation into the operational characteristics of a unifi
ed parallel processing strategy for implicit fluid mechanics formulati
ons. This Hierarchical Poly Tree (HPT) strategy is based on multilevel
substructural decomposition. The Tree morphology is chosen to minimiz
e memory, communications and computational effort. The methodology is
general enough to apply to existing finite difference (FD), finite ele
ment (FEM), finite volume (FV) or spectral element (SE) based computer
programs without an extensive rewrite of code. In addition to finding
large reductions in memory, communications and computational effort a
ssociated with a parallel computing environment, substantial reduction
s are generated in the sequential mode of application. Such improvemen
ts grow with increasing problem size. Along with a theoretical develop
ment of general 2-D and 3-D HPT, several techniques for expanding the
problem size that the current generation of computers are capable of s
olving, are presented and discussed. Among these techniques are severa
l interpolative reduction methods. It was found that by combining seve
ral of these techniques that a relatively small interpolative reductio
n resulted in substantial performance gains. Several other unique feat
ures/benefits are discussed in this paper. Along with Part I's theoret
ical development, Part II presents a numerical approach to the HPT alo
ng with four prototype CFD Applications. These demonstrate the potenti
al of the HPT strategy.