Parallel quadrant interlocking factorization on hypercube computers. (English) Zbl 0707.65012

García, I.; Merelo, J. J.; Bruguera, J. D.; Zapata, E. L.

Parallel Comput. 15, No. 1-3, 87-100 (1990).

Starting from the WZ factorization of matrices the quadrant interlocking factorization method (QIF) for solving systems of linear equations is presented in a brief summary. Next the authors discuss those aspects of hypercube computer architecture that are essential for their numerical application and present a general technique, how to decompose a sequential algorithm into parallel processes on a hypercube. They also introduce an extension of the computer language C for programming their hypercube. Applying those techniques they develop the parallel WZ and QIF algorithms and give their code listing. The paper concludes with a complexity analysis of the algorithm and a discussion of the effect of the hypercube dimension on efficiency.

Reviewer: C.H.Cap

MSC:

65F05	Direct numerical methods for linear systems and matrix inversion
65Y05	Parallel numerical computation
65Y20	Complexity and performance of numerical algorithms

Keywords:

parallel computation; WZ factorization; quadrant interlocking factorization method; hypercube computer architecture; computer language C; complexity; efficiency