×

Performance issues for frontal schemes on a cache-based high-performance computer. (English) Zbl 0909.65019

An implementation on a cache-based high-performance computer of a frontal code for the solution of large sparse unsymmetric linear systems is discussed. A modification of the frontal solution algorithm is proposed in order to enhance the proportion of floating-point operations performed using Level 3 BLAS. After a short description of frontal methods in Section 1, Section 2 discusses the features of the frontal scheme. Section 3 presents the proposed modification. Section 4 debates the effect of a cache and indicates the effect of data reuse by looking at the performance results on a multiprocessor machine. Section 5 illustrates the effects of exploiting Level 3 BLAS in the frontal solver through numerical experiments using practical problems.

MSC:

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

Software:

NAMMU; HSL; MA42
PDF BibTeX XML Cite
Full Text: DOI

References:

[1] Irons, Int. J. Numer. Meth. Engng. 2 pp 5– (1970) · Zbl 0252.73050
[2] Hood, Int. J. Numer. Meth. Engng. 10 pp 379– (1976) · Zbl 0322.65013
[3] Duff, Parallel Comput. 5 pp 55– (1984)
[4] Duff, ACM Trans. Math. Software 22 pp 30– (1996) · Zbl 0884.65018
[5] Harwell Subroutine Library, A Catalogue of Subroutines (Release 12), AEA Technology, Harwell Laboratory, Harwell, U.K., 1996.
[6] and , ’A comparison of frontal software with other sparse direct solvers’, Technical Report RAL-96-102 (revised), Rutherford Appleton Laboratory, 1996.
[7] Dongarra, ACM Trans. Math. Software 16 pp 1– (1990) · Zbl 0900.65115
[8] , and , Solving Linear Systems on Vector and Shared Memory Computers, SIAM, Philadelphia, PA, 1991.
[9] Duff, Int. J. Numer. Meth. Engng. 28 pp 2555– (1989) · Zbl 0725.65045
[10] and , ’MA42–a new frontal code for solving sparse unsymmetric systems’, Technical Report RAL-93-064, Rutherford Appleton Laboratory, 1993.
[11] Zitney, Comput. Chem. Engng. 17 pp 319– (1993)
[12] Dave, SIAM J. Sci. Statist. Comput. 5 pp 270– (1987)
[13] and , ’NAMMU (Release 6.3) User Guide’, Technical Report AEA-ES-0138, AEA Technology, 1996.
[14] Zitney, Comput. Chem. Engng. 20 pp 641– (1996)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.