We propose the Ensparsed LU decomposition (ELU) method as a linear solver for large-scale circuit transient analysis. Ensparsing is an algorithmic technique of ignoring small values in computation. While some researchers have used ensparsing by value in linear solvers for circuit simulation, the ELU method incorporates ensparsing by time as well. A high performance implementation of the ELU method for transient analysis is also investigated. The ELU method is faster than the conventional linear solvers, and the advantages of the ELU method will still increase for larger circuits.

This paper investigates an efficient implementation of circuit simulator sparta based on the preconditioned Jacobi method on a loosely coupled distributed memory machine Fujitsu AP1000. The preconditioned relaxation methods are linear solvers effective in large scale circuit simulations. Because the preconditioned Jacobi method has a high parallelism, the only problem of parallelization of sparta is the application of the preconditioner to the residual vector. This paper investigates the matrix mapping schemes and the communication methods for efficient preconditioner application. The peculiarity of the preconditioner in sparta is a few full rows and columns, in spite of its rather high sparsity. The column-row mapping is the best scheme by far, where the decisive point is the low communication requirements. The best communication method is the far-first cascade method, despite the out-of-order reception of messages of the one-to-one method. The parallel efficiency of sparta is reporte...

A multigrid preconditioned conjugate gradient (MGCG) method, which uses the multigrid method as a preconditioner for the conjugate gradient method, has a good convergence rate even for problems on which the standard multigrid method does not converge efficiently. This paper considers a parallelization of the MGCG method and proposes an efficient parallel MGCG method on distributed memory machines. For a good convergence rate of the MGCG method, several difficulties in parallelizing the multigrid method are successfully settled. It is also shown that the parallel MGCG method has high performance on the Fujitsu AP1000 multicomputer, and it is more than 10 times faster than the scaled conjugate gradient (SCG) method.>

Article Free Access Share on QCDPAX-an MIMD array of vector processors for the numerical simulation of quantum chromodynamics Authors: T. Shirakawa Institute of Engineering Mechanics Institute of Engineering MechanicsView Profile , T. Hoshino Institute of Engineering Mechanics Institute of Engineering MechanicsView Profile , Y. Oyanagi Institute of Information Sciences Institute of Information SciencesView Profile , Y. Iwasaki Institute of PiIysics, University of Tsukuba, Tsukuba, Ibaraki 305, JAPAN Institute of PiIysics, University of Tsukuba, Tsukuba, Ibaraki 305, JAPANView Profile , T. Yoshie Institute of PiIysics, University of Tsukuba, Tsukuba, Ibaraki 305, JAPAN Institute of PiIysics, University of Tsukuba, Tsukuba, Ibaraki 305, JAPANView Profile Authors Info & Claims Supercomputing '89: Proceedings of the 1989 ACM/IEEE conference on SupercomputingAugust 1989 Pages 495–504https://doi.org/10.1145/76263.76318Published:01 August 1989Publication History 9citation168DownloadsMetricsTotal Citations9Total Downloads168Last 12 Months4Last 6 weeks1 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my Alerts New Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteeReaderPDF