Abstract:SUMMARYThe Peaceman-Rachford AD1 technique is a well-known iterative method for solving large systems of linear equations arising from finite difference methods of solution of partial differential equations. The present note demonstrates how the algorithm can be conveniently rearranged for computation in a virtual storage system with paging. Timing results are given for a sample problem typical of those encountered in nuclear diffusion computation.
“…He ignores the 1960 paper by Barron and Swinnerton-Dyerm [4] that showed a partitioned schedule that accesses entire columns and does not require a block layout. Papers describing similar schedules for other problems were published in the following years [18,19,59].…”
Section: Dense Matrix Computationsmentioning
confidence: 99%
“…Schedules that access data mostly sequentially have been developed for virtual memory systems, however. Wang, for example, shows such a schedule for alternating directions [59]. Another idea that has emerged recently is to modify the data flow of iterative algorithms that cannot be scheduled for data reuse so that they can be scheduled.…”
Abstract. This paper surveys algorithms that efficiently solve linear equations or compute eigenvalues even when the matrices involved are too large to fit in the main memory of the computer and must be stored on disks. The paper focuses on scheduling techniques that result in mostly sequential data accesses and in data reuse, and on techniques for transforming algorithms that cannot be effectively scheduled. The survey covers out-of-core algorithms for solving dense systems of linear equations, for the direct and iterative solution of sparse systems, for computing eigenvalues, for fast Fourier transforms, and for N-body computations. The paper also discusses reasonable assumptions on memory size, approaches for the analysis of out-of-core algorithms, and relationships between out-of-core, cache-aware, and parallel algorithms.
“…He ignores the 1960 paper by Barron and Swinnerton-Dyerm [4] that showed a partitioned schedule that accesses entire columns and does not require a block layout. Papers describing similar schedules for other problems were published in the following years [18,19,59].…”
Section: Dense Matrix Computationsmentioning
confidence: 99%
“…Schedules that access data mostly sequentially have been developed for virtual memory systems, however. Wang, for example, shows such a schedule for alternating directions [59]. Another idea that has emerged recently is to modify the data flow of iterative algorithms that cannot be scheduled for data reuse so that they can be scheduled.…”
Abstract. This paper surveys algorithms that efficiently solve linear equations or compute eigenvalues even when the matrices involved are too large to fit in the main memory of the computer and must be stored on disks. The paper focuses on scheduling techniques that result in mostly sequential data accesses and in data reuse, and on techniques for transforming algorithms that cannot be effectively scheduled. The survey covers out-of-core algorithms for solving dense systems of linear equations, for the direct and iterative solution of sparse systems, for computing eigenvalues, for fast Fourier transforms, and for N-body computations. The paper also discusses reasonable assumptions on memory size, approaches for the analysis of out-of-core algorithms, and relationships between out-of-core, cache-aware, and parallel algorithms.
A B S T R A C TMUFTI, I.R. 1985, Seismic Modeling in the Implicit Mode, Geophysical Prospecting 33, Finite-difference seismic models are often saddled with huge memory requirements for data manipulation, a prohibitive amount of CPU time and even approximate results. At least part of these costs may be due to the fact that most of the work reported on this subject is devoted to the development of explicit models which suffer from severe limitations of stability and necessitate extremely fine time sampling of the wavefield.A new method of seismic modeling which works in the implicit mode and is unconditionally stable is put forward. It is based on the self-adjoint version of the acoustic wave equation. The evaluation of the wavefield is done by using a highly efficient splitting algorithm which does not require transposing the field data at the various time steps. Moreover, it can accommodate anisotropic media as well as three-dimensional structures. Computational efficiency is achieved by introducing an unconventional procedure which yields the sum of the values of the wavefield corresponding to a new time step and a previous time step. The new value can be obtained from this sum by a simple subtraction. I 2 1 c J J t I
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.