scientific article; zbMATH DE number 4064411
coarse grid correctionComplexity estimatesFourier analysishypercubesidle processor problemimplementationmultigrid algorithms
Laplace operator, Helmholtz equation (reduced wave equation), Poisson equation (35J05) Iterative numerical methods for linear systems (65F10) Numerical solution of discretized equations for boundary value problems involving PDEs (65N22) Parallel numerical computation (65Y05) Analysis of algorithms and problem complexity (68Q25)
[For the entire collection see Zbl 0641.00031.] The implementation of multigrid algorithms for solving model partial differential equations on hypercubes is investigated. It is shown that one can design and map on the hypercube asymptotically optimal (with respect to the execution time) multigrid algorithms in both cases if there are many grid points per processor and if there is only one grid point per processor. Complexity estimates confirmed by experimental results on an Intel iPSC system and allowing us to predict the execution time are presented. To avoid the ``idle processor problem the authors proposed a new interesting modification of the multigrid algorithm. The basic idea of this modification consists in splitting the residual equation (before doing the coarse grid correction) appropriately into some auxiliary equations in order to keep the processors busy. The Fourier analysis of the model 1D-problem \(-u=f\) shows that this technique additionally improves the two-grid rate in comparison with the standard two-grid method.
- A multigrid algorithm for parallel computers: CPMG
- VML: a class of virtual multi-level iterative methods for solving partial differential equations
- A multigrid method for multiprocessors
- Multigrid solvers in reconfigurable hardware
- Multigrid on massively parallel architectures
- Implementing the multigrid method on a massively parallel computer
- Parallel multigrid technique: reduction to independent problems
- TOTALLY PARALLEL MULTILEVEL ALGORITHMS
- A MODEL AND IMPLEMENTATION OF MULTIGRID FOR MASSIVELY PARALLEL COMPUTERS
- scientific article; zbMATH DE number 3919773 (Why is no real title available?)
- scientific article; zbMATH DE number 3949987 (Why is no real title available?)
- Parallel Networks for Multi-Grid Algorithms: Architecture and Complexity
- scientific article; zbMATH DE number 4033076 (Why is no real title available?)
- scientific article; zbMATH DE number 4064410 (Why is no real title available?)
- A Grid-Based Subtree-Subcube Assignment Strategy for Solving Partial Differential Equations on Hypercubes
- Semicoarsening Multigrid on a Hypercube
- scientific article; zbMATH DE number 4001393 (Why is no real title available?)
- scientific article; zbMATH DE number 2226864 (Why is no real title available?)
- Multiplicative and additive parallel multigrid algorithms for the acceleration of compressible flow computations on unstructured meshes.
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3798178)