An Overview of the Carp-cg Algorithm
نویسندگان
چکیده
Recent work shows that the block-parallel CARP-CG algorithm [Parallel Computing 36, 2010] is extremely effective on sparse nonsymmetric linear systems with very small diagonal elements, including cases with discontinuous coefficients. In contrast to most known solvers, the effectiveness of CARP-CG often improves as the diagonal elements become smaller. This property is shown to follow from the foundations on which CARP-CG is based. The unique behavior of CARP-CG is demonstrated with some old and new results. Previously studied problems consist of convection-dominated elliptic PDEs, with and without discontinuous coefficients. New results using high-frequency Helmholtz equations in the heterogeneous Marmousi model, discretized with 2nd, 4th and 6th order finite difference schemes, indicate the insufficiency of low-order schemes and the good parallel scalability of CARP-CG at high frequencies.
منابع مشابه
CARP-CG: A robust and efficient parallel solver for linear systems, applied to strongly convection dominated PDEs
CARP-CG is a conjugate gradient (CG) acceleration of CARP, which was introduced by Gordon and Gordon as a robust block-parallel scheme for sparse linear systems. CARP performs Kaczmarz (KACZ) row projections within the blocks, and the results from the separate blocks are merged by averaging, for each component, its updated values from the different blocks. The averaging operations are equivalen...
متن کاملA robust and efficient, highly scalable parallel solution of the Helmholtz equation with large wave numbers
Numerical solution of the Helmholtz equation is a challenging computational task, particularly when the wave number is large. Recent years have seen great improvements in the finite difference approach to this problem through enhancements of the shifted Laplacian preconditioner, originally introduced in [1]. For a recent survey and some new results, see [4]. In some cases, this approach may be ...
متن کاملRobust and highly scalable parallel solution of the Helmholtz equation with large wave numbers
Numerical solution of the Helmholtz equation is a challenging computational task, particularly when the wave number is large. For two-dimensional problems, direct methods provide satisfactory solutions, but large three-dimensional problems become unmanageable. In this work, the block-parallel CARP-CG algorithm [Parallel Computing 36, 2010] is applied to the Helmholtz equation with large wave nu...
متن کاملGPU computing with Kaczmarz's and other iterative algorithms for linear systems
The graphics processing unit (GPU) is used to solve large linear systems derived from partial differential equations. The differential equations studied are strongly convection-dominated, of various sizes, and common to many fields, including computational fluid dynamics, heat transfer, and structural mechanics. The paper presents comparisons between GPU and CPU implementations of several well-...
متن کاملPlanning in a cross dock network with an operational scheduling overview
Nowadays, cross docking plays an important role in the supply chain networks especially in transportation systems. According to the cross dock system advantages such as reducing transportation costs, lead times, and inventories, scheduling in a cross-dock center would be more complicated by increasing the number of suppliers, customers and product types. Considering the cross dock limited capac...
متن کامل