A Dense Out-of-Core Solver (DOCS) for Complex-Valued Linear Systems
نویسنده
چکیده
Dense systems of linear equations are quite common in many science and engineering applications. Such linear systems place extreme storage and computational demands on computer resources and, in many cases, may severely limit the subsequent analysis. A dense out-of-core solver (DOCS) that operates on a partitioned coefficient matrix can reduce the in-core storage requirements of the linear system while spreading the associated computational burden over multiple processors (which reduces run time as well). In this report, I describe a DOCS that operates on a partitioned coefficient matrix that may be distributed over multiple external storage devices. I have implemented this solver using Message-Passing Interface (MPI) protocols. This report presents performance data from a series of run time trials that compare the run time of both sequential and parallel implementations of the DOCS.
منابع مشابه
Anatomy of a Parallel Out-of-Core Dense Linear Solver
{ In this paper, we describe the design and implementation of the Platform Independent Parallel Solver (PIPSolver) package for the out-of-core (OOC) solution of complex dense linear systems. Our approach is unique in that it allows essentially all of RAM to be lled with the current portion of the matrix (slab) to be updated and fac-tored, thereby greatly improving the computation to I/O ratio o...
متن کاملAn efficient distributed randomized solver with application to large dense linear systems
Randomized algorithms are gaining ground in high performance computing applications as they have the potential to outperform deterministic methods, while still providing accurate results. In this paper, we propose a randomized algorithm for distributed multicore architectures to efficiently solve large dense symmetric indefinite linear systems that are encountered, for instance, in parameter es...
متن کاملImplementing Complex Krylov Solvers with Minimal Pain
Most algorithms used in preconditioned iterative methods are generally applicable to complex valued linear systems, with real valued linear systems simply being a special case. However, most iterative solver packages available today focus exclusively on real valued systems, or deal with complex valued systems as an afterthought. One obvious approach to addressing this problem is to recast the c...
متن کاملINTERSECTION OF ESSENTIAL IDEALS IN THE RING OF REAL-VALUED CONTINUOUS FUNCTIONS ON A FRAME
A frame $L$ is called {it coz-dense} if $Sigma_{coz(alpha)}=emptyset$ implies $alpha=mathbf 0$. Let $mathcal RL$ be the ring of real-valued continuous functions on a coz-dense and completely regular frame $L$. We present a description of the socle of the ring $mathcal RL$ based on minimal ideals of $mathcal RL$ and zero sets in pointfree topology. We show that socle of $mathcal RL$ is an essent...
متن کاملAn efficient out-of-core multifrontal solver for large-scale unsymmetric element problems
In many applications where the efficient solution of large sparse linear systems of equations is required, a direct method is frequently the method of choice. Unfortunately, direct methods have a potentially severe limitation: as the problem size grows, the memory needed generally increases rapidly. However, the in-core memory requirements can be limited by storing the matrix and its factors ex...
متن کامل