ePubs

The open archive for STFC research publications

Full Record Details

DOI 10.5286/raltr.2017010
Persistent URL http://purl.org/net/epubs/work/34746569
Record Status Checked
Record Id 34746569
Title Sparse direct solution on parallel computers
Contributors
Abstract We describe our recent work on designing algorithms and software for solving sparse systems using direct methods on parallel computers. This work has been conducted within an EU Horizon 2020 Project called NLAFET. We first discuss the solution of large sparse symmetric positive definite systems. We use a runtime system to express and execute a DAG-based Cholesky factorization. The runtime system plays the role of a software layer between the application and the architecture and handles the management of task dependencies as well as task scheduling and maintaining data coherency. Although runtime systems are widely used in dense linear algebra, this approach is challenging for sparse algorithms because of the irregularity and variable granularity of the DAGs arising in these systems. We have implemented our software using the OpenMP standard and the runtime systems StarPU and PaRSEC. We compare these implementations to HSL MA87, a state-of-the-art DAG-based solver for positive definite systems. We demonstrate comparable performance on a multicore architecture. We also consider the case when thematrix is symmetric indefinite. For highly unsymmetric systems we use a completely different approachbased on developing a parallel version of a Markowitz threshold ordering. This work is less advancedbut we discuss some of the algorithmic challenges involved. Finally, we briefly discuss using a hybriddirect-iterative solver that combines the best of the two approaches and enables the solution of even larger problems in parallel.
Organisation STFC , SCI-COMP , SCI-COMP-CM
Keywords sparse Cholesky , symmetric indefinite , block Cimmino , hybrid methods , runtime systems , highly unsymmetric matrices
Funding Information EU, Horizon 2020/NLAFET (671633)
Related Research Object(s):
Licence Information: Creative Commons Attribution 3.0 Unported (CC BY 3.0)
Language English (EN)
Type Details URI(s) Local file(s) Year
Report RAL Technical Reports RAL-TR-2017-010. 2017. RAL-TR-2017-010.pdf 2017