The open archive for STFC research publications

Full Record Details

Persistent URL http://purl.org/net/epubs/work/29771
Record Status Checked
Record Id 29771
Title Task scheduling in an asynchronous distributed memory multifrontal solver
Abstract We describe the improvements to the task scheduling for MUMPS, an asynchronous distributed memory direct solver for sparse linear systems. In the new approach, we determine, during the analysis of the matrix, candidate processes for the tasks that will be dynamically scheduled during the subsequent factorization. This approach significantly improves the scalability of the solver in tems of execution time and storage. By comparison with the previous version of MUMPS, we demonstrate the efficiency and the scalability of the new algorithm on up to 512 processors. Our test cases include matrices from regular 3D grids and irregular ones from real-life applications.
Organisation CCLRC
Funding Information
Related Research Object(s):
Licence Information:
Language English (EN)
Type Details URI(s) Local file(s) Year
Report RAL Technical Reports RAL-TR-2002-028. 2002. raltr-2002028.pdf 2002
Journal Article SIAM J Matrix Anal A 26, no. 2 (2005): 544-565. doi:10.1137/S0895479802419877 2005
Report TR/PA/02/105. 2002