Distribution and Scheduling Strategies for Parallel Sparse System Solvers in Device Simulation Applications

We discuss strategies for partitioning data and scheduling computation and communication suitable for the parallel solution of sparse triangular systems arising in the iterative solution of non-linear PDEs. For partitioning, we consider recursive spectral bisection (RSB) and its modification, RSB with vertex separator. To extract available parallelism and to minimize the effects of communication delays, we consider five different schemes for ordering the variables within each partition. To further minimize the overheads, we use a micro scheduler that preschedules the computation and communication steps for each triangular solve by taking into account both the problem specific and architecture specific parameters. We present performance results of a solver used in a general purpose device simulation code called Fielday on a 32-processor IBM SP2.

By: Vladimir Menkov and Vijay K. Naik

Published in: RC20667 in 1996

LIMITED DISTRIBUTION NOTICE:

This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.

8467.ps.gz

Questions about this service can be mailed to reports@us.ibm.com .