DocumentCode :
2495605
Title :
A parallel graph partitioner on a distributed memory multiprocessor
Author :
Buch, Premal ; Sanghavi, Jagesh ; Sangiovanni-Vincentelli, Alberto
Author_Institution :
California Univ., Berkeley, CA, USA
fYear :
1995
fDate :
6-9 Feb 1995
Firstpage :
360
Lastpage :
366
Abstract :
In order to realize the full potential of speed-up by parallelization, it is essential to partition a problem into small tasks with minimal interactions without making this process itself a bottleneck. We present a method for graph partitioning that is suitable for parallel implementation and scales well with the number of processors and the problem size. Our algorithm uses hierarchical partitioning. It exploits the parallel resources to minimize the dependence on the starting point with multiple starts at the higher levels of the hierarchy. These decrease at the lower levels as it zeroes in on the final partitioning. This is followed by a last-gasp phase that randomly collapses partitions and repartitions to further improve the quality of the fmat solution. Each individual 2-way partitioning step can be performed by any standard partitioning algorithm. Results are presented on a set of benchmarks representing connectivity graphs of device and circuit simulation problems
Keywords :
computational complexity; graph theory; message passing; parallel algorithms; benchmarks; connectivity graphs; distributed memory multiprocessor; graph partitioning; hierarchical partitioning; parallel graph partitioner; partitioning algorithm; Central Processing Unit; Circuit simulation; Concurrent computing; Costs; Load management; Message passing; Parallel algorithms; Parallel processing; Partitioning algorithms; Simulated annealing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Frontiers of Massively Parallel Computation, 1995. Proceedings. Frontiers '95., Fifth Symposium on the
Conference_Location :
McLean, VA
Print_ISBN :
0-8186-6965-9
Type :
conf
DOI :
10.1109/FMPC.1995.380434
Filename :
380434
Link To Document :
بازگشت