Home // FUTURE COMPUTING 2012, The Fourth International Conference on Future Computational Technologies and Applications // View article

Balancing Communication and Execution Technique for Parallelized Sparse Matrix-Vector Multiplication

Authors:
Seiji Fujino
Takeshi Nanri
Ken'itiro Kusaba

Keywords: Sparse Matrix-Vector Multiplication, Load-Balancing, Bi_IDR(s) method

Abstract:
This paper proposes a runtime optimization technique for load balancing parallelized SpMxV (sparse matrix-vector multiplication) with consideration of cost for both communication and execution time. In state-of-the-art iterative methods that call SpMxV repeatedly during iterations, this optimization utilizes the estimated time of communication and the measured time of execution for adjustment of balance of load among processes. Through numerical experiments of an iterative solver for linear systems, it will make it clear that the proposed optimization technique outperforms compared with the conventional technique.

Pages: 28 to 31

Copyright: Copyright (c) IARIA, 2012

Publication date: July 22, 2012

Published in: conference

ISSN: 2308-3735

ISBN: 978-1-61208-217-2

Location: Nice, France

Dates: from July 22, 2012 to July 27, 2012