Abstract
A new matrix multiplication algorithm is proposed for massively parallel supercomputers with 2D/3D, all-port torus interconnection networks. The proposed algorithm is based on the traditional row-by-column multiplication matrix product model and employs a special routing pattern for better scalability. It compares favorably to the variants of Cannon’s and DNS algorithms since it allows matrices of the same size to be multiplied on a higher number of processors due to lower data communications overhead.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
A review of these methods can be found in Chap. 47, Handbook of Linear Algebra [1].
- 2.
Throughout the paper, we assume that the matrix multiplication operation is to be performed such that this initial mapping is preserved at the end of the operation.
- 3.
For a good intro to routing in general and wormhole routing in particular, see [5].
References
Bini DA (2007) Fast matrix multiplication. In: Hogben L (ed) Handbook of linear algebra. Chapman & Hall/CRC press, Boca Raton (Chap. 47)
Cannon LE (1969) A cellular computer to implement the kalman filter algorithm. Ph. D. Thesis, Montana State University
Grama A, Gupta A, Karypis G, Kumar V (2003) Introduction to parallel computing, 2nd edn. Addison Wesle, Eugene
Imre KM, Baransel C, Artuner H (2010) Efficient and scalable routing algorithms for collective communication operations on 2D All-Port Torus networks. Int J Parallel Progr 39(6):746–782
Ni LM, McKinley PK (1993, February) A survey of wormhole routing techniques in direct networks. Computer 26(2):62–76
Strassen V (1969) Gaussian elimination is not optimal. Numer Math 13:354–356
Winograd S (1971) On multiplication of 2 × 2 matrices. Linear Algebra Appl 4:381–388
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer Science+Business Media Dordrecht
About this paper
Cite this paper
Baransel, C., Imre, K., Artuner, H. (2014). A New Parallel Matrix Multiplication Algorithm for Wormhole-Routed All-Port 2D/3D Torus Networks. In: Fonseca Ferreira, N., Tenreiro Machado, J. (eds) Mathematical Methods in Engineering. Springer, Dordrecht. https://doi.org/10.1007/978-94-007-7183-3_2
Download citation
DOI: https://doi.org/10.1007/978-94-007-7183-3_2
Published:
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-007-7182-6
Online ISBN: 978-94-007-7183-3
eBook Packages: EngineeringEngineering (R0)