Hybrid MPI/Pthread implementation of 1-D FFT on SMP

Author

Na´mneh, R. AL ; Pan, W.D.

Author_Institution

Dept. of Electr. & Comput. Eng., Alabama Univ., Huntsville, AL

fYear

2006

Firstpage

114

Lastpage

117

Abstract

SMP (shared memory multiprocessors) can support two programming models: message passing and shared address space. Shard address space offers easy programming, particularly for irregularly structured computations, but suffers from performance limitation due to the protocol overhead and poor spatial locality (false sharing). On the other hand, it has been shown that there are performance advantages in integrating message passing in cache coherent multiprocessors. In this paper, we propose the hybrid MPI/Pthread implementation of 1-D FFT on SMP. We compare the results of hybrid MPI/Pthread implementation with pure MPI implementation. The results show that MPI/Pthread implementation perform better than pure MPI implementation for large data sizes

Keywords

fast Fourier transforms; message passing; multi-threading; shared memory systems; 1D FFT; MPI; Pthread implementation; message passing; shared address space; shared memory multiprocessors; Computational modeling; Concurrent computing; Fast Fourier transforms; Memory architecture; Message passing; Partitioning algorithms; Plasma simulation; Signal processing algorithms; Telecommunication computing; Yarn;

fLanguage

English

Publisher

ieee

Conference_Titel

System Theory, 2006. SSST '06. Proceeding of the Thirty-Eighth Southeastern Symposium on

Conference_Location

Cookeville, TN

Print_ISBN

0-7803-9457-7

Type

conf

DOI

10.1109/SSST.2006.1619106

Filename

1619106