Implementation/acceleration of vectorized acoustic source localization algorithms on the CELL multi-core processor

Neena Imam, Jacob Barhen, Travis Humble

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper describes the optimized design and implementation of sonar signal processing algorithms (matched filter) on the CELL multi-core processor. The algorithm is modified to achieve maximum parallelism to enhance time performance on the multi-core platform. In this proof-of-concept effort, we show a factor of 10 speedup by parallelizing the matched filter code on 16 separate computing nodes on the CELL processor for broadband matched filter implementation. This demonstration of considerably faster signal processing capability should be of substantial interest to the signal processing community in general, and the oceanic engineering community in particular.

Original languageEnglish
Title of host publicationOCEANS'10 IEEE Sydney, OCEANSSYD 2010
DOIs
StatePublished - 2010
EventOCEANS'10 IEEE Sydney, OCEANSSYD 2010 - Sydney, NSW, Australia
Duration: May 24 2010May 27 2010

Publication series

NameOCEANS'10 IEEE Sydney, OCEANSSYD 2010

Conference

ConferenceOCEANS'10 IEEE Sydney, OCEANSSYD 2010
Country/TerritoryAustralia
CitySydney, NSW
Period05/24/1005/27/10

Keywords

  • Ambiguity function (AF)
  • Doppler sensitive waveform (DSW)
  • Fast fourier transform (FFT)
  • Matched filter (MF)
  • Multi-core

Fingerprint

Dive into the research topics of 'Implementation/acceleration of vectorized acoustic source localization algorithms on the CELL multi-core processor'. Together they form a unique fingerprint.

Cite this