Reduction operator for wide-SIMDs reconsidered

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

1 Citation (Scopus)
91 Downloads (Pure)

Abstract

It has been shown that wide Single Instruction Multiple Data architectures (wide-SIMDs) can achieve high energy efficiency, especially in domains such as image and vision processing. In these and various other application domains, reduction is a frequently encountered operation, where multiple input elements need to be combined into a single element by an associative operation, e.g. addition or multiplication. There are many applications that require reduction such as: partial histogram merging, matrix multiplication and min/max-finding. Wide-SIMDs contain a large number of processing elements (PEs), which in general are connected by a minimal form of interconnect for scalability reasons. To efficiently support reduction operations on wide-SIMDs with such a minimal interconnect, we introduce two novel reduction algorithms which do not rely on complex communication networks or any dedicated hardware. The proposed approaches are compared with both dedicated hardware and other software solutions in terms of performance, area, and energy consumption. A practical case study demonstrates that the proposed software approach has much better generality, flexibility and no additional hardware cost. Compared to a dedicated hardware adder tree, the proposed software approach saves 6.8% area with a performance penalty of only 6.5%.
Original languageEnglish
Title of host publicationProceedings of the 51st Annual Design Automation Conference (DAC '14), 1-5 june 2014, San Francisco CA, United States
Place of PublicationSan Francisco, CA, USA
PublisherAssociation for Computing Machinery, Inc
ISBN (Print)978-1-4503-2730-5
DOIs
Publication statusPublished - 2014
EventDesign Automation Conference (DAC'14), May 1-5, 2014, San Francisco, California, United States - San Francisco, United States
Duration: 1 Jun 20145 Jun 2014

Conference

ConferenceDesign Automation Conference (DAC'14), May 1-5, 2014, San Francisco, California, United States
Abbreviated titleDAC'14
CountryUnited States
CitySan Francisco
Period1/06/145/06/14
OtherDesign Automation Conference (DAC'14)

Fingerprint

Dive into the research topics of 'Reduction operator for wide-SIMDs reconsidered'. Together they form a unique fingerprint.

Cite this