Exploring the possibility of operating in the compressed domain

V. De La Luz, M. Kandemir, A. Sivasubramaniam, M. J. Irwin

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Data compression in caches has been studied from the performance and energy consumption points of view. In this paper, we study the possible benefits of operating with compressed operands. Operating in the compressed domain (i.e., executing instructions with compressed operands) as far as possible can offer several advantages. First, since it is less reliant on decompression, it may be possible to employ a fancier compression/decompression strategy, without the associated performance/power penalties, that can offer higher compression rates to further boost cache locality. Second, the transfers between cache and datapath can use fewer bits (to transmit codes instead of data values) and provide dynamic energy savings in the corresponding bus. In this paper, we demonstrate that there are cases where operations can be executed in the compressed domain, which can lead to performance improvements and energy savings. In particular, we show that by operating in the compressed domain, the effectiveness of prior techniques can be further improved.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
EditorsMarco Danelutto, Marco Vanneschi, Domenico Laforenza
PublisherSpringer Verlag
Pages507-515
Number of pages9
ISBN (Print)3540229248
DOIs
StatePublished - Jan 1 2004

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3149
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Exploring the possibility of operating in the compressed domain'. Together they form a unique fingerprint.

  • Cite this

    De La Luz, V., Kandemir, M., Sivasubramaniam, A., & Irwin, M. J. (2004). Exploring the possibility of operating in the compressed domain. In M. Danelutto, M. Vanneschi, & D. Laforenza (Eds.), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp. 507-515). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3149). Springer Verlag. https://doi.org/10.1007/978-3-540-27866-5_67