Improving offset assignment on embedded processors using transformations

Sunil Atri, J. Ramanujam, Mahmut Kandemir

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Scopus citations


Embedded systems consisting of the application program ROM, RAM, the embedded processor core and any custom hardware on a single wafer are becoming increasingly common in areas such as signal processing. In this paper, we address new code optimization techniques for embedded fixed point DSP processors which have limited on-chip program ROM and include indirect addressing modes usingpost increment and decrement operations. These addressingmodes allow for efficient sequential access but the addressinginstructions increase code size. Most of the previous approaches to the problem aim to find a placement or layout of variables in the memory so that it is possible to subsume explicit address pointer manipulation instructions into other instructions as a post-increment or post-decrement operation. Our solution is aimed at transformingthe access pattern by usingproperties of operators such as commutativity so that current algorithms for variable placement are more effective.

Original languageEnglish (US)
Title of host publicationHigh Performance Computing - HiPC 2000 - 7th International Conference, Proceedings
EditorsMateo Valero, Viktor K. Prasanna, Sriram Vajapeyam
PublisherSpringer Verlag
Number of pages8
ISBN (Print)3540414290, 9783540414292
StatePublished - 2000
Event7th International Conference on High Performance Computing, HiPC 2000 - Bangalore, India
Duration: Dec 17 2000Dec 20 2000

Publication series

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


Other7th International Conference on High Performance Computing, HiPC 2000

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Improving offset assignment on embedded processors using transformations'. Together they form a unique fingerprint.

Cite this