SPM management using markov chain based data access prediction

Taylan Yemliha, Shekhar Srikantaiah, Mahmut Kandemir, Ozean Ozturk

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

4 Citations (Scopus)

Abstract

Leveraging the power of scratchpad memories (SPMs) available in most embedded systems today is crucial to extract maximum performance from application programs. While regular accesses like scalar values and array expressions with affine subscript functions have been tractable for compiler analysis (to be prefetched into SPM), irregular accesses like pointer accesses and indexed array accesses have not been easily amenable for compiler analysis. This paper presents an SPM management technique using Markov chain based data access prediction for such irregular accesses. Our approach takes advantage of inherent, but hidden reuse in data accesses made by irregular references. We have implemented our proposed approach using an optimizing compiler. In this paper, we also present a thorough comparison of our different dynamic prediction schemes with other SPM management schemes. SPM management using our approaches produces 12.7% to 28.5% improvements in performance across a range of applications with both regular and irregular access patterns, with an average improvement of 20.8%.

Original languageEnglish (US)
Title of host publication2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008
Pages566-569
Number of pages4
DOIs
StatePublished - Dec 26 2008
Event2008 International Conference on Computer-Aided Design, ICCAD - San Jose, CA, United States
Duration: Nov 10 2008Nov 13 2008

Publication series

NameIEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD
ISSN (Print)1092-3152

Other

Other2008 International Conference on Computer-Aided Design, ICCAD
CountryUnited States
CitySan Jose, CA
Period11/10/0811/13/08

Fingerprint

Markov processes
Data storage equipment
Application programs
Embedded systems

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Science Applications
  • Computer Graphics and Computer-Aided Design

Cite this

Yemliha, T., Srikantaiah, S., Kandemir, M., & Ozturk, O. (2008). SPM management using markov chain based data access prediction. In 2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008 (pp. 566-569). [4681632] (IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD). https://doi.org/10.1109/ICCAD.2008.4681632
Yemliha, Taylan ; Srikantaiah, Shekhar ; Kandemir, Mahmut ; Ozturk, Ozean. / SPM management using markov chain based data access prediction. 2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008. 2008. pp. 566-569 (IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD).
@inproceedings{349c400641c7440ba045f10bd65431b8,
title = "SPM management using markov chain based data access prediction",
abstract = "Leveraging the power of scratchpad memories (SPMs) available in most embedded systems today is crucial to extract maximum performance from application programs. While regular accesses like scalar values and array expressions with affine subscript functions have been tractable for compiler analysis (to be prefetched into SPM), irregular accesses like pointer accesses and indexed array accesses have not been easily amenable for compiler analysis. This paper presents an SPM management technique using Markov chain based data access prediction for such irregular accesses. Our approach takes advantage of inherent, but hidden reuse in data accesses made by irregular references. We have implemented our proposed approach using an optimizing compiler. In this paper, we also present a thorough comparison of our different dynamic prediction schemes with other SPM management schemes. SPM management using our approaches produces 12.7{\%} to 28.5{\%} improvements in performance across a range of applications with both regular and irregular access patterns, with an average improvement of 20.8{\%}.",
author = "Taylan Yemliha and Shekhar Srikantaiah and Mahmut Kandemir and Ozean Ozturk",
year = "2008",
month = "12",
day = "26",
doi = "10.1109/ICCAD.2008.4681632",
language = "English (US)",
isbn = "9781424428205",
series = "IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD",
pages = "566--569",
booktitle = "2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008",

}

Yemliha, T, Srikantaiah, S, Kandemir, M & Ozturk, O 2008, SPM management using markov chain based data access prediction. in 2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008., 4681632, IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD, pp. 566-569, 2008 International Conference on Computer-Aided Design, ICCAD, San Jose, CA, United States, 11/10/08. https://doi.org/10.1109/ICCAD.2008.4681632

SPM management using markov chain based data access prediction. / Yemliha, Taylan; Srikantaiah, Shekhar; Kandemir, Mahmut; Ozturk, Ozean.

2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008. 2008. p. 566-569 4681632 (IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD).

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

TY - GEN

T1 - SPM management using markov chain based data access prediction

AU - Yemliha, Taylan

AU - Srikantaiah, Shekhar

AU - Kandemir, Mahmut

AU - Ozturk, Ozean

PY - 2008/12/26

Y1 - 2008/12/26

N2 - Leveraging the power of scratchpad memories (SPMs) available in most embedded systems today is crucial to extract maximum performance from application programs. While regular accesses like scalar values and array expressions with affine subscript functions have been tractable for compiler analysis (to be prefetched into SPM), irregular accesses like pointer accesses and indexed array accesses have not been easily amenable for compiler analysis. This paper presents an SPM management technique using Markov chain based data access prediction for such irregular accesses. Our approach takes advantage of inherent, but hidden reuse in data accesses made by irregular references. We have implemented our proposed approach using an optimizing compiler. In this paper, we also present a thorough comparison of our different dynamic prediction schemes with other SPM management schemes. SPM management using our approaches produces 12.7% to 28.5% improvements in performance across a range of applications with both regular and irregular access patterns, with an average improvement of 20.8%.

AB - Leveraging the power of scratchpad memories (SPMs) available in most embedded systems today is crucial to extract maximum performance from application programs. While regular accesses like scalar values and array expressions with affine subscript functions have been tractable for compiler analysis (to be prefetched into SPM), irregular accesses like pointer accesses and indexed array accesses have not been easily amenable for compiler analysis. This paper presents an SPM management technique using Markov chain based data access prediction for such irregular accesses. Our approach takes advantage of inherent, but hidden reuse in data accesses made by irregular references. We have implemented our proposed approach using an optimizing compiler. In this paper, we also present a thorough comparison of our different dynamic prediction schemes with other SPM management schemes. SPM management using our approaches produces 12.7% to 28.5% improvements in performance across a range of applications with both regular and irregular access patterns, with an average improvement of 20.8%.

UR - http://www.scopus.com/inward/record.url?scp=57849143765&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=57849143765&partnerID=8YFLogxK

U2 - 10.1109/ICCAD.2008.4681632

DO - 10.1109/ICCAD.2008.4681632

M3 - Conference contribution

SN - 9781424428205

T3 - IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD

SP - 566

EP - 569

BT - 2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008

ER -

Yemliha T, Srikantaiah S, Kandemir M, Ozturk O. SPM management using markov chain based data access prediction. In 2008 IEEE/ACM International Conference on Computer-Aided Design Digest of Technical Papers, ICCAD 2008. 2008. p. 566-569. 4681632. (IEEE/ACM International Conference on Computer-Aided Design, Digest of Technical Papers, ICCAD). https://doi.org/10.1109/ICCAD.2008.4681632