Data space oriented tiling

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

3 Citations (Scopus)

Abstract

An optimizing compiler can play an important role in enhancing data locality in array-intensive applications with regular data access patterns. This paper presents a compiler-based data space oriented tiling approach (DST). In this strategy, the data space (i.e., thearray index space) is logically divided into chunks (called data tiles) andeach data tile is processed in turn. In processing a data tile, our approach traverses the entire iteration space of all nests in the code and executes all iterations (potentially coming from different nests) that access thedata tile being processed. In doing so, it also takes data dependences into account. Since a data space is common across all nests that accessit, DST can potentially achieve better results than traditional tiling by exploiting inter-nest data locality. This paper also shows how data spaceoriented tiling can be used for improving the performance of softwaremanaged scratch pad memories.

Original languageEnglish (US)
Title of host publicationProgramming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings
EditorsDaniel Le Métayer
PublisherSpringer Verlag
Pages178-193
Number of pages16
ISBN (Print)3540433635, 9783540433637
StatePublished - Jan 1 2002
Event11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002 - Grenoble, France
Duration: Apr 8 2002Apr 12 2002

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2305
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002
CountryFrance
CityGrenoble
Period4/8/024/12/02

Fingerprint

Tile
Tiling
Nest
Data Locality
Optimizing Compilers
Iteration
Data storage equipment
Data Dependence
Processing
Compiler
Entire

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Kandemir, M. (2002). Data space oriented tiling. In D. Le Métayer (Ed.), Programming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings (pp. 178-193). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2305). Springer Verlag.
Kandemir, Mahmut. / Data space oriented tiling. Programming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings. editor / Daniel Le Métayer. Springer Verlag, 2002. pp. 178-193 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{8cc67b1a11fe4e1dbefd385f4b27c32f,
title = "Data space oriented tiling",
abstract = "An optimizing compiler can play an important role in enhancing data locality in array-intensive applications with regular data access patterns. This paper presents a compiler-based data space oriented tiling approach (DST). In this strategy, the data space (i.e., thearray index space) is logically divided into chunks (called data tiles) andeach data tile is processed in turn. In processing a data tile, our approach traverses the entire iteration space of all nests in the code and executes all iterations (potentially coming from different nests) that access thedata tile being processed. In doing so, it also takes data dependences into account. Since a data space is common across all nests that accessit, DST can potentially achieve better results than traditional tiling by exploiting inter-nest data locality. This paper also shows how data spaceoriented tiling can be used for improving the performance of softwaremanaged scratch pad memories.",
author = "Mahmut Kandemir",
year = "2002",
month = "1",
day = "1",
language = "English (US)",
isbn = "3540433635",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "178--193",
editor = "{Le M{\'e}tayer}, Daniel",
booktitle = "Programming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings",
address = "Germany",

}

Kandemir, M 2002, Data space oriented tiling. in D Le Métayer (ed.), Programming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2305, Springer Verlag, pp. 178-193, 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Grenoble, France, 4/8/02.

Data space oriented tiling. / Kandemir, Mahmut.

Programming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings. ed. / Daniel Le Métayer. Springer Verlag, 2002. p. 178-193 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2305).

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

TY - GEN

T1 - Data space oriented tiling

AU - Kandemir, Mahmut

PY - 2002/1/1

Y1 - 2002/1/1

N2 - An optimizing compiler can play an important role in enhancing data locality in array-intensive applications with regular data access patterns. This paper presents a compiler-based data space oriented tiling approach (DST). In this strategy, the data space (i.e., thearray index space) is logically divided into chunks (called data tiles) andeach data tile is processed in turn. In processing a data tile, our approach traverses the entire iteration space of all nests in the code and executes all iterations (potentially coming from different nests) that access thedata tile being processed. In doing so, it also takes data dependences into account. Since a data space is common across all nests that accessit, DST can potentially achieve better results than traditional tiling by exploiting inter-nest data locality. This paper also shows how data spaceoriented tiling can be used for improving the performance of softwaremanaged scratch pad memories.

AB - An optimizing compiler can play an important role in enhancing data locality in array-intensive applications with regular data access patterns. This paper presents a compiler-based data space oriented tiling approach (DST). In this strategy, the data space (i.e., thearray index space) is logically divided into chunks (called data tiles) andeach data tile is processed in turn. In processing a data tile, our approach traverses the entire iteration space of all nests in the code and executes all iterations (potentially coming from different nests) that access thedata tile being processed. In doing so, it also takes data dependences into account. Since a data space is common across all nests that accessit, DST can potentially achieve better results than traditional tiling by exploiting inter-nest data locality. This paper also shows how data spaceoriented tiling can be used for improving the performance of softwaremanaged scratch pad memories.

UR - http://www.scopus.com/inward/record.url?scp=84876590330&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876590330&partnerID=8YFLogxK

M3 - Conference contribution

SN - 3540433635

SN - 9783540433637

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 178

EP - 193

BT - Programming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings

A2 - Le Métayer, Daniel

PB - Springer Verlag

ER -

Kandemir M. Data space oriented tiling. In Le Métayer D, editor, Programming Languages and Systems - 11th European Symposium on Programming, ESOP 2002 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2002, Proceedings. Springer Verlag. 2002. p. 178-193. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).