On-chip memory space partitioning for chip multiprocessors using polyhedral algebra

O. Ozturk, M. Kandemir, M. J. Irwin

Research output: Contribution to journalArticle

Abstract

One of the most important issues in designing a chip multiprocessor is to decide its on-chip memory organisation. While it is possible to design an application-specific memory architecture, this may not necessarily be the best option, in particular when storage demands of individual processors and/or their data sharing patterns can change from one point in execution to another for the same application. Here, two problems are formulated. First, we show how a polyhedral method can be used to design, for array-based data-intensive embedded applications, an application-specific hybrid memory architecture that has both shared and private components. We evaluate the resulting memory configurations using a set of benchmarks and compare them to pure private and pure shared memory on-chip multiprocessor architectures. The second approach proposed consider dynamic configuration of software-managed on-chip memory space to adapt to the runtime variations in data storage demand and interprocessor sharing patterns. The proposed framework is fully implemented using an optimising compiler, a polyhedral tool, and a memory partitioner (based on integer linear programming), and is tested using a suite of eight data-intensive embedded applications.

Original languageEnglish (US)
Pages (from-to)484-498
Number of pages15
JournalIET Computers and Digital Techniques
Volume4
Issue number6
DOIs
StatePublished - Nov 1 2010

Fingerprint

Algebra
Data storage equipment
Memory architecture
Linear programming

All Science Journal Classification (ASJC) codes

  • Software
  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

@article{5e96dab0b56e4adbbcee28a99123e3ea,
title = "On-chip memory space partitioning for chip multiprocessors using polyhedral algebra",
abstract = "One of the most important issues in designing a chip multiprocessor is to decide its on-chip memory organisation. While it is possible to design an application-specific memory architecture, this may not necessarily be the best option, in particular when storage demands of individual processors and/or their data sharing patterns can change from one point in execution to another for the same application. Here, two problems are formulated. First, we show how a polyhedral method can be used to design, for array-based data-intensive embedded applications, an application-specific hybrid memory architecture that has both shared and private components. We evaluate the resulting memory configurations using a set of benchmarks and compare them to pure private and pure shared memory on-chip multiprocessor architectures. The second approach proposed consider dynamic configuration of software-managed on-chip memory space to adapt to the runtime variations in data storage demand and interprocessor sharing patterns. The proposed framework is fully implemented using an optimising compiler, a polyhedral tool, and a memory partitioner (based on integer linear programming), and is tested using a suite of eight data-intensive embedded applications.",
author = "O. Ozturk and M. Kandemir and Irwin, {M. J.}",
year = "2010",
month = "11",
day = "1",
doi = "10.1049/iet-cdt.2009.0089",
language = "English (US)",
volume = "4",
pages = "484--498",
journal = "IET Computers and Digital Techniques",
issn = "1751-8601",
publisher = "Institution of Engineering and Technology",
number = "6",

}

On-chip memory space partitioning for chip multiprocessors using polyhedral algebra. / Ozturk, O.; Kandemir, M.; Irwin, M. J.

In: IET Computers and Digital Techniques, Vol. 4, No. 6, 01.11.2010, p. 484-498.

Research output: Contribution to journalArticle

TY - JOUR

T1 - On-chip memory space partitioning for chip multiprocessors using polyhedral algebra

AU - Ozturk, O.

AU - Kandemir, M.

AU - Irwin, M. J.

PY - 2010/11/1

Y1 - 2010/11/1

N2 - One of the most important issues in designing a chip multiprocessor is to decide its on-chip memory organisation. While it is possible to design an application-specific memory architecture, this may not necessarily be the best option, in particular when storage demands of individual processors and/or their data sharing patterns can change from one point in execution to another for the same application. Here, two problems are formulated. First, we show how a polyhedral method can be used to design, for array-based data-intensive embedded applications, an application-specific hybrid memory architecture that has both shared and private components. We evaluate the resulting memory configurations using a set of benchmarks and compare them to pure private and pure shared memory on-chip multiprocessor architectures. The second approach proposed consider dynamic configuration of software-managed on-chip memory space to adapt to the runtime variations in data storage demand and interprocessor sharing patterns. The proposed framework is fully implemented using an optimising compiler, a polyhedral tool, and a memory partitioner (based on integer linear programming), and is tested using a suite of eight data-intensive embedded applications.

AB - One of the most important issues in designing a chip multiprocessor is to decide its on-chip memory organisation. While it is possible to design an application-specific memory architecture, this may not necessarily be the best option, in particular when storage demands of individual processors and/or their data sharing patterns can change from one point in execution to another for the same application. Here, two problems are formulated. First, we show how a polyhedral method can be used to design, for array-based data-intensive embedded applications, an application-specific hybrid memory architecture that has both shared and private components. We evaluate the resulting memory configurations using a set of benchmarks and compare them to pure private and pure shared memory on-chip multiprocessor architectures. The second approach proposed consider dynamic configuration of software-managed on-chip memory space to adapt to the runtime variations in data storage demand and interprocessor sharing patterns. The proposed framework is fully implemented using an optimising compiler, a polyhedral tool, and a memory partitioner (based on integer linear programming), and is tested using a suite of eight data-intensive embedded applications.

UR - http://www.scopus.com/inward/record.url?scp=78149397609&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78149397609&partnerID=8YFLogxK

U2 - 10.1049/iet-cdt.2009.0089

DO - 10.1049/iet-cdt.2009.0089

M3 - Article

AN - SCOPUS:78149397609

VL - 4

SP - 484

EP - 498

JO - IET Computers and Digital Techniques

JF - IET Computers and Digital Techniques

SN - 1751-8601

IS - 6

ER -