An energy-efficient partitioned instruction cache architecture for embedded processors

CheolHong Kim, Sung Woo Jung, ChuShik Jhon

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Energy efficiency of cache memories is crucial in designing embedded processors. Reducing energy consumption in the instruction cache is especially important, since the instruction cache consumes a significant portion of total processor energy. This paper proposes a new instruction cache architecture, named Partitioned Instruction Cache (PI-Cache), for reducing dynamic energy consumption in the instruction cache by partitioning it to smaller (less power-consuming) sub-caches. When the proposed PI-Cache is accessed, only one sub-cache is accessed by utilizing the temporal/spatial locality of applications. In the meantime, other sub-caches are not accessed, leading to dynamic energy reduction. The PI-Cache also reduces dynamic energy consumption by eliminating the energy consumed in tag lookup and comparison. Moreover, the performance gap between the conventional instruction cache and the proposed PI-Cache becomes little when the physical cache access time is considered. We evaluated the energy efficiency by running a cycle accurate simulator. SimpleScalar, with power parameters obtained from CACTI. Simulation results show that the PI-Cache improves the energy-delay product by 20%-54% compared to the conventional direct-mapped instruction cache.

Original languageEnglish
Pages (from-to)1450-1458
Number of pages9
JournalIEICE Transactions on Information and Systems
VolumeE89-D
Issue number4
DOIs
Publication statusPublished - 2006 Apr 1

Fingerprint

Energy utilization
Energy efficiency
Cache memory
Simulators

Keywords

  • Dynamic energy
  • Embedded processor
  • Instruction cache
  • Low power design
  • Partitioned cache

ASJC Scopus subject areas

  • Information Systems
  • Computer Graphics and Computer-Aided Design
  • Software

Cite this

An energy-efficient partitioned instruction cache architecture for embedded processors. / Kim, CheolHong; Jung, Sung Woo; Jhon, ChuShik.

In: IEICE Transactions on Information and Systems, Vol. E89-D, No. 4, 01.04.2006, p. 1450-1458.

Research output: Contribution to journalArticle

@article{ebb38d7f67634d2f8d19fe4e378187b8,
title = "An energy-efficient partitioned instruction cache architecture for embedded processors",
abstract = "Energy efficiency of cache memories is crucial in designing embedded processors. Reducing energy consumption in the instruction cache is especially important, since the instruction cache consumes a significant portion of total processor energy. This paper proposes a new instruction cache architecture, named Partitioned Instruction Cache (PI-Cache), for reducing dynamic energy consumption in the instruction cache by partitioning it to smaller (less power-consuming) sub-caches. When the proposed PI-Cache is accessed, only one sub-cache is accessed by utilizing the temporal/spatial locality of applications. In the meantime, other sub-caches are not accessed, leading to dynamic energy reduction. The PI-Cache also reduces dynamic energy consumption by eliminating the energy consumed in tag lookup and comparison. Moreover, the performance gap between the conventional instruction cache and the proposed PI-Cache becomes little when the physical cache access time is considered. We evaluated the energy efficiency by running a cycle accurate simulator. SimpleScalar, with power parameters obtained from CACTI. Simulation results show that the PI-Cache improves the energy-delay product by 20{\%}-54{\%} compared to the conventional direct-mapped instruction cache.",
keywords = "Dynamic energy, Embedded processor, Instruction cache, Low power design, Partitioned cache",
author = "CheolHong Kim and Jung, {Sung Woo} and ChuShik Jhon",
year = "2006",
month = "4",
day = "1",
doi = "10.1093/ietisy/e89-d.4.1450",
language = "English",
volume = "E89-D",
pages = "1450--1458",
journal = "IEICE Transactions on Information and Systems",
issn = "0916-8532",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "4",

}

TY - JOUR

T1 - An energy-efficient partitioned instruction cache architecture for embedded processors

AU - Kim, CheolHong

AU - Jung, Sung Woo

AU - Jhon, ChuShik

PY - 2006/4/1

Y1 - 2006/4/1

N2 - Energy efficiency of cache memories is crucial in designing embedded processors. Reducing energy consumption in the instruction cache is especially important, since the instruction cache consumes a significant portion of total processor energy. This paper proposes a new instruction cache architecture, named Partitioned Instruction Cache (PI-Cache), for reducing dynamic energy consumption in the instruction cache by partitioning it to smaller (less power-consuming) sub-caches. When the proposed PI-Cache is accessed, only one sub-cache is accessed by utilizing the temporal/spatial locality of applications. In the meantime, other sub-caches are not accessed, leading to dynamic energy reduction. The PI-Cache also reduces dynamic energy consumption by eliminating the energy consumed in tag lookup and comparison. Moreover, the performance gap between the conventional instruction cache and the proposed PI-Cache becomes little when the physical cache access time is considered. We evaluated the energy efficiency by running a cycle accurate simulator. SimpleScalar, with power parameters obtained from CACTI. Simulation results show that the PI-Cache improves the energy-delay product by 20%-54% compared to the conventional direct-mapped instruction cache.

AB - Energy efficiency of cache memories is crucial in designing embedded processors. Reducing energy consumption in the instruction cache is especially important, since the instruction cache consumes a significant portion of total processor energy. This paper proposes a new instruction cache architecture, named Partitioned Instruction Cache (PI-Cache), for reducing dynamic energy consumption in the instruction cache by partitioning it to smaller (less power-consuming) sub-caches. When the proposed PI-Cache is accessed, only one sub-cache is accessed by utilizing the temporal/spatial locality of applications. In the meantime, other sub-caches are not accessed, leading to dynamic energy reduction. The PI-Cache also reduces dynamic energy consumption by eliminating the energy consumed in tag lookup and comparison. Moreover, the performance gap between the conventional instruction cache and the proposed PI-Cache becomes little when the physical cache access time is considered. We evaluated the energy efficiency by running a cycle accurate simulator. SimpleScalar, with power parameters obtained from CACTI. Simulation results show that the PI-Cache improves the energy-delay product by 20%-54% compared to the conventional direct-mapped instruction cache.

KW - Dynamic energy

KW - Embedded processor

KW - Instruction cache

KW - Low power design

KW - Partitioned cache

UR - http://www.scopus.com/inward/record.url?scp=33646229288&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33646229288&partnerID=8YFLogxK

U2 - 10.1093/ietisy/e89-d.4.1450

DO - 10.1093/ietisy/e89-d.4.1450

M3 - Article

AN - SCOPUS:33646229288

VL - E89-D

SP - 1450

EP - 1458

JO - IEICE Transactions on Information and Systems

JF - IEICE Transactions on Information and Systems

SN - 0916-8532

IS - 4

ER -