Return data interleaving for multi-channel embedded CMPs systems

Fei Hong, Aviral Shrivastava, Jongeun Lee

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Using multi-channel memory subsystems is an efficient way of satisfying high volume memory requests from CMPs. At the same time, the imbalance between memory bandwidth and bus performance opens up new possibility of optimization before they are sent to bus. This paper presents a new memory controller design for embedded CMPs systems when the return data from the return buffer is sent back to bus. Our scheduling policy, called return data interleaving (RDI) interleaves the return data of each request in a round robin manner. Further, for each request, it sends the critical word first. To evaluate our technique, we model an Intel XScale-based CMPs using M5 simulator for CMPs simulation and DRAMsim for memory subsystem simulation and examine the performance of MiBench and SPEC2000 benchmarks. Simulation results show that for memory-bound benchmarks running on the CMPs systems with the number of cores from 6 to 16, RDI can improve the execution time by average 11% and up to 16.9%.

Original languageEnglish (US)
Article number5936661
Pages (from-to)1351-1354
Number of pages4
JournalIEEE Transactions on Very Large Scale Integration (VLSI) Systems
Volume20
Issue number7
DOIs
StatePublished - 2012

Fingerprint

Embedded systems
Data storage equipment
Simulators
Scheduling
Bandwidth
Controllers

Keywords

  • Chip multi-core processor
  • multi-channel memory
  • return data interleaving (RDI)

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Hardware and Architecture
  • Software

Cite this

Return data interleaving for multi-channel embedded CMPs systems. / Hong, Fei; Shrivastava, Aviral; Lee, Jongeun.

In: IEEE Transactions on Very Large Scale Integration (VLSI) Systems, Vol. 20, No. 7, 5936661, 2012, p. 1351-1354.

Research output: Contribution to journalArticle

@article{85e32b79c50943df8b835122fedaeab5,
title = "Return data interleaving for multi-channel embedded CMPs systems",
abstract = "Using multi-channel memory subsystems is an efficient way of satisfying high volume memory requests from CMPs. At the same time, the imbalance between memory bandwidth and bus performance opens up new possibility of optimization before they are sent to bus. This paper presents a new memory controller design for embedded CMPs systems when the return data from the return buffer is sent back to bus. Our scheduling policy, called return data interleaving (RDI) interleaves the return data of each request in a round robin manner. Further, for each request, it sends the critical word first. To evaluate our technique, we model an Intel XScale-based CMPs using M5 simulator for CMPs simulation and DRAMsim for memory subsystem simulation and examine the performance of MiBench and SPEC2000 benchmarks. Simulation results show that for memory-bound benchmarks running on the CMPs systems with the number of cores from 6 to 16, RDI can improve the execution time by average 11{\%} and up to 16.9{\%}.",
keywords = "Chip multi-core processor, multi-channel memory, return data interleaving (RDI)",
author = "Fei Hong and Aviral Shrivastava and Jongeun Lee",
year = "2012",
doi = "10.1109/TVLSI.2011.2157368",
language = "English (US)",
volume = "20",
pages = "1351--1354",
journal = "IEEE Transactions on Very Large Scale Integration (VLSI) Systems",
issn = "1063-8210",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "7",

}

TY - JOUR

T1 - Return data interleaving for multi-channel embedded CMPs systems

AU - Hong, Fei

AU - Shrivastava, Aviral

AU - Lee, Jongeun

PY - 2012

Y1 - 2012

N2 - Using multi-channel memory subsystems is an efficient way of satisfying high volume memory requests from CMPs. At the same time, the imbalance between memory bandwidth and bus performance opens up new possibility of optimization before they are sent to bus. This paper presents a new memory controller design for embedded CMPs systems when the return data from the return buffer is sent back to bus. Our scheduling policy, called return data interleaving (RDI) interleaves the return data of each request in a round robin manner. Further, for each request, it sends the critical word first. To evaluate our technique, we model an Intel XScale-based CMPs using M5 simulator for CMPs simulation and DRAMsim for memory subsystem simulation and examine the performance of MiBench and SPEC2000 benchmarks. Simulation results show that for memory-bound benchmarks running on the CMPs systems with the number of cores from 6 to 16, RDI can improve the execution time by average 11% and up to 16.9%.

AB - Using multi-channel memory subsystems is an efficient way of satisfying high volume memory requests from CMPs. At the same time, the imbalance between memory bandwidth and bus performance opens up new possibility of optimization before they are sent to bus. This paper presents a new memory controller design for embedded CMPs systems when the return data from the return buffer is sent back to bus. Our scheduling policy, called return data interleaving (RDI) interleaves the return data of each request in a round robin manner. Further, for each request, it sends the critical word first. To evaluate our technique, we model an Intel XScale-based CMPs using M5 simulator for CMPs simulation and DRAMsim for memory subsystem simulation and examine the performance of MiBench and SPEC2000 benchmarks. Simulation results show that for memory-bound benchmarks running on the CMPs systems with the number of cores from 6 to 16, RDI can improve the execution time by average 11% and up to 16.9%.

KW - Chip multi-core processor

KW - multi-channel memory

KW - return data interleaving (RDI)

UR - http://www.scopus.com/inward/record.url?scp=84862233003&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84862233003&partnerID=8YFLogxK

U2 - 10.1109/TVLSI.2011.2157368

DO - 10.1109/TVLSI.2011.2157368

M3 - Article

VL - 20

SP - 1351

EP - 1354

JO - IEEE Transactions on Very Large Scale Integration (VLSI) Systems

JF - IEEE Transactions on Very Large Scale Integration (VLSI) Systems

SN - 1063-8210

IS - 7

M1 - 5936661

ER -