Regression prefetcher with preprocessing for DRAM-PCM hybrid main memory

Ji Tae Yun, Su Kyung Yoon, Jeong Geun Kim, bernd Burgstaller, Shin-Dug Kim

Research output: Contribution to journalArticle

Abstract

This research is to design an effective hybrid main memory structure for graph processing applications, because it is quite expensive to use only high-speed DRAM for such applications. Thus, we propose a DRAM-PCM hybrid main memory structure to reduce the cost and energy consumption and design regression prefetch scheme to cope with irregular access patterns in large graph processing workloads. In addition, the prefetch includes preprocessing algorithm to maximize prefetching performance. Our experimental evaluation shows a performance improvement of 36 percent over a conventional DRAM model, 15 percent over existing prefetch models such as GHB/PC, SMS, and AMPM, and 6 percent over the latest model.

Original languageEnglish
Pages (from-to)163-166
Number of pages4
JournalIEEE Computer Architecture Letters
Volume17
Issue number2
DOIs
Publication statusPublished - 2018 Jul 1

Fingerprint

Pulse code modulation
Dynamic random access storage
Data storage equipment
Processing
Energy utilization
Costs

All Science Journal Classification (ASJC) codes

  • Hardware and Architecture

Cite this

@article{4775d3798a99433d99e50f6d39c0d848,
title = "Regression prefetcher with preprocessing for DRAM-PCM hybrid main memory",
abstract = "This research is to design an effective hybrid main memory structure for graph processing applications, because it is quite expensive to use only high-speed DRAM for such applications. Thus, we propose a DRAM-PCM hybrid main memory structure to reduce the cost and energy consumption and design regression prefetch scheme to cope with irregular access patterns in large graph processing workloads. In addition, the prefetch includes preprocessing algorithm to maximize prefetching performance. Our experimental evaluation shows a performance improvement of 36 percent over a conventional DRAM model, 15 percent over existing prefetch models such as GHB/PC, SMS, and AMPM, and 6 percent over the latest model.",
author = "Yun, {Ji Tae} and Yoon, {Su Kyung} and Kim, {Jeong Geun} and bernd Burgstaller and Shin-Dug Kim",
year = "2018",
month = "7",
day = "1",
doi = "10.1109/LCA.2018.2841835",
language = "English",
volume = "17",
pages = "163--166",
journal = "IEEE Computer Architecture Letters",
issn = "1556-6056",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "2",

}

Regression prefetcher with preprocessing for DRAM-PCM hybrid main memory. / Yun, Ji Tae; Yoon, Su Kyung; Kim, Jeong Geun; Burgstaller, bernd; Kim, Shin-Dug.

In: IEEE Computer Architecture Letters, Vol. 17, No. 2, 01.07.2018, p. 163-166.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Regression prefetcher with preprocessing for DRAM-PCM hybrid main memory

AU - Yun, Ji Tae

AU - Yoon, Su Kyung

AU - Kim, Jeong Geun

AU - Burgstaller, bernd

AU - Kim, Shin-Dug

PY - 2018/7/1

Y1 - 2018/7/1

N2 - This research is to design an effective hybrid main memory structure for graph processing applications, because it is quite expensive to use only high-speed DRAM for such applications. Thus, we propose a DRAM-PCM hybrid main memory structure to reduce the cost and energy consumption and design regression prefetch scheme to cope with irregular access patterns in large graph processing workloads. In addition, the prefetch includes preprocessing algorithm to maximize prefetching performance. Our experimental evaluation shows a performance improvement of 36 percent over a conventional DRAM model, 15 percent over existing prefetch models such as GHB/PC, SMS, and AMPM, and 6 percent over the latest model.

AB - This research is to design an effective hybrid main memory structure for graph processing applications, because it is quite expensive to use only high-speed DRAM for such applications. Thus, we propose a DRAM-PCM hybrid main memory structure to reduce the cost and energy consumption and design regression prefetch scheme to cope with irregular access patterns in large graph processing workloads. In addition, the prefetch includes preprocessing algorithm to maximize prefetching performance. Our experimental evaluation shows a performance improvement of 36 percent over a conventional DRAM model, 15 percent over existing prefetch models such as GHB/PC, SMS, and AMPM, and 6 percent over the latest model.

UR - http://www.scopus.com/inward/record.url?scp=85047840923&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047840923&partnerID=8YFLogxK

U2 - 10.1109/LCA.2018.2841835

DO - 10.1109/LCA.2018.2841835

M3 - Article

AN - SCOPUS:85047840923

VL - 17

SP - 163

EP - 166

JO - IEEE Computer Architecture Letters

JF - IEEE Computer Architecture Letters

SN - 1556-6056

IS - 2

ER -