DUANG

Fast and lightweight page migration in asymmetric memory systems

Hao Wang, Jie Zhang, Sharmila Shridhar, Gieseo Park, Myoungsoo Jung, Nam Sung Kim

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

5 Citations (Scopus)

Abstract

Main memory systems have gone through dramatic increases in bandwidth and capacity. At the same time, their random access latency has remained relatively constant. For given memory technology, optimizing the latency typically requires sacrificing the density (i.e., cost per bit), which is one of the most critical concerns for memory industry. Recent studies have proposed memory architectures comprised of asymmetric (fast/low-density and slow/high-density) regions to optimize between overall latency and negative impact on density. Such memory architectures attempt to cost-effectively offer both high capacity and high performance. Yet they present a unique challenge, requiring direct placements of hot memory pages1 in the fast region and/or expensive runtime page migrations. In this paper, we propose a novel resistive memory architecture sharing a set of row buffers between a pair of neighboring banks. It enables two attractive techniques: (1) migrating memory pages between slow and fast banks with little performance overhead and (2) adaptively allocating more row buffers to busier banks based on memory access patterns. For an asymmetric memory architecture with both slow/high-density and fast/low-density banks, our shared row-buffer architecture can capture 87-93% of the performance of a memory architecture with only fast banks.

Original languageEnglish
Title of host publicationProceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016
PublisherIEEE Computer Society
Pages481-493
Number of pages13
ISBN (Electronic)9781467392112
DOIs
Publication statusPublished - 2016 Apr 1
Event22nd IEEE International Symposium on High Performance Computer Architecture, HPCA 2016 - Barcelona, Spain
Duration: 2016 Mar 122016 Mar 16

Publication series

NameProceedings - International Symposium on High-Performance Computer Architecture
Volume2016-April
ISSN (Print)1530-0897

Other

Other22nd IEEE International Symposium on High Performance Computer Architecture, HPCA 2016
CountrySpain
CityBarcelona
Period16/3/1216/3/16

Fingerprint

Memory architecture
Computer systems
Data storage equipment
Costs
Bandwidth
Industry

All Science Journal Classification (ASJC) codes

  • Hardware and Architecture

Cite this

Wang, H., Zhang, J., Shridhar, S., Park, G., Jung, M., & Kim, N. S. (2016). DUANG: Fast and lightweight page migration in asymmetric memory systems. In Proceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016 (pp. 481-493). [7446088] (Proceedings - International Symposium on High-Performance Computer Architecture; Vol. 2016-April). IEEE Computer Society. https://doi.org/10.1109/HPCA.2016.7446088
Wang, Hao ; Zhang, Jie ; Shridhar, Sharmila ; Park, Gieseo ; Jung, Myoungsoo ; Kim, Nam Sung. / DUANG : Fast and lightweight page migration in asymmetric memory systems. Proceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016. IEEE Computer Society, 2016. pp. 481-493 (Proceedings - International Symposium on High-Performance Computer Architecture).
@inproceedings{bab10e7b4a3a4ca28cdf74157165248c,
title = "DUANG: Fast and lightweight page migration in asymmetric memory systems",
abstract = "Main memory systems have gone through dramatic increases in bandwidth and capacity. At the same time, their random access latency has remained relatively constant. For given memory technology, optimizing the latency typically requires sacrificing the density (i.e., cost per bit), which is one of the most critical concerns for memory industry. Recent studies have proposed memory architectures comprised of asymmetric (fast/low-density and slow/high-density) regions to optimize between overall latency and negative impact on density. Such memory architectures attempt to cost-effectively offer both high capacity and high performance. Yet they present a unique challenge, requiring direct placements of hot memory pages1 in the fast region and/or expensive runtime page migrations. In this paper, we propose a novel resistive memory architecture sharing a set of row buffers between a pair of neighboring banks. It enables two attractive techniques: (1) migrating memory pages between slow and fast banks with little performance overhead and (2) adaptively allocating more row buffers to busier banks based on memory access patterns. For an asymmetric memory architecture with both slow/high-density and fast/low-density banks, our shared row-buffer architecture can capture 87-93{\%} of the performance of a memory architecture with only fast banks.",
author = "Hao Wang and Jie Zhang and Sharmila Shridhar and Gieseo Park and Myoungsoo Jung and Kim, {Nam Sung}",
year = "2016",
month = "4",
day = "1",
doi = "10.1109/HPCA.2016.7446088",
language = "English",
series = "Proceedings - International Symposium on High-Performance Computer Architecture",
publisher = "IEEE Computer Society",
pages = "481--493",
booktitle = "Proceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016",
address = "United States",

}

Wang, H, Zhang, J, Shridhar, S, Park, G, Jung, M & Kim, NS 2016, DUANG: Fast and lightweight page migration in asymmetric memory systems. in Proceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016., 7446088, Proceedings - International Symposium on High-Performance Computer Architecture, vol. 2016-April, IEEE Computer Society, pp. 481-493, 22nd IEEE International Symposium on High Performance Computer Architecture, HPCA 2016, Barcelona, Spain, 16/3/12. https://doi.org/10.1109/HPCA.2016.7446088

DUANG : Fast and lightweight page migration in asymmetric memory systems. / Wang, Hao; Zhang, Jie; Shridhar, Sharmila; Park, Gieseo; Jung, Myoungsoo; Kim, Nam Sung.

Proceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016. IEEE Computer Society, 2016. p. 481-493 7446088 (Proceedings - International Symposium on High-Performance Computer Architecture; Vol. 2016-April).

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

TY - GEN

T1 - DUANG

T2 - Fast and lightweight page migration in asymmetric memory systems

AU - Wang, Hao

AU - Zhang, Jie

AU - Shridhar, Sharmila

AU - Park, Gieseo

AU - Jung, Myoungsoo

AU - Kim, Nam Sung

PY - 2016/4/1

Y1 - 2016/4/1

N2 - Main memory systems have gone through dramatic increases in bandwidth and capacity. At the same time, their random access latency has remained relatively constant. For given memory technology, optimizing the latency typically requires sacrificing the density (i.e., cost per bit), which is one of the most critical concerns for memory industry. Recent studies have proposed memory architectures comprised of asymmetric (fast/low-density and slow/high-density) regions to optimize between overall latency and negative impact on density. Such memory architectures attempt to cost-effectively offer both high capacity and high performance. Yet they present a unique challenge, requiring direct placements of hot memory pages1 in the fast region and/or expensive runtime page migrations. In this paper, we propose a novel resistive memory architecture sharing a set of row buffers between a pair of neighboring banks. It enables two attractive techniques: (1) migrating memory pages between slow and fast banks with little performance overhead and (2) adaptively allocating more row buffers to busier banks based on memory access patterns. For an asymmetric memory architecture with both slow/high-density and fast/low-density banks, our shared row-buffer architecture can capture 87-93% of the performance of a memory architecture with only fast banks.

AB - Main memory systems have gone through dramatic increases in bandwidth and capacity. At the same time, their random access latency has remained relatively constant. For given memory technology, optimizing the latency typically requires sacrificing the density (i.e., cost per bit), which is one of the most critical concerns for memory industry. Recent studies have proposed memory architectures comprised of asymmetric (fast/low-density and slow/high-density) regions to optimize between overall latency and negative impact on density. Such memory architectures attempt to cost-effectively offer both high capacity and high performance. Yet they present a unique challenge, requiring direct placements of hot memory pages1 in the fast region and/or expensive runtime page migrations. In this paper, we propose a novel resistive memory architecture sharing a set of row buffers between a pair of neighboring banks. It enables two attractive techniques: (1) migrating memory pages between slow and fast banks with little performance overhead and (2) adaptively allocating more row buffers to busier banks based on memory access patterns. For an asymmetric memory architecture with both slow/high-density and fast/low-density banks, our shared row-buffer architecture can capture 87-93% of the performance of a memory architecture with only fast banks.

UR - http://www.scopus.com/inward/record.url?scp=84965032112&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84965032112&partnerID=8YFLogxK

U2 - 10.1109/HPCA.2016.7446088

DO - 10.1109/HPCA.2016.7446088

M3 - Conference contribution

T3 - Proceedings - International Symposium on High-Performance Computer Architecture

SP - 481

EP - 493

BT - Proceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016

PB - IEEE Computer Society

ER -

Wang H, Zhang J, Shridhar S, Park G, Jung M, Kim NS. DUANG: Fast and lightweight page migration in asymmetric memory systems. In Proceedings of the 2016 IEEE International Symposium on High-Performance Computer Architecture, HPCA 2016. IEEE Computer Society. 2016. p. 481-493. 7446088. (Proceedings - International Symposium on High-Performance Computer Architecture). https://doi.org/10.1109/HPCA.2016.7446088