LazyPIM: An Efficient Cache Coherence Mechanism for Processing-in-Memory

Amirali Boroumand; Saugata Ghose; Minesh Patel; Hasan Hassan; Brandon Lucia; Kevin Hsieh; Krishna T. Malladi; Hongzhong Zheng; Onur Mutlu

doi:10.1109/LCA.2016.2577557

LazyPIM: An Efficient Cache Coherence Mechanism for Processing-in-Memory

Amirali Boroumand, Saugata Ghose, Minesh Patel, Hasan Hassan, Brandon Lucia, Kevin Hsieh, Krishna T. Malladi, Hongzhong Zheng, Onur Mutlu

Research output: Contribution to journal › Article › peer-review

Abstract

Processing-in-memory (PIM) architectures cannot use traditional approaches to cache coherence due to the high off-chip traffic consumed by coherence messages. We propose LazyPIM, a new hardware cache coherence mechanism designed specifically for PIM. LazyPIM uses a combination of speculative cache coherence and compressed coherence signatures to greatly reduce the overhead of keeping PIM coherent with the processor. We find that LazyPIM improves average performance across a range of PIM applications by 49.1 percent over the best prior approach, coming within 5.5 percent of an ideal PIM mechanism.

Original language	English (US)
Article number	7485993
Pages (from-to)	46-50
Number of pages	5
Journal	IEEE Computer Architecture Letters
Volume	16
Issue number	1
DOIs	https://doi.org/10.1109/LCA.2016.2577557
State	Published - Jan 1 2017
Externally published	Yes

ASJC Scopus subject areas

Hardware and Architecture

Online availability

10.1109/LCA.2016.2577557

Library availability

Discover UIUC Full Text

Cite this

@article{d258f67211a244739cb1621ae6dc2164,

title = "LazyPIM: An Efficient Cache Coherence Mechanism for Processing-in-Memory",

abstract = "Processing-in-memory (PIM) architectures cannot use traditional approaches to cache coherence due to the high off-chip traffic consumed by coherence messages. We propose LazyPIM, a new hardware cache coherence mechanism designed specifically for PIM. LazyPIM uses a combination of speculative cache coherence and compressed coherence signatures to greatly reduce the overhead of keeping PIM coherent with the processor. We find that LazyPIM improves average performance across a range of PIM applications by 49.1 percent over the best prior approach, coming within 5.5 percent of an ideal PIM mechanism.",

author = "Amirali Boroumand and Saugata Ghose and Minesh Patel and Hasan Hassan and Brandon Lucia and Kevin Hsieh and Malladi, {Krishna T.} and Hongzhong Zheng and Onur Mutlu",

year = "2017",

month = jan,

day = "1",

doi = "10.1109/LCA.2016.2577557",

language = "English (US)",

volume = "16",

pages = "46--50",

journal = "IEEE Computer Architecture Letters",

issn = "1556-6056",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "1",

}

TY - JOUR

T1 - LazyPIM

T2 - An Efficient Cache Coherence Mechanism for Processing-in-Memory

AU - Boroumand, Amirali

AU - Ghose, Saugata

AU - Patel, Minesh

AU - Hassan, Hasan

AU - Lucia, Brandon

AU - Hsieh, Kevin

AU - Malladi, Krishna T.

AU - Zheng, Hongzhong

AU - Mutlu, Onur

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Processing-in-memory (PIM) architectures cannot use traditional approaches to cache coherence due to the high off-chip traffic consumed by coherence messages. We propose LazyPIM, a new hardware cache coherence mechanism designed specifically for PIM. LazyPIM uses a combination of speculative cache coherence and compressed coherence signatures to greatly reduce the overhead of keeping PIM coherent with the processor. We find that LazyPIM improves average performance across a range of PIM applications by 49.1 percent over the best prior approach, coming within 5.5 percent of an ideal PIM mechanism.

AB - Processing-in-memory (PIM) architectures cannot use traditional approaches to cache coherence due to the high off-chip traffic consumed by coherence messages. We propose LazyPIM, a new hardware cache coherence mechanism designed specifically for PIM. LazyPIM uses a combination of speculative cache coherence and compressed coherence signatures to greatly reduce the overhead of keeping PIM coherent with the processor. We find that LazyPIM improves average performance across a range of PIM applications by 49.1 percent over the best prior approach, coming within 5.5 percent of an ideal PIM mechanism.

UR - http://www.scopus.com/inward/record.url?scp=85021344519&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85021344519&partnerID=8YFLogxK

U2 - 10.1109/LCA.2016.2577557

DO - 10.1109/LCA.2016.2577557

M3 - Article

AN - SCOPUS:85021344519

SN - 1556-6056

VL - 16

SP - 46

EP - 50

JO - IEEE Computer Architecture Letters

JF - IEEE Computer Architecture Letters

IS - 1

M1 - 7485993

ER -

LazyPIM: An Efficient Cache Coherence Mechanism for Processing-in-Memory

Abstract

ASJC Scopus subject areas

Online availability

Library availability

Related links

Fingerprint

Cite this