Project PBerry: FPGA Acceleration for Remote Memory

Irina Calciu, Ivan Puddu, Aasheesh Kolli, Andreas Nowatzyk, Jayneel Gandhi, Onur Mutlu, Pratap Subrahmanyam

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

Abstract

Recent research efforts propose remote memory systems that pool memory from multiple hosts. These systems rely on the virtual memory subsystem to track application memory accesses and transparently offer remote memory to applications. We outline several limitations of this approach, such as page fault overheads and dirty data amplification. Instead, we argue for a fundamentally different approach: leverage the local host's cache coherence traffic to track application memory accesses at cache line granularity. Our approach uses emerging cache-coherent FPGAs to expose cache coherence events to the operating system. This approach not only accelerates remote memory systems by reducing dirty data amplification and by eliminating page faults, but also enables other use cases, such as live virtual machine migration, unified virtual memory, security and code analysis. All of these use cases open up many promising research directions.

Original languageEnglish (US)
Title of host publicationProceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019
PublisherAssociation for Computing Machinery, Inc
Pages127-135
Number of pages9
ISBN (Electronic)9781450367271
DOIs
StatePublished - May 12 2019
Event2019 Workshop on Hot Topics in Operating Systems, HotOS 2019 - Bertinoro, Italy
Duration: May 13 2019May 15 2019

Publication series

NameProceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019

Conference

Conference2019 Workshop on Hot Topics in Operating Systems, HotOS 2019
CountryItaly
CityBertinoro
Period5/13/195/15/19

Fingerprint

Field programmable gate arrays (FPGA)
Data storage equipment
Amplification
Computer systems

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications
  • Hardware and Architecture
  • Information Systems

Cite this

Calciu, I., Puddu, I., Kolli, A., Nowatzyk, A., Gandhi, J., Mutlu, O., & Subrahmanyam, P. (2019). Project PBerry: FPGA Acceleration for Remote Memory. In Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019 (pp. 127-135). [3321424] (Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019). Association for Computing Machinery, Inc. https://doi.org/10.1145/3317550.3321424
Calciu, Irina ; Puddu, Ivan ; Kolli, Aasheesh ; Nowatzyk, Andreas ; Gandhi, Jayneel ; Mutlu, Onur ; Subrahmanyam, Pratap. / Project PBerry : FPGA Acceleration for Remote Memory. Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019. Association for Computing Machinery, Inc, 2019. pp. 127-135 (Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019).
@inproceedings{1742542fa2d84efcb11e9cdaa40dd4dc,
title = "Project PBerry: FPGA Acceleration for Remote Memory",
abstract = "Recent research efforts propose remote memory systems that pool memory from multiple hosts. These systems rely on the virtual memory subsystem to track application memory accesses and transparently offer remote memory to applications. We outline several limitations of this approach, such as page fault overheads and dirty data amplification. Instead, we argue for a fundamentally different approach: leverage the local host's cache coherence traffic to track application memory accesses at cache line granularity. Our approach uses emerging cache-coherent FPGAs to expose cache coherence events to the operating system. This approach not only accelerates remote memory systems by reducing dirty data amplification and by eliminating page faults, but also enables other use cases, such as live virtual machine migration, unified virtual memory, security and code analysis. All of these use cases open up many promising research directions.",
author = "Irina Calciu and Ivan Puddu and Aasheesh Kolli and Andreas Nowatzyk and Jayneel Gandhi and Onur Mutlu and Pratap Subrahmanyam",
year = "2019",
month = "5",
day = "12",
doi = "10.1145/3317550.3321424",
language = "English (US)",
series = "Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019",
publisher = "Association for Computing Machinery, Inc",
pages = "127--135",
booktitle = "Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019",

}

Calciu, I, Puddu, I, Kolli, A, Nowatzyk, A, Gandhi, J, Mutlu, O & Subrahmanyam, P 2019, Project PBerry: FPGA Acceleration for Remote Memory. in Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019., 3321424, Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019, Association for Computing Machinery, Inc, pp. 127-135, 2019 Workshop on Hot Topics in Operating Systems, HotOS 2019, Bertinoro, Italy, 5/13/19. https://doi.org/10.1145/3317550.3321424

Project PBerry : FPGA Acceleration for Remote Memory. / Calciu, Irina; Puddu, Ivan; Kolli, Aasheesh; Nowatzyk, Andreas; Gandhi, Jayneel; Mutlu, Onur; Subrahmanyam, Pratap.

Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019. Association for Computing Machinery, Inc, 2019. p. 127-135 3321424 (Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019).

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

TY - GEN

T1 - Project PBerry

T2 - FPGA Acceleration for Remote Memory

AU - Calciu, Irina

AU - Puddu, Ivan

AU - Kolli, Aasheesh

AU - Nowatzyk, Andreas

AU - Gandhi, Jayneel

AU - Mutlu, Onur

AU - Subrahmanyam, Pratap

PY - 2019/5/12

Y1 - 2019/5/12

N2 - Recent research efforts propose remote memory systems that pool memory from multiple hosts. These systems rely on the virtual memory subsystem to track application memory accesses and transparently offer remote memory to applications. We outline several limitations of this approach, such as page fault overheads and dirty data amplification. Instead, we argue for a fundamentally different approach: leverage the local host's cache coherence traffic to track application memory accesses at cache line granularity. Our approach uses emerging cache-coherent FPGAs to expose cache coherence events to the operating system. This approach not only accelerates remote memory systems by reducing dirty data amplification and by eliminating page faults, but also enables other use cases, such as live virtual machine migration, unified virtual memory, security and code analysis. All of these use cases open up many promising research directions.

AB - Recent research efforts propose remote memory systems that pool memory from multiple hosts. These systems rely on the virtual memory subsystem to track application memory accesses and transparently offer remote memory to applications. We outline several limitations of this approach, such as page fault overheads and dirty data amplification. Instead, we argue for a fundamentally different approach: leverage the local host's cache coherence traffic to track application memory accesses at cache line granularity. Our approach uses emerging cache-coherent FPGAs to expose cache coherence events to the operating system. This approach not only accelerates remote memory systems by reducing dirty data amplification and by eliminating page faults, but also enables other use cases, such as live virtual machine migration, unified virtual memory, security and code analysis. All of these use cases open up many promising research directions.

UR - http://www.scopus.com/inward/record.url?scp=85066896060&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066896060&partnerID=8YFLogxK

U2 - 10.1145/3317550.3321424

DO - 10.1145/3317550.3321424

M3 - Conference contribution

AN - SCOPUS:85066896060

T3 - Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019

SP - 127

EP - 135

BT - Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019

PB - Association for Computing Machinery, Inc

ER -

Calciu I, Puddu I, Kolli A, Nowatzyk A, Gandhi J, Mutlu O et al. Project PBerry: FPGA Acceleration for Remote Memory. In Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019. Association for Computing Machinery, Inc. 2019. p. 127-135. 3321424. (Proceedings of the Workshop on Hot Topics in Operating Systems, HotOS 2019). https://doi.org/10.1145/3317550.3321424