Hiding communication latency and coherence overhead in software DSMs

In this paper we propose the use of a PCI-based programmable protocol controller for hiding communication and coherence overheads in software DSMs. Our protocol controller provides three different types of overhead tolerance: a) moving basic communication and coherence tasks away from computation pr...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Computer architecture news 1996-01, Vol.24 (Special Issu), p.198-209
Hauptverfasser: Bianchini, R, Kontothanassis, L I, Pinto, R, De Maria, M, Abud, M, Amorim, C L
Format: Artikel
Sprache:eng
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 209
container_issue Special Issu
container_start_page 198
container_title Computer architecture news
container_volume 24
creator Bianchini, R
Kontothanassis, L I
Pinto, R
De Maria, M
Abud, M
Amorim, C L
description In this paper we propose the use of a PCI-based programmable protocol controller for hiding communication and coherence overheads in software DSMs. Our protocol controller provides three different types of overhead tolerance: a) moving basic communication and coherence tasks away from computation processors; b) prefetching of diffs; and c) generating and applying diffs with hardware assistance. We evaluate the isolated and combined impact of these features on the performance of TreadMarks. We also compare performance against two versions of the Shrimp-based AURC protocol. Using detailed execution-driven simulations of a 16-node network of workstations, we show that the greatest performance benefits provided by our protocol controller come from our hardware-supported diffs. Reducing the burden of communication and coherence transactions on the computation processor is also beneficial but to a smaller extent. Prefetching is not always profitable. Our results show that our protocol controller can improve running time performance by up to 50% for TreadMarks, which means that it can double the TreadMarks speedups. The overlapping implementation of TreadMarks performs as well or better than AURC for 5 of our 6 applications. We conclude that the simple hardware support we propose allows for the implementation of high-performance software DSMs at low cost. Based on this conclusion, we are building the NCP sub(2) parallel system at COPPE/UFRJ.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_miscellaneous_26217609</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>26217609</sourcerecordid><originalsourceid>FETCH-proquest_miscellaneous_262176093</originalsourceid><addsrcrecordid>eNqNirsOgjAUQDtoIj7-oZMbSaFawiwaFifdyU25SE1ptbdo_HsZ_ACnk5NzZiwRmZLpvlS7BVsS3cXkhRQJq2rTGnfj2g_D6IyGaLzjFiI6_eHg2qn0GCZD7l8YeoSWG8fJd_ENAXl1OdOazTuwhJsfV2x7Ol4PdfoI_jkixWYwpNFacOhHanKVZ4USpfx7_AJ6kD1H</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>26217609</pqid></control><display><type>article</type><title>Hiding communication latency and coherence overhead in software DSMs</title><source>Access via ACM Digital Library</source><creator>Bianchini, R ; Kontothanassis, L I ; Pinto, R ; De Maria, M ; Abud, M ; Amorim, C L</creator><creatorcontrib>Bianchini, R ; Kontothanassis, L I ; Pinto, R ; De Maria, M ; Abud, M ; Amorim, C L</creatorcontrib><description>In this paper we propose the use of a PCI-based programmable protocol controller for hiding communication and coherence overheads in software DSMs. Our protocol controller provides three different types of overhead tolerance: a) moving basic communication and coherence tasks away from computation processors; b) prefetching of diffs; and c) generating and applying diffs with hardware assistance. We evaluate the isolated and combined impact of these features on the performance of TreadMarks. We also compare performance against two versions of the Shrimp-based AURC protocol. Using detailed execution-driven simulations of a 16-node network of workstations, we show that the greatest performance benefits provided by our protocol controller come from our hardware-supported diffs. Reducing the burden of communication and coherence transactions on the computation processor is also beneficial but to a smaller extent. Prefetching is not always profitable. Our results show that our protocol controller can improve running time performance by up to 50% for TreadMarks, which means that it can double the TreadMarks speedups. The overlapping implementation of TreadMarks performs as well or better than AURC for 5 of our 6 applications. We conclude that the simple hardware support we propose allows for the implementation of high-performance software DSMs at low cost. Based on this conclusion, we are building the NCP sub(2) parallel system at COPPE/UFRJ.</description><identifier>ISSN: 0163-5964</identifier><language>eng</language><ispartof>Computer architecture news, 1996-01, Vol.24 (Special Issu), p.198-209</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784</link.rule.ids></links><search><creatorcontrib>Bianchini, R</creatorcontrib><creatorcontrib>Kontothanassis, L I</creatorcontrib><creatorcontrib>Pinto, R</creatorcontrib><creatorcontrib>De Maria, M</creatorcontrib><creatorcontrib>Abud, M</creatorcontrib><creatorcontrib>Amorim, C L</creatorcontrib><title>Hiding communication latency and coherence overhead in software DSMs</title><title>Computer architecture news</title><description>In this paper we propose the use of a PCI-based programmable protocol controller for hiding communication and coherence overheads in software DSMs. Our protocol controller provides three different types of overhead tolerance: a) moving basic communication and coherence tasks away from computation processors; b) prefetching of diffs; and c) generating and applying diffs with hardware assistance. We evaluate the isolated and combined impact of these features on the performance of TreadMarks. We also compare performance against two versions of the Shrimp-based AURC protocol. Using detailed execution-driven simulations of a 16-node network of workstations, we show that the greatest performance benefits provided by our protocol controller come from our hardware-supported diffs. Reducing the burden of communication and coherence transactions on the computation processor is also beneficial but to a smaller extent. Prefetching is not always profitable. Our results show that our protocol controller can improve running time performance by up to 50% for TreadMarks, which means that it can double the TreadMarks speedups. The overlapping implementation of TreadMarks performs as well or better than AURC for 5 of our 6 applications. We conclude that the simple hardware support we propose allows for the implementation of high-performance software DSMs at low cost. Based on this conclusion, we are building the NCP sub(2) parallel system at COPPE/UFRJ.</description><issn>0163-5964</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>1996</creationdate><recordtype>article</recordtype><recordid>eNqNirsOgjAUQDtoIj7-oZMbSaFawiwaFifdyU25SE1ptbdo_HsZ_ACnk5NzZiwRmZLpvlS7BVsS3cXkhRQJq2rTGnfj2g_D6IyGaLzjFiI6_eHg2qn0GCZD7l8YeoSWG8fJd_ENAXl1OdOazTuwhJsfV2x7Ol4PdfoI_jkixWYwpNFacOhHanKVZ4USpfx7_AJ6kD1H</recordid><startdate>19960101</startdate><enddate>19960101</enddate><creator>Bianchini, R</creator><creator>Kontothanassis, L I</creator><creator>Pinto, R</creator><creator>De Maria, M</creator><creator>Abud, M</creator><creator>Amorim, C L</creator><scope>7SC</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope></search><sort><creationdate>19960101</creationdate><title>Hiding communication latency and coherence overhead in software DSMs</title><author>Bianchini, R ; Kontothanassis, L I ; Pinto, R ; De Maria, M ; Abud, M ; Amorim, C L</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_miscellaneous_262176093</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>1996</creationdate><toplevel>online_resources</toplevel><creatorcontrib>Bianchini, R</creatorcontrib><creatorcontrib>Kontothanassis, L I</creatorcontrib><creatorcontrib>Pinto, R</creatorcontrib><creatorcontrib>De Maria, M</creatorcontrib><creatorcontrib>Abud, M</creatorcontrib><creatorcontrib>Amorim, C L</creatorcontrib><collection>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>Computer architecture news</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Bianchini, R</au><au>Kontothanassis, L I</au><au>Pinto, R</au><au>De Maria, M</au><au>Abud, M</au><au>Amorim, C L</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Hiding communication latency and coherence overhead in software DSMs</atitle><jtitle>Computer architecture news</jtitle><date>1996-01-01</date><risdate>1996</risdate><volume>24</volume><issue>Special Issu</issue><spage>198</spage><epage>209</epage><pages>198-209</pages><issn>0163-5964</issn><abstract>In this paper we propose the use of a PCI-based programmable protocol controller for hiding communication and coherence overheads in software DSMs. Our protocol controller provides three different types of overhead tolerance: a) moving basic communication and coherence tasks away from computation processors; b) prefetching of diffs; and c) generating and applying diffs with hardware assistance. We evaluate the isolated and combined impact of these features on the performance of TreadMarks. We also compare performance against two versions of the Shrimp-based AURC protocol. Using detailed execution-driven simulations of a 16-node network of workstations, we show that the greatest performance benefits provided by our protocol controller come from our hardware-supported diffs. Reducing the burden of communication and coherence transactions on the computation processor is also beneficial but to a smaller extent. Prefetching is not always profitable. Our results show that our protocol controller can improve running time performance by up to 50% for TreadMarks, which means that it can double the TreadMarks speedups. The overlapping implementation of TreadMarks performs as well or better than AURC for 5 of our 6 applications. We conclude that the simple hardware support we propose allows for the implementation of high-performance software DSMs at low cost. Based on this conclusion, we are building the NCP sub(2) parallel system at COPPE/UFRJ.</abstract></addata></record>
fulltext fulltext
identifier ISSN: 0163-5964
ispartof Computer architecture news, 1996-01, Vol.24 (Special Issu), p.198-209
issn 0163-5964
language eng
recordid cdi_proquest_miscellaneous_26217609
source Access via ACM Digital Library
title Hiding communication latency and coherence overhead in software DSMs
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-22T09%3A00%3A47IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Hiding%20communication%20latency%20and%20coherence%20overhead%20in%20software%20DSMs&rft.jtitle=Computer%20architecture%20news&rft.au=Bianchini,%20R&rft.date=1996-01-01&rft.volume=24&rft.issue=Special%20Issu&rft.spage=198&rft.epage=209&rft.pages=198-209&rft.issn=0163-5964&rft_id=info:doi/&rft_dat=%3Cproquest%3E26217609%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=26217609&rft_id=info:pmid/&rfr_iscdi=true