ACHIEVING BALANCED EXECUTION THROUGH RUNTIME DETECTION OF PERFORMANCE VARIATION

ACHIEVING BALANCED EXECUTION THROUGH RUNTIME DETECTION OF PERFORMANCE VARIATION

Systems, apparatuses, and methods for achieving balanced execution in a multi-node cluster through runtime detection of performance variation are described. During a training phase, performance counters and an amount of time spent waiting for synchronization is monitored for a plurality of tasks for...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Kocoloski Brian J, Piga Leonardo, Huang Wei, Paul Indrani
Format: Patent
Sprache:eng
Schlagworte:
ELECTRIC COMMUNICATION TECHNIQUE
ELECTRICITY
TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Kocoloski Brian J
Piga Leonardo
Huang Wei
Paul Indrani
description Systems, apparatuses, and methods for achieving balanced execution in a multi-node cluster through runtime detection of performance variation are described. During a training phase, performance counters and an amount of time spent waiting for synchronization is monitored for a plurality of tasks for each node of the multi-node cluster. These values are utilized to generate a model which correlates the values of the performance counters to the amount of time spent waiting for synchronization. Once the model is built, the values of the performance counters are monitored for a period of time at the start of each task, and these values are input into the model. The model generates a prediction of whether a given node is on the critical path. If the given node is predicted to be on the critical path, the power allocation of the given node is increased.
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_US2017373955A1</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>US2017373955A1</sourcerecordid><originalsourceid>FETCH-epo_espacenet_US2017373955A13</originalsourceid><addsrcrecordid>eNrjZPB3dPbwdA3z9HNXcHL0cfRzdnVRcI1wdQ4N8fT3UwjxCPIPdfdQCAr1C_H0dVVwcQ1xdQbL-LspBLgGufkH-YL0KIQ5Bnk6giR4GFjTEnOKU3mhNDeDsptriLOHbmpBfnxqcUFicmpeakl8aLCRgaG5sbmxpampo6ExcaoAGbEwAg</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>ACHIEVING BALANCED EXECUTION THROUGH RUNTIME DETECTION OF PERFORMANCE VARIATION</title><source>esp@cenet</source><creator>Kocoloski Brian J ; Piga Leonardo ; Huang Wei ; Paul Indrani</creator><creatorcontrib>Kocoloski Brian J ; Piga Leonardo ; Huang Wei ; Paul Indrani</creatorcontrib><description>Systems, apparatuses, and methods for achieving balanced execution in a multi-node cluster through runtime detection of performance variation are described. During a training phase, performance counters and an amount of time spent waiting for synchronization is monitored for a plurality of tasks for each node of the multi-node cluster. These values are utilized to generate a model which correlates the values of the performance counters to the amount of time spent waiting for synchronization. Once the model is built, the values of the performance counters are monitored for a period of time at the start of each task, and these values are input into the model. The model generates a prediction of whether a given node is on the critical path. If the given node is predicted to be on the critical path, the power allocation of the given node is increased.</description><language>eng</language><subject>ELECTRIC COMMUNICATION TECHNIQUE ; ELECTRICITY ; TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION</subject><creationdate>2017</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20171228&amp;DB=EPODOC&amp;CC=US&amp;NR=2017373955A1$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,777,882,25545,76296</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20171228&amp;DB=EPODOC&amp;CC=US&amp;NR=2017373955A1$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>Kocoloski Brian J</creatorcontrib><creatorcontrib>Piga Leonardo</creatorcontrib><creatorcontrib>Huang Wei</creatorcontrib><creatorcontrib>Paul Indrani</creatorcontrib><title>ACHIEVING BALANCED EXECUTION THROUGH RUNTIME DETECTION OF PERFORMANCE VARIATION</title><description>Systems, apparatuses, and methods for achieving balanced execution in a multi-node cluster through runtime detection of performance variation are described. During a training phase, performance counters and an amount of time spent waiting for synchronization is monitored for a plurality of tasks for each node of the multi-node cluster. These values are utilized to generate a model which correlates the values of the performance counters to the amount of time spent waiting for synchronization. Once the model is built, the values of the performance counters are monitored for a period of time at the start of each task, and these values are input into the model. The model generates a prediction of whether a given node is on the critical path. If the given node is predicted to be on the critical path, the power allocation of the given node is increased.</description><subject>ELECTRIC COMMUNICATION TECHNIQUE</subject><subject>ELECTRICITY</subject><subject>TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2017</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZPB3dPbwdA3z9HNXcHL0cfRzdnVRcI1wdQ4N8fT3UwjxCPIPdfdQCAr1C_H0dVVwcQ1xdQbL-LspBLgGufkH-YL0KIQ5Bnk6giR4GFjTEnOKU3mhNDeDsptriLOHbmpBfnxqcUFicmpeakl8aLCRgaG5sbmxpampo6ExcaoAGbEwAg</recordid><startdate>20171228</startdate><enddate>20171228</enddate><creator>Kocoloski Brian J</creator><creator>Piga Leonardo</creator><creator>Huang Wei</creator><creator>Paul Indrani</creator><scope>EVB</scope></search><sort><creationdate>20171228</creationdate><title>ACHIEVING BALANCED EXECUTION THROUGH RUNTIME DETECTION OF PERFORMANCE VARIATION</title><author>Kocoloski Brian J ; Piga Leonardo ; Huang Wei ; Paul Indrani</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_US2017373955A13</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>eng</language><creationdate>2017</creationdate><topic>ELECTRIC COMMUNICATION TECHNIQUE</topic><topic>ELECTRICITY</topic><topic>TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION</topic><toplevel>online_resources</toplevel><creatorcontrib>Kocoloski Brian J</creatorcontrib><creatorcontrib>Piga Leonardo</creatorcontrib><creatorcontrib>Huang Wei</creatorcontrib><creatorcontrib>Paul Indrani</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Kocoloski Brian J</au><au>Piga Leonardo</au><au>Huang Wei</au><au>Paul Indrani</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>ACHIEVING BALANCED EXECUTION THROUGH RUNTIME DETECTION OF PERFORMANCE VARIATION</title><date>2017-12-28</date><risdate>2017</risdate><abstract>Systems, apparatuses, and methods for achieving balanced execution in a multi-node cluster through runtime detection of performance variation are described. During a training phase, performance counters and an amount of time spent waiting for synchronization is monitored for a plurality of tasks for each node of the multi-node cluster. These values are utilized to generate a model which correlates the values of the performance counters to the amount of time spent waiting for synchronization. Once the model is built, the values of the performance counters are monitored for a period of time at the start of each task, and these values are input into the model. The model generates a prediction of whether a given node is on the critical path. If the given node is predicted to be on the critical path, the power allocation of the given node is increased.</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language eng
recordid cdi_epo_espacenet_US2017373955A1
source esp@cenet
subjects ELECTRIC COMMUNICATION TECHNIQUE
ELECTRICITY
TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION
title ACHIEVING BALANCED EXECUTION THROUGH RUNTIME DETECTION OF PERFORMANCE VARIATION
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-18T02%3A52%3A26IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=Kocoloski%20Brian%20J&rft.date=2017-12-28&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3EUS2017373955A1%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true