Accurate and Scalable Many-Node Simulation

Accurate performance estimation of future many-node machines is challenging because it requires detailed simulation models of both node and network. However, simulating the full system in detail is unfeasible in terms of compute and memory resources. State-of-the-art techniques use a two-phase appro...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Eyerman, Stijn, Heirman, Wim, Bois, Kristof Du, Hur, Ibrahim
Format: Artikel
Sprache:eng
Schlagworte:
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 Eyerman, Stijn
Heirman, Wim
Bois, Kristof Du
Hur, Ibrahim
description Accurate performance estimation of future many-node machines is challenging because it requires detailed simulation models of both node and network. However, simulating the full system in detail is unfeasible in terms of compute and memory resources. State-of-the-art techniques use a two-phase approach that combines detailed simulation of a single node with network-only simulation of the full system. We show that these techniques, where the detailed node simulation is done in isolation, are inaccurate because they ignore two important node-level effects: compute time variability, and inter-node communication. We propose a novel three-stage simulation method to allow scalable and accurate many-node simulation, combining native profiling, detailed node simulation and high-level network simulation. By including timing variability and the impact of external nodes, our method leads to more accurate estimates. We validate our technique against measurements on a multi-node cluster, and report an average 6.7% error on 64 nodes (maximum error of 12%), compared to on average 27% error and up to 54% when timing variability and the scaling overhead are ignored. At higher node counts, the prediction error of ignoring variable timings and scaling overhead continues to increase compared to our technique, and may lead to selecting the wrong optimal cluster configuration. Using our technique, we are able to accurately project performance to thousands of nodes within a day of simulation time, using only a single or a few simulation hosts. Our method can be used to quickly explore large many-node design spaces, including node micro-architecture, node count and network configuration.
doi_str_mv 10.48550/arxiv.2401.09877
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2401_09877</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2401_09877</sourcerecordid><originalsourceid>FETCH-LOGICAL-a677-878066b8b0bdc663816aff4c02ae6383f954bb9c30e2ff7e86fec99feb0e72b13</originalsourceid><addsrcrecordid>eNotzrkOgkAUheFpLIz6AFZSm4AXBmYpjXFLXArsyZ3hTkKCaBCMvL2KVid_c_IxNg0hiFWSwALrV_EMohjCALSScsjmS2vbGhvysMq91GKJpiTviFXnn245eWlxbUtsils1ZgOH5YMm_x2xy2Z9We38w3m7Xy0PPgopfSUVCGGUAZNbIbgKBToXW4iQPsWdTmJjtOVAkXOSlHBktXZkgGRkQj5is99tj83udXHFusu-6KxH8zc3qTvJ</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Accurate and Scalable Many-Node Simulation</title><source>arXiv.org</source><creator>Eyerman, Stijn ; Heirman, Wim ; Bois, Kristof Du ; Hur, Ibrahim</creator><creatorcontrib>Eyerman, Stijn ; Heirman, Wim ; Bois, Kristof Du ; Hur, Ibrahim</creatorcontrib><description>Accurate performance estimation of future many-node machines is challenging because it requires detailed simulation models of both node and network. However, simulating the full system in detail is unfeasible in terms of compute and memory resources. State-of-the-art techniques use a two-phase approach that combines detailed simulation of a single node with network-only simulation of the full system. We show that these techniques, where the detailed node simulation is done in isolation, are inaccurate because they ignore two important node-level effects: compute time variability, and inter-node communication. We propose a novel three-stage simulation method to allow scalable and accurate many-node simulation, combining native profiling, detailed node simulation and high-level network simulation. By including timing variability and the impact of external nodes, our method leads to more accurate estimates. We validate our technique against measurements on a multi-node cluster, and report an average 6.7% error on 64 nodes (maximum error of 12%), compared to on average 27% error and up to 54% when timing variability and the scaling overhead are ignored. At higher node counts, the prediction error of ignoring variable timings and scaling overhead continues to increase compared to our technique, and may lead to selecting the wrong optimal cluster configuration. Using our technique, we are able to accurately project performance to thousands of nodes within a day of simulation time, using only a single or a few simulation hosts. Our method can be used to quickly explore large many-node design spaces, including node micro-architecture, node count and network configuration.</description><identifier>DOI: 10.48550/arxiv.2401.09877</identifier><language>eng</language><subject>Computer Science - Performance</subject><creationdate>2024-01</creationdate><rights>http://creativecommons.org/licenses/by-nc-nd/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2401.09877$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2401.09877$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Eyerman, Stijn</creatorcontrib><creatorcontrib>Heirman, Wim</creatorcontrib><creatorcontrib>Bois, Kristof Du</creatorcontrib><creatorcontrib>Hur, Ibrahim</creatorcontrib><title>Accurate and Scalable Many-Node Simulation</title><description>Accurate performance estimation of future many-node machines is challenging because it requires detailed simulation models of both node and network. However, simulating the full system in detail is unfeasible in terms of compute and memory resources. State-of-the-art techniques use a two-phase approach that combines detailed simulation of a single node with network-only simulation of the full system. We show that these techniques, where the detailed node simulation is done in isolation, are inaccurate because they ignore two important node-level effects: compute time variability, and inter-node communication. We propose a novel three-stage simulation method to allow scalable and accurate many-node simulation, combining native profiling, detailed node simulation and high-level network simulation. By including timing variability and the impact of external nodes, our method leads to more accurate estimates. We validate our technique against measurements on a multi-node cluster, and report an average 6.7% error on 64 nodes (maximum error of 12%), compared to on average 27% error and up to 54% when timing variability and the scaling overhead are ignored. At higher node counts, the prediction error of ignoring variable timings and scaling overhead continues to increase compared to our technique, and may lead to selecting the wrong optimal cluster configuration. Using our technique, we are able to accurately project performance to thousands of nodes within a day of simulation time, using only a single or a few simulation hosts. Our method can be used to quickly explore large many-node design spaces, including node micro-architecture, node count and network configuration.</description><subject>Computer Science - Performance</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotzrkOgkAUheFpLIz6AFZSm4AXBmYpjXFLXArsyZ3hTkKCaBCMvL2KVid_c_IxNg0hiFWSwALrV_EMohjCALSScsjmS2vbGhvysMq91GKJpiTviFXnn245eWlxbUtsils1ZgOH5YMm_x2xy2Z9We38w3m7Xy0PPgopfSUVCGGUAZNbIbgKBToXW4iQPsWdTmJjtOVAkXOSlHBktXZkgGRkQj5is99tj83udXHFusu-6KxH8zc3qTvJ</recordid><startdate>20240118</startdate><enddate>20240118</enddate><creator>Eyerman, Stijn</creator><creator>Heirman, Wim</creator><creator>Bois, Kristof Du</creator><creator>Hur, Ibrahim</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240118</creationdate><title>Accurate and Scalable Many-Node Simulation</title><author>Eyerman, Stijn ; Heirman, Wim ; Bois, Kristof Du ; Hur, Ibrahim</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a677-878066b8b0bdc663816aff4c02ae6383f954bb9c30e2ff7e86fec99feb0e72b13</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Performance</topic><toplevel>online_resources</toplevel><creatorcontrib>Eyerman, Stijn</creatorcontrib><creatorcontrib>Heirman, Wim</creatorcontrib><creatorcontrib>Bois, Kristof Du</creatorcontrib><creatorcontrib>Hur, Ibrahim</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Eyerman, Stijn</au><au>Heirman, Wim</au><au>Bois, Kristof Du</au><au>Hur, Ibrahim</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Accurate and Scalable Many-Node Simulation</atitle><date>2024-01-18</date><risdate>2024</risdate><abstract>Accurate performance estimation of future many-node machines is challenging because it requires detailed simulation models of both node and network. However, simulating the full system in detail is unfeasible in terms of compute and memory resources. State-of-the-art techniques use a two-phase approach that combines detailed simulation of a single node with network-only simulation of the full system. We show that these techniques, where the detailed node simulation is done in isolation, are inaccurate because they ignore two important node-level effects: compute time variability, and inter-node communication. We propose a novel three-stage simulation method to allow scalable and accurate many-node simulation, combining native profiling, detailed node simulation and high-level network simulation. By including timing variability and the impact of external nodes, our method leads to more accurate estimates. We validate our technique against measurements on a multi-node cluster, and report an average 6.7% error on 64 nodes (maximum error of 12%), compared to on average 27% error and up to 54% when timing variability and the scaling overhead are ignored. At higher node counts, the prediction error of ignoring variable timings and scaling overhead continues to increase compared to our technique, and may lead to selecting the wrong optimal cluster configuration. Using our technique, we are able to accurately project performance to thousands of nodes within a day of simulation time, using only a single or a few simulation hosts. Our method can be used to quickly explore large many-node design spaces, including node micro-architecture, node count and network configuration.</abstract><doi>10.48550/arxiv.2401.09877</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2401.09877
ispartof
issn
language eng
recordid cdi_arxiv_primary_2401_09877
source arXiv.org
subjects Computer Science - Performance
title Accurate and Scalable Many-Node Simulation
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-11T04%3A17%3A39IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Accurate%20and%20Scalable%20Many-Node%20Simulation&rft.au=Eyerman,%20Stijn&rft.date=2024-01-18&rft_id=info:doi/10.48550/arxiv.2401.09877&rft_dat=%3Carxiv_GOX%3E2401_09877%3C/arxiv_GOX%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