CXL Memory as Persistent Memory for Disaggregated HPC: A Practical Approach

In the landscape of High-Performance Computing (HPC), the quest for efficient and scalable memory solutions remains paramount. The advent of Compute Express Link (CXL) introduces a promising avenue with its potential to function as a Persistent Memory (PMem) solution in the context of disaggregated...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Fridman, Yehonatan, Desai, Suprasad Mutalik, Singh, Navneet, Willhalm, Thomas, Oren, Gal
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 Fridman, Yehonatan
Desai, Suprasad Mutalik
Singh, Navneet
Willhalm, Thomas
Oren, Gal
description In the landscape of High-Performance Computing (HPC), the quest for efficient and scalable memory solutions remains paramount. The advent of Compute Express Link (CXL) introduces a promising avenue with its potential to function as a Persistent Memory (PMem) solution in the context of disaggregated HPC systems. This paper presents a comprehensive exploration of CXL memory's viability as a candidate for PMem, supported by physical experiments conducted on cutting-edge multi-NUMA nodes equipped with CXL-attached memory prototypes. Our study not only benchmarks the performance of CXL memory but also illustrates the seamless transition from traditional PMem programming models to CXL, reinforcing its practicality. To substantiate our claims, we establish a tangible CXL prototype using an FPGA card embodying CXL 1.1/2.0 compliant endpoint designs (Intel FPGA CXL IP). Performance evaluations, executed through the STREAM and STREAM-PMem benchmarks, showcase CXL memory's ability to mirror PMem characteristics in App-Direct and Memory Mode while achieving impressive bandwidth metrics with Intel 4th generation Xeon (Sapphire Rapids) processors. The results elucidate the feasibility of CXL memory as a persistent memory solution, outperforming previously established benchmarks. In contrast to published DCPMM results, our CXL-DDR4 memory module offers comparable bandwidth to local DDR4 memory configurations, albeit with a moderate decrease in performance. The modified STREAM-PMem application underscores the ease of transitioning programming models from PMem to CXL, thus underscoring the practicality of adopting CXL memory.
doi_str_mv 10.48550/arxiv.2308.10714
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2308_10714</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2308_10714</sourcerecordid><originalsourceid>FETCH-LOGICAL-a674-e0cdf14ebdad5ae3317d725e7245e98c7382a2d384de063c14e3fb2ff016441d3</originalsourceid><addsrcrecordid>eNo1z71OwzAUBWAvHVDpAzDhF0jwb2zYolAoIogMHdiiW_s6WGqbyI4QfXugwHSko6MjfYRccVYqqzW7gfQZP0ohmS05M1xdkOfmraUveBjTiUKmHaYc84zH-b8MY6L3McMwJBxgRk83XXNHa9olcHN0sKf1NKUR3PslWQTYZ1z95ZJsH9bbZlO0r49PTd0WUBlVIHM-cIU7D14DSsmNN0KjEUrjrXVGWgHCS6s8skq676kMOxEC45VS3Msluf69PWv6KcUDpFP_o-rPKvkF4wxGww</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>CXL Memory as Persistent Memory for Disaggregated HPC: A Practical Approach</title><source>arXiv.org</source><creator>Fridman, Yehonatan ; Desai, Suprasad Mutalik ; Singh, Navneet ; Willhalm, Thomas ; Oren, Gal</creator><creatorcontrib>Fridman, Yehonatan ; Desai, Suprasad Mutalik ; Singh, Navneet ; Willhalm, Thomas ; Oren, Gal</creatorcontrib><description>In the landscape of High-Performance Computing (HPC), the quest for efficient and scalable memory solutions remains paramount. The advent of Compute Express Link (CXL) introduces a promising avenue with its potential to function as a Persistent Memory (PMem) solution in the context of disaggregated HPC systems. This paper presents a comprehensive exploration of CXL memory's viability as a candidate for PMem, supported by physical experiments conducted on cutting-edge multi-NUMA nodes equipped with CXL-attached memory prototypes. Our study not only benchmarks the performance of CXL memory but also illustrates the seamless transition from traditional PMem programming models to CXL, reinforcing its practicality. To substantiate our claims, we establish a tangible CXL prototype using an FPGA card embodying CXL 1.1/2.0 compliant endpoint designs (Intel FPGA CXL IP). Performance evaluations, executed through the STREAM and STREAM-PMem benchmarks, showcase CXL memory's ability to mirror PMem characteristics in App-Direct and Memory Mode while achieving impressive bandwidth metrics with Intel 4th generation Xeon (Sapphire Rapids) processors. The results elucidate the feasibility of CXL memory as a persistent memory solution, outperforming previously established benchmarks. In contrast to published DCPMM results, our CXL-DDR4 memory module offers comparable bandwidth to local DDR4 memory configurations, albeit with a moderate decrease in performance. The modified STREAM-PMem application underscores the ease of transitioning programming models from PMem to CXL, thus underscoring the practicality of adopting CXL memory.</description><identifier>DOI: 10.48550/arxiv.2308.10714</identifier><language>eng</language><subject>Computer Science - Distributed, Parallel, and Cluster Computing</subject><creationdate>2023-08</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2308.10714$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2308.10714$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Fridman, Yehonatan</creatorcontrib><creatorcontrib>Desai, Suprasad Mutalik</creatorcontrib><creatorcontrib>Singh, Navneet</creatorcontrib><creatorcontrib>Willhalm, Thomas</creatorcontrib><creatorcontrib>Oren, Gal</creatorcontrib><title>CXL Memory as Persistent Memory for Disaggregated HPC: A Practical Approach</title><description>In the landscape of High-Performance Computing (HPC), the quest for efficient and scalable memory solutions remains paramount. The advent of Compute Express Link (CXL) introduces a promising avenue with its potential to function as a Persistent Memory (PMem) solution in the context of disaggregated HPC systems. This paper presents a comprehensive exploration of CXL memory's viability as a candidate for PMem, supported by physical experiments conducted on cutting-edge multi-NUMA nodes equipped with CXL-attached memory prototypes. Our study not only benchmarks the performance of CXL memory but also illustrates the seamless transition from traditional PMem programming models to CXL, reinforcing its practicality. To substantiate our claims, we establish a tangible CXL prototype using an FPGA card embodying CXL 1.1/2.0 compliant endpoint designs (Intel FPGA CXL IP). Performance evaluations, executed through the STREAM and STREAM-PMem benchmarks, showcase CXL memory's ability to mirror PMem characteristics in App-Direct and Memory Mode while achieving impressive bandwidth metrics with Intel 4th generation Xeon (Sapphire Rapids) processors. The results elucidate the feasibility of CXL memory as a persistent memory solution, outperforming previously established benchmarks. In contrast to published DCPMM results, our CXL-DDR4 memory module offers comparable bandwidth to local DDR4 memory configurations, albeit with a moderate decrease in performance. The modified STREAM-PMem application underscores the ease of transitioning programming models from PMem to CXL, thus underscoring the practicality of adopting CXL memory.</description><subject>Computer Science - Distributed, Parallel, and Cluster Computing</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNo1z71OwzAUBWAvHVDpAzDhF0jwb2zYolAoIogMHdiiW_s6WGqbyI4QfXugwHSko6MjfYRccVYqqzW7gfQZP0ohmS05M1xdkOfmraUveBjTiUKmHaYc84zH-b8MY6L3McMwJBxgRk83XXNHa9olcHN0sKf1NKUR3PslWQTYZ1z95ZJsH9bbZlO0r49PTd0WUBlVIHM-cIU7D14DSsmNN0KjEUrjrXVGWgHCS6s8skq676kMOxEC45VS3Msluf69PWv6KcUDpFP_o-rPKvkF4wxGww</recordid><startdate>20230821</startdate><enddate>20230821</enddate><creator>Fridman, Yehonatan</creator><creator>Desai, Suprasad Mutalik</creator><creator>Singh, Navneet</creator><creator>Willhalm, Thomas</creator><creator>Oren, Gal</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20230821</creationdate><title>CXL Memory as Persistent Memory for Disaggregated HPC: A Practical Approach</title><author>Fridman, Yehonatan ; Desai, Suprasad Mutalik ; Singh, Navneet ; Willhalm, Thomas ; Oren, Gal</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a674-e0cdf14ebdad5ae3317d725e7245e98c7382a2d384de063c14e3fb2ff016441d3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Distributed, Parallel, and Cluster Computing</topic><toplevel>online_resources</toplevel><creatorcontrib>Fridman, Yehonatan</creatorcontrib><creatorcontrib>Desai, Suprasad Mutalik</creatorcontrib><creatorcontrib>Singh, Navneet</creatorcontrib><creatorcontrib>Willhalm, Thomas</creatorcontrib><creatorcontrib>Oren, Gal</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Fridman, Yehonatan</au><au>Desai, Suprasad Mutalik</au><au>Singh, Navneet</au><au>Willhalm, Thomas</au><au>Oren, Gal</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>CXL Memory as Persistent Memory for Disaggregated HPC: A Practical Approach</atitle><date>2023-08-21</date><risdate>2023</risdate><abstract>In the landscape of High-Performance Computing (HPC), the quest for efficient and scalable memory solutions remains paramount. The advent of Compute Express Link (CXL) introduces a promising avenue with its potential to function as a Persistent Memory (PMem) solution in the context of disaggregated HPC systems. This paper presents a comprehensive exploration of CXL memory's viability as a candidate for PMem, supported by physical experiments conducted on cutting-edge multi-NUMA nodes equipped with CXL-attached memory prototypes. Our study not only benchmarks the performance of CXL memory but also illustrates the seamless transition from traditional PMem programming models to CXL, reinforcing its practicality. To substantiate our claims, we establish a tangible CXL prototype using an FPGA card embodying CXL 1.1/2.0 compliant endpoint designs (Intel FPGA CXL IP). Performance evaluations, executed through the STREAM and STREAM-PMem benchmarks, showcase CXL memory's ability to mirror PMem characteristics in App-Direct and Memory Mode while achieving impressive bandwidth metrics with Intel 4th generation Xeon (Sapphire Rapids) processors. The results elucidate the feasibility of CXL memory as a persistent memory solution, outperforming previously established benchmarks. In contrast to published DCPMM results, our CXL-DDR4 memory module offers comparable bandwidth to local DDR4 memory configurations, albeit with a moderate decrease in performance. The modified STREAM-PMem application underscores the ease of transitioning programming models from PMem to CXL, thus underscoring the practicality of adopting CXL memory.</abstract><doi>10.48550/arxiv.2308.10714</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2308.10714
ispartof
issn
language eng
recordid cdi_arxiv_primary_2308_10714
source arXiv.org
subjects Computer Science - Distributed, Parallel, and Cluster Computing
title CXL Memory as Persistent Memory for Disaggregated HPC: A Practical Approach
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-19T21%3A18%3A20IST&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=CXL%20Memory%20as%20Persistent%20Memory%20for%20Disaggregated%20HPC:%20A%20Practical%20Approach&rft.au=Fridman,%20Yehonatan&rft.date=2023-08-21&rft_id=info:doi/10.48550/arxiv.2308.10714&rft_dat=%3Carxiv_GOX%3E2308_10714%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