Low latency optical switch for high performance computing with minimized processor energy load [invited]
Power density and cooling issues are limiting the performance of high performance chip multiprocessors (CMPs), and off-chip communications currently consume more than 20% of power for memory, coherence, PCI, and Ethernet links. Photonic transceivers integrated with CMPs are being developed to overco...
Gespeichert in:
Veröffentlicht in: | Journal of optical communications and networking 2015-03, Vol.7 (3), p.A498-A510 |
---|---|
Hauptverfasser: | , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | A510 |
---|---|
container_issue | 3 |
container_start_page | A498 |
container_title | Journal of optical communications and networking |
container_volume | 7 |
creator | Liu, Shiyun Cheng, Qixiang Madarbux, Muhammad Ridwan Wonfor, Adrian Penty, Richard V. White, Ian H. Watts, Philip M. |
description | Power density and cooling issues are limiting the performance of high performance chip multiprocessors (CMPs), and off-chip communications currently consume more than 20% of power for memory, coherence, PCI, and Ethernet links. Photonic transceivers integrated with CMPs are being developed to overcome these issues, potentially allowing low hop count switched connections between chips or data center servers. However, latency in setting up optical connections is critically important in all computing applications, and having transceivers integrated on the processor chip also pushes other network functions and their associated power consumption onto the chip. In this paper, we propose a low latency optical switch architecture that minimizes the power consumed on the processor chip for two scenarios: multiple-socket shared memory coherence networks and optical top-of-rack switches for data centers. The switch architecture reduces power consumed on the CMP using a control plane with a simplified send and forget server interface and the use of a hybrid Mach-Zehnder interferometer and semiconductor optical amplifier integrated optical switch with electronic buffering. Results show that the proposed architecture offers a 42% reduction in head latency at low loads compared with a conventional scheduled optical switch as well as offering increased performance for streaming and incast traffic patterns. Power dissipated on the server chip is shown to be reduced by more than 60% compared with a scheduled optical switch architecture with ring resonator switching. |
doi_str_mv | 10.1364/JOCN.7.00A498 |
format | Article |
fullrecord | <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_crossref_primary_10_1364_JOCN_7_00A498</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>7063694</ieee_id><sourcerecordid>3635773351</sourcerecordid><originalsourceid>FETCH-LOGICAL-c324t-a77928866a9979a372667427ef06b1c55d6a1ef14d3fee46b7c87db957decfcc3</originalsourceid><addsrcrecordid>eNo9kD1PwzAQhi0EEqUwMrFYYk6x82HHY1XxqYouMCFkuc6lcZXEwXapyq_HVRDTvcNz750ehK4pmdGM5Xcvq8XrjM8ImeeiPEETKvIsISwTp_85JefowvstIYxTWkxQs7R73KoAvT5gOwSjVYv93gTd4No63JhNgwdwMXeq14C17YZdMP0GR6jBnelNZ36gwoOzGryPO9CD2xxwa1WFP0z_bQJUn5forFath6u_OUXvD_dvi6dkuXp8XsyXic7SPCSKc5GWJWNKCC5UxlPGeJ5yqAlbU10UFVMUappXWQ2QszXXJa_WouAV6FrrbIpux974z9cOfJBbu3N9PCkpi11UkCKNVDJS2lnvHdRycKZT7iApkUeZ8ihTcjnKjPzNyBsA-Gd5lMui2F-ZFHJG</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>1666719052</pqid></control><display><type>article</type><title>Low latency optical switch for high performance computing with minimized processor energy load [invited]</title><source>IEEE Electronic Library (IEL)</source><creator>Liu, Shiyun ; Cheng, Qixiang ; Madarbux, Muhammad Ridwan ; Wonfor, Adrian ; Penty, Richard V. ; White, Ian H. ; Watts, Philip M.</creator><creatorcontrib>Liu, Shiyun ; Cheng, Qixiang ; Madarbux, Muhammad Ridwan ; Wonfor, Adrian ; Penty, Richard V. ; White, Ian H. ; Watts, Philip M.</creatorcontrib><description>Power density and cooling issues are limiting the performance of high performance chip multiprocessors (CMPs), and off-chip communications currently consume more than 20% of power for memory, coherence, PCI, and Ethernet links. Photonic transceivers integrated with CMPs are being developed to overcome these issues, potentially allowing low hop count switched connections between chips or data center servers. However, latency in setting up optical connections is critically important in all computing applications, and having transceivers integrated on the processor chip also pushes other network functions and their associated power consumption onto the chip. In this paper, we propose a low latency optical switch architecture that minimizes the power consumed on the processor chip for two scenarios: multiple-socket shared memory coherence networks and optical top-of-rack switches for data centers. The switch architecture reduces power consumed on the CMP using a control plane with a simplified send and forget server interface and the use of a hybrid Mach-Zehnder interferometer and semiconductor optical amplifier integrated optical switch with electronic buffering. Results show that the proposed architecture offers a 42% reduction in head latency at low loads compared with a conventional scheduled optical switch as well as offering increased performance for streaming and incast traffic patterns. Power dissipated on the server chip is shown to be reduced by more than 60% compared with a scheduled optical switch architecture with ring resonator switching.</description><identifier>ISSN: 1943-0620</identifier><identifier>EISSN: 1943-0639</identifier><identifier>DOI: 10.1364/JOCN.7.00A498</identifier><identifier>CODEN: JOCNBB</identifier><language>eng</language><publisher>Piscataway: Optica Publishing Group</publisher><subject>Assignment and routing algorithms ; Fiber optic communications ; Networks ; Optical buffering ; Optical interconnects ; Optical switches ; Optical transmitters ; Ports (Computers) ; Resource management ; Semiconductors ; Servers</subject><ispartof>Journal of optical communications and networking, 2015-03, Vol.7 (3), p.A498-A510</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) Mar 2015</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c324t-a77928866a9979a372667427ef06b1c55d6a1ef14d3fee46b7c87db957decfcc3</citedby><cites>FETCH-LOGICAL-c324t-a77928866a9979a372667427ef06b1c55d6a1ef14d3fee46b7c87db957decfcc3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/7063694$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,778,782,794,27907,27908,54741</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/7063694$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Liu, Shiyun</creatorcontrib><creatorcontrib>Cheng, Qixiang</creatorcontrib><creatorcontrib>Madarbux, Muhammad Ridwan</creatorcontrib><creatorcontrib>Wonfor, Adrian</creatorcontrib><creatorcontrib>Penty, Richard V.</creatorcontrib><creatorcontrib>White, Ian H.</creatorcontrib><creatorcontrib>Watts, Philip M.</creatorcontrib><title>Low latency optical switch for high performance computing with minimized processor energy load [invited]</title><title>Journal of optical communications and networking</title><addtitle>jocn</addtitle><description>Power density and cooling issues are limiting the performance of high performance chip multiprocessors (CMPs), and off-chip communications currently consume more than 20% of power for memory, coherence, PCI, and Ethernet links. Photonic transceivers integrated with CMPs are being developed to overcome these issues, potentially allowing low hop count switched connections between chips or data center servers. However, latency in setting up optical connections is critically important in all computing applications, and having transceivers integrated on the processor chip also pushes other network functions and their associated power consumption onto the chip. In this paper, we propose a low latency optical switch architecture that minimizes the power consumed on the processor chip for two scenarios: multiple-socket shared memory coherence networks and optical top-of-rack switches for data centers. The switch architecture reduces power consumed on the CMP using a control plane with a simplified send and forget server interface and the use of a hybrid Mach-Zehnder interferometer and semiconductor optical amplifier integrated optical switch with electronic buffering. Results show that the proposed architecture offers a 42% reduction in head latency at low loads compared with a conventional scheduled optical switch as well as offering increased performance for streaming and incast traffic patterns. Power dissipated on the server chip is shown to be reduced by more than 60% compared with a scheduled optical switch architecture with ring resonator switching.</description><subject>Assignment and routing algorithms</subject><subject>Fiber optic communications</subject><subject>Networks</subject><subject>Optical buffering</subject><subject>Optical interconnects</subject><subject>Optical switches</subject><subject>Optical transmitters</subject><subject>Ports (Computers)</subject><subject>Resource management</subject><subject>Semiconductors</subject><subject>Servers</subject><issn>1943-0620</issn><issn>1943-0639</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2015</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNo9kD1PwzAQhi0EEqUwMrFYYk6x82HHY1XxqYouMCFkuc6lcZXEwXapyq_HVRDTvcNz750ehK4pmdGM5Xcvq8XrjM8ImeeiPEETKvIsISwTp_85JefowvstIYxTWkxQs7R73KoAvT5gOwSjVYv93gTd4No63JhNgwdwMXeq14C17YZdMP0GR6jBnelNZ36gwoOzGryPO9CD2xxwa1WFP0z_bQJUn5forFath6u_OUXvD_dvi6dkuXp8XsyXic7SPCSKc5GWJWNKCC5UxlPGeJ5yqAlbU10UFVMUappXWQ2QszXXJa_WouAV6FrrbIpux974z9cOfJBbu3N9PCkpi11UkCKNVDJS2lnvHdRycKZT7iApkUeZ8ihTcjnKjPzNyBsA-Gd5lMui2F-ZFHJG</recordid><startdate>201503</startdate><enddate>201503</enddate><creator>Liu, Shiyun</creator><creator>Cheng, Qixiang</creator><creator>Madarbux, Muhammad Ridwan</creator><creator>Wonfor, Adrian</creator><creator>Penty, Richard V.</creator><creator>White, Ian H.</creator><creator>Watts, Philip M.</creator><general>Optica Publishing Group</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>7SP</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope></search><sort><creationdate>201503</creationdate><title>Low latency optical switch for high performance computing with minimized processor energy load [invited]</title><author>Liu, Shiyun ; Cheng, Qixiang ; Madarbux, Muhammad Ridwan ; Wonfor, Adrian ; Penty, Richard V. ; White, Ian H. ; Watts, Philip M.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c324t-a77928866a9979a372667427ef06b1c55d6a1ef14d3fee46b7c87db957decfcc3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2015</creationdate><topic>Assignment and routing algorithms</topic><topic>Fiber optic communications</topic><topic>Networks</topic><topic>Optical buffering</topic><topic>Optical interconnects</topic><topic>Optical switches</topic><topic>Optical transmitters</topic><topic>Ports (Computers)</topic><topic>Resource management</topic><topic>Semiconductors</topic><topic>Servers</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Liu, Shiyun</creatorcontrib><creatorcontrib>Cheng, Qixiang</creatorcontrib><creatorcontrib>Madarbux, Muhammad Ridwan</creatorcontrib><creatorcontrib>Wonfor, Adrian</creatorcontrib><creatorcontrib>Penty, Richard V.</creatorcontrib><creatorcontrib>White, Ian H.</creatorcontrib><creatorcontrib>Watts, Philip M.</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Electronics & Communications 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>Journal of optical communications and networking</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Liu, Shiyun</au><au>Cheng, Qixiang</au><au>Madarbux, Muhammad Ridwan</au><au>Wonfor, Adrian</au><au>Penty, Richard V.</au><au>White, Ian H.</au><au>Watts, Philip M.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Low latency optical switch for high performance computing with minimized processor energy load [invited]</atitle><jtitle>Journal of optical communications and networking</jtitle><stitle>jocn</stitle><date>2015-03</date><risdate>2015</risdate><volume>7</volume><issue>3</issue><spage>A498</spage><epage>A510</epage><pages>A498-A510</pages><issn>1943-0620</issn><eissn>1943-0639</eissn><coden>JOCNBB</coden><abstract>Power density and cooling issues are limiting the performance of high performance chip multiprocessors (CMPs), and off-chip communications currently consume more than 20% of power for memory, coherence, PCI, and Ethernet links. Photonic transceivers integrated with CMPs are being developed to overcome these issues, potentially allowing low hop count switched connections between chips or data center servers. However, latency in setting up optical connections is critically important in all computing applications, and having transceivers integrated on the processor chip also pushes other network functions and their associated power consumption onto the chip. In this paper, we propose a low latency optical switch architecture that minimizes the power consumed on the processor chip for two scenarios: multiple-socket shared memory coherence networks and optical top-of-rack switches for data centers. The switch architecture reduces power consumed on the CMP using a control plane with a simplified send and forget server interface and the use of a hybrid Mach-Zehnder interferometer and semiconductor optical amplifier integrated optical switch with electronic buffering. Results show that the proposed architecture offers a 42% reduction in head latency at low loads compared with a conventional scheduled optical switch as well as offering increased performance for streaming and incast traffic patterns. Power dissipated on the server chip is shown to be reduced by more than 60% compared with a scheduled optical switch architecture with ring resonator switching.</abstract><cop>Piscataway</cop><pub>Optica Publishing Group</pub><doi>10.1364/JOCN.7.00A498</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | ISSN: 1943-0620 |
ispartof | Journal of optical communications and networking, 2015-03, Vol.7 (3), p.A498-A510 |
issn | 1943-0620 1943-0639 |
language | eng |
recordid | cdi_crossref_primary_10_1364_JOCN_7_00A498 |
source | IEEE Electronic Library (IEL) |
subjects | Assignment and routing algorithms Fiber optic communications Networks Optical buffering Optical interconnects Optical switches Optical transmitters Ports (Computers) Resource management Semiconductors Servers |
title | Low latency optical switch for high performance computing with minimized processor energy load [invited] |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-16T18%3A00%3A43IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Low%20latency%20optical%20switch%20for%20high%20performance%20computing%20with%20minimized%20processor%20energy%20load%20%5Binvited%5D&rft.jtitle=Journal%20of%20optical%20communications%20and%20networking&rft.au=Liu,%20Shiyun&rft.date=2015-03&rft.volume=7&rft.issue=3&rft.spage=A498&rft.epage=A510&rft.pages=A498-A510&rft.issn=1943-0620&rft.eissn=1943-0639&rft.coden=JOCNBB&rft_id=info:doi/10.1364/JOCN.7.00A498&rft_dat=%3Cproquest_RIE%3E3635773351%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=1666719052&rft_id=info:pmid/&rft_ieee_id=7063694&rfr_iscdi=true |