COMPaS: A pentium Pro PC-based SMP cluster and its experience
We have built an eight node SMP cluster called COMPaS (Cluster Of Multi-Processor Systems), each node of which is a quadprocessor Pentium Pro PC. We have designed and implemented a remote memory based user-level communication layer which provides lowover-head and high bandwidth using Myrinet. We des...
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Tagungsbericht |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 497 |
|---|---|
| container_issue | |
| container_start_page | 486 |
| container_title | |
| container_volume | |
| creator | Tanaka, Yoshio Matsuda, Motohiko Ando, Makoto Kubota, Kazuto Sato, Mitsuhisa |
| description | We have built an eight node SMP cluster called COMPaS (Cluster Of Multi-Processor Systems), each node of which is a quadprocessor Pentium Pro PC. We have designed and implemented a remote memory based user-level communication layer which provides lowover-head and high bandwidth using Myrinet. We designed a hybrid programming model in order to take advantage of locality in each SMP node. Intra-node computations utilize a multi-threaded programming style (Solaris threads) and inter-node programming is based on message passing and remote memory operations. In this paper we report on this hybrid shared memory/distributed memory programming on COMPaS and its preliminary evaluation. The performance of COMPaS is affected by data size and access patterns, and the proportion of inter-node communication. If the data size is small enough to all fit on the cache, parallel efficiency exceeds 1.0 using the hybrid programming model on COMPaS. But the performance is limited by the low memory bus bandwidth of PC-based SMP nodes for some memory intensive workloads. |
| doi_str_mv | 10.1007/3-540-64359-1_722 |
| format | Conference Proceeding |
| fullrecord | <record><control><sourceid>pascalfrancis_sprin</sourceid><recordid>TN_cdi_pascalfrancis_primary_2040801</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2040801</sourcerecordid><originalsourceid>FETCH-LOGICAL-p224t-d768ed959ea9287819a4bcfa641a2993755b3bd070dee3aab2689af577fbfe0f3</originalsourceid><addsrcrecordid>eNotkMtqwzAURNUXNEnzAd1p0a2aq5dlFboIpi9IiCHtWlzbcnGbOEZyoP37OklXAzOHWRxCbjnccwAzk0wrYImS2jLujBBnZGpNKoc2sUYn-pyMeMI5k1LZCzI-DgeaX5IRSBDMGiWvyTjGLwAQxooRecxWyxzXD3ROO9_2zX5L87CjecYKjL6i62VOy80-9j5QbCva9JH6n86HxrelvyFXNW6in_7nhHw8P71nr2yxennL5gvWCaF6Vpkk9ZXV1qMVqUm5RVWUNSaKo7BWGq0LWVRgoPJeIhYiSS3W2pi6qD3UckLuTr8dxhI3dcC2bKLrQrPF8OsEKEiBD9jshMVhaT99cMVu9x0dB3cw6KQblLijE3c0KP8AbZlc9A</addsrcrecordid><sourcetype>Index Database</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>COMPaS: A pentium Pro PC-based SMP cluster and its experience</title><source>Springer Books</source><source>IEEE Electronic Library (IEL) Conference Proceedings</source><creator>Tanaka, Yoshio ; Matsuda, Motohiko ; Ando, Makoto ; Kubota, Kazuto ; Sato, Mitsuhisa</creator><contributor>Rolim, José</contributor><creatorcontrib>Tanaka, Yoshio ; Matsuda, Motohiko ; Ando, Makoto ; Kubota, Kazuto ; Sato, Mitsuhisa ; Rolim, José</creatorcontrib><description>We have built an eight node SMP cluster called COMPaS (Cluster Of Multi-Processor Systems), each node of which is a quadprocessor Pentium Pro PC. We have designed and implemented a remote memory based user-level communication layer which provides lowover-head and high bandwidth using Myrinet. We designed a hybrid programming model in order to take advantage of locality in each SMP node. Intra-node computations utilize a multi-threaded programming style (Solaris threads) and inter-node programming is based on message passing and remote memory operations. In this paper we report on this hybrid shared memory/distributed memory programming on COMPaS and its preliminary evaluation. The performance of COMPaS is affected by data size and access patterns, and the proportion of inter-node communication. If the data size is small enough to all fit on the cache, parallel efficiency exceeds 1.0 using the hybrid programming model on COMPaS. But the performance is limited by the low memory bus bandwidth of PC-based SMP nodes for some memory intensive workloads.</description><identifier>ISSN: 0302-9743</identifier><identifier>ISBN: 3540643591</identifier><identifier>ISBN: 9783540643593</identifier><identifier>EISSN: 1611-3349</identifier><identifier>EISBN: 9783540697565</identifier><identifier>EISBN: 354069756X</identifier><identifier>DOI: 10.1007/3-540-64359-1_722</identifier><language>eng</language><publisher>Berlin, Heidelberg: Springer Berlin Heidelberg</publisher><subject>Active Message ; Applied sciences ; Computer science; control theory; systems ; Computer systems performance. Reliability ; Exact sciences and technology ; Message Copying ; Message Passing ; Remote Memory ; Shared Memory System ; Software</subject><ispartof>Lecture notes in computer science, 1998, p.486-497</ispartof><rights>Springer-Verlag Berlin Heidelberg 1998</rights><rights>1998 INIST-CNRS</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/3-540-64359-1_722$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/3-540-64359-1_722$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>309,310,779,780,784,789,790,793,27925,38255,41442,42511</link.rule.ids><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=2040801$$DView record in Pascal Francis$$Hfree_for_read</backlink></links><search><contributor>Rolim, José</contributor><creatorcontrib>Tanaka, Yoshio</creatorcontrib><creatorcontrib>Matsuda, Motohiko</creatorcontrib><creatorcontrib>Ando, Makoto</creatorcontrib><creatorcontrib>Kubota, Kazuto</creatorcontrib><creatorcontrib>Sato, Mitsuhisa</creatorcontrib><title>COMPaS: A pentium Pro PC-based SMP cluster and its experience</title><title>Lecture notes in computer science</title><description>We have built an eight node SMP cluster called COMPaS (Cluster Of Multi-Processor Systems), each node of which is a quadprocessor Pentium Pro PC. We have designed and implemented a remote memory based user-level communication layer which provides lowover-head and high bandwidth using Myrinet. We designed a hybrid programming model in order to take advantage of locality in each SMP node. Intra-node computations utilize a multi-threaded programming style (Solaris threads) and inter-node programming is based on message passing and remote memory operations. In this paper we report on this hybrid shared memory/distributed memory programming on COMPaS and its preliminary evaluation. The performance of COMPaS is affected by data size and access patterns, and the proportion of inter-node communication. If the data size is small enough to all fit on the cache, parallel efficiency exceeds 1.0 using the hybrid programming model on COMPaS. But the performance is limited by the low memory bus bandwidth of PC-based SMP nodes for some memory intensive workloads.</description><subject>Active Message</subject><subject>Applied sciences</subject><subject>Computer science; control theory; systems</subject><subject>Computer systems performance. Reliability</subject><subject>Exact sciences and technology</subject><subject>Message Copying</subject><subject>Message Passing</subject><subject>Remote Memory</subject><subject>Shared Memory System</subject><subject>Software</subject><issn>0302-9743</issn><issn>1611-3349</issn><isbn>3540643591</isbn><isbn>9783540643593</isbn><isbn>9783540697565</isbn><isbn>354069756X</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>1998</creationdate><recordtype>conference_proceeding</recordtype><recordid>eNotkMtqwzAURNUXNEnzAd1p0a2aq5dlFboIpi9IiCHtWlzbcnGbOEZyoP37OklXAzOHWRxCbjnccwAzk0wrYImS2jLujBBnZGpNKoc2sUYn-pyMeMI5k1LZCzI-DgeaX5IRSBDMGiWvyTjGLwAQxooRecxWyxzXD3ROO9_2zX5L87CjecYKjL6i62VOy80-9j5QbCva9JH6n86HxrelvyFXNW6in_7nhHw8P71nr2yxennL5gvWCaF6Vpkk9ZXV1qMVqUm5RVWUNSaKo7BWGq0LWVRgoPJeIhYiSS3W2pi6qD3UckLuTr8dxhI3dcC2bKLrQrPF8OsEKEiBD9jshMVhaT99cMVu9x0dB3cw6KQblLijE3c0KP8AbZlc9A</recordid><startdate>19980101</startdate><enddate>19980101</enddate><creator>Tanaka, Yoshio</creator><creator>Matsuda, Motohiko</creator><creator>Ando, Makoto</creator><creator>Kubota, Kazuto</creator><creator>Sato, Mitsuhisa</creator><general>Springer Berlin Heidelberg</general><general>Springer-Verlag</general><scope>IQODW</scope></search><sort><creationdate>19980101</creationdate><title>COMPaS: A pentium Pro PC-based SMP cluster and its experience</title><author>Tanaka, Yoshio ; Matsuda, Motohiko ; Ando, Makoto ; Kubota, Kazuto ; Sato, Mitsuhisa</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-p224t-d768ed959ea9287819a4bcfa641a2993755b3bd070dee3aab2689af577fbfe0f3</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>1998</creationdate><topic>Active Message</topic><topic>Applied sciences</topic><topic>Computer science; control theory; systems</topic><topic>Computer systems performance. Reliability</topic><topic>Exact sciences and technology</topic><topic>Message Copying</topic><topic>Message Passing</topic><topic>Remote Memory</topic><topic>Shared Memory System</topic><topic>Software</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Tanaka, Yoshio</creatorcontrib><creatorcontrib>Matsuda, Motohiko</creatorcontrib><creatorcontrib>Ando, Makoto</creatorcontrib><creatorcontrib>Kubota, Kazuto</creatorcontrib><creatorcontrib>Sato, Mitsuhisa</creatorcontrib><collection>Pascal-Francis</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Tanaka, Yoshio</au><au>Matsuda, Motohiko</au><au>Ando, Makoto</au><au>Kubota, Kazuto</au><au>Sato, Mitsuhisa</au><au>Rolim, José</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>COMPaS: A pentium Pro PC-based SMP cluster and its experience</atitle><btitle>Lecture notes in computer science</btitle><date>1998-01-01</date><risdate>1998</risdate><spage>486</spage><epage>497</epage><pages>486-497</pages><issn>0302-9743</issn><eissn>1611-3349</eissn><isbn>3540643591</isbn><isbn>9783540643593</isbn><eisbn>9783540697565</eisbn><eisbn>354069756X</eisbn><abstract>We have built an eight node SMP cluster called COMPaS (Cluster Of Multi-Processor Systems), each node of which is a quadprocessor Pentium Pro PC. We have designed and implemented a remote memory based user-level communication layer which provides lowover-head and high bandwidth using Myrinet. We designed a hybrid programming model in order to take advantage of locality in each SMP node. Intra-node computations utilize a multi-threaded programming style (Solaris threads) and inter-node programming is based on message passing and remote memory operations. In this paper we report on this hybrid shared memory/distributed memory programming on COMPaS and its preliminary evaluation. The performance of COMPaS is affected by data size and access patterns, and the proportion of inter-node communication. If the data size is small enough to all fit on the cache, parallel efficiency exceeds 1.0 using the hybrid programming model on COMPaS. But the performance is limited by the low memory bus bandwidth of PC-based SMP nodes for some memory intensive workloads.</abstract><cop>Berlin, Heidelberg</cop><pub>Springer Berlin Heidelberg</pub><doi>10.1007/3-540-64359-1_722</doi><tpages>12</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 0302-9743 |
| ispartof | Lecture notes in computer science, 1998, p.486-497 |
| issn | 0302-9743 1611-3349 |
| language | eng |
| recordid | cdi_pascalfrancis_primary_2040801 |
| source | Springer Books; IEEE Electronic Library (IEL) Conference Proceedings |
| subjects | Active Message Applied sciences Computer science control theory systems Computer systems performance. Reliability Exact sciences and technology Message Copying Message Passing Remote Memory Shared Memory System Software |
| title | COMPaS: A pentium Pro PC-based SMP cluster and its experience |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-22T21%3A24%3A23IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-pascalfrancis_sprin&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=COMPaS:%20A%20pentium%20Pro%20PC-based%20SMP%20cluster%20and%20its%20experience&rft.btitle=Lecture%20notes%20in%20computer%20science&rft.au=Tanaka,%20Yoshio&rft.date=1998-01-01&rft.spage=486&rft.epage=497&rft.pages=486-497&rft.issn=0302-9743&rft.eissn=1611-3349&rft.isbn=3540643591&rft.isbn_list=9783540643593&rft_id=info:doi/10.1007/3-540-64359-1_722&rft_dat=%3Cpascalfrancis_sprin%3E2040801%3C/pascalfrancis_sprin%3E%3Curl%3E%3C/url%3E&rft.eisbn=9783540697565&rft.eisbn_list=354069756X&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |