Extending magny-cours cache coherence
One cost-effective way to meet the increasing demand for larger high-performance shared-memory servers is to build clusters with off-the-shelf processors connected with low-latency point-to-point interconnections like HyperTransport. Unfortunately, HyperTransport addressing limitations prevent build...
Published in: | IEEE Transactions on Computers |
---|---|
Main Authors: | , , , , , , , |
Other Authors: | , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2012
|
Subjects: | |
Online Access: | http://hdl.handle.net/10251/36257 https://doi.org/10.1109/TC.2011.65 |
id |
ftunivpvalencia:oai:riunet.upv.es:10251/36257 |
---|---|
record_format |
openpolar |
spelling |
ftunivpvalencia:oai:riunet.upv.es:10251/36257 2023-05-15T18:32:44+02:00 Extending magny-cours cache coherence Ros Bardisa, Alberto Cuesta Sáez, Blas Antonio Fernández-Pascual, Ricardo Gómez Requena, María Engracia Acacio Sánchez, Manuel E. Robles Martínez, Antonio García Carrasco, José Manuel Duato Marín, José Francisco Universitat Politècnica de València. Departamento de Informática de Sistemas y Computadores - Departament d'Informàtica de Sistemes i Computadors Ministerio de Educación y Ciencia Generalitat Valenciana Ministerio de Ciencia e Innovación 2012-05 http://hdl.handle.net/10251/36257 https://doi.org/10.1109/TC.2011.65 eng eng Institute of Electrical and Electronics Engineers (IEEE) IEEE Transactions on Computers info:eu-repo/grantAgreement/MICINN//TIN2009-14475-C04-01/ES/Arquitecturas De Servidores, Aplicaciones Y Servicios/ info:eu-repo/grantAgreement/Generalitat Valenciana//PROMETEO08%2F2008%2F060/ES/Extensión de la tecnología de red hypertransport para la mejora de la escalabilidad de los servidores de internet/ info:eu-repo/grantAgreement/MICINN//TIN2009-14475-C04-03/ES/Arquitectura de servidores, aplicaciones y servicios/ info:eu-repo/grantAgreement/MEC//CSD2006-00046/ES/High-performance, reliable architectures for data centers and Internet servers/ http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5740853 urn:issn:0018-9340 http://hdl.handle.net/10251/36257 doi:10.1109/TC.2011.65 http://rightsstatements.org/vocab/InC/1.0/ info:eu-repo/semantics/openAccess High-performance computing Cache coherence Coherence extension Directory protocol Scalability Shared memory Traffic filtering ARQUITECTURA Y TECNOLOGIA DE COMPUTADORES info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion 2012 ftunivpvalencia https://doi.org/10.1109/TC.2011.65 2022-06-12T20:32:11Z One cost-effective way to meet the increasing demand for larger high-performance shared-memory servers is to build clusters with off-the-shelf processors connected with low-latency point-to-point interconnections like HyperTransport. Unfortunately, HyperTransport addressing limitations prevent building systems with more than eight nodes. While the recent High-Node Count HyperTransport specification overcomes this limitation, recently launched twelve-core Magny-Cours processors have already inherited it and provide only 3 bits to encode the pointers used by the directory cache which they include to increase the scalability of their coherence protocol. In this work, we propose and develop an external device to extend the coherence domain of Magny-Cours processors beyond the 8-node limit while maintaining the advantages provided by the directory cache. Evaluation results for systems with up to 32 nodes show that the performance offered by our solution scales with the number of nodes, enhancing the directory cache effectiveness by filtering additional messages. Particularly, we reduce execution time by 47 percent in a 32-die system with respect to the 8-die Magny-Cours configuration. This work was supported by the Spanish MICINN, Consolider Programme and Plan E funds, as well as European Commission FEDER funds, under Grants CSD2006-00046 and TIN2009-14475-C04-01/03. It was also partly supported by (PROMETEO from Generalitat Valenciana (GVA) under Grant PROMETEO/2008/060). Ros Bardisa, A.; Cuesta Sáez, BA.; Fernández-Pascual, R.; Gómez Requena, ME.; Acacio Sánchez, ME.; Robles Martínez, A.; García Carrasco, JM. (2012). Extending magny-cours cache coherence. IEEE Transactions on Computers. 61(5):593-606. https://doi.org/10.1109/TC.2011.65 S 593 606 61 5 Article in Journal/Newspaper The Pointers Politechnical University of Valencia: RiuNet Carrasco ENVELOPE(-63.367,-63.367,-64.800,-64.800) Martínez ENVELOPE(-62.183,-62.183,-64.650,-64.650) Robles ENVELOPE(-61.450,-61.450,-64.367,-64.367) IEEE Transactions on Computers 61 5 593 606 |
institution |
Open Polar |
collection |
Politechnical University of Valencia: RiuNet |
op_collection_id |
ftunivpvalencia |
language |
English |
topic |
High-performance computing Cache coherence Coherence extension Directory protocol Scalability Shared memory Traffic filtering ARQUITECTURA Y TECNOLOGIA DE COMPUTADORES |
spellingShingle |
High-performance computing Cache coherence Coherence extension Directory protocol Scalability Shared memory Traffic filtering ARQUITECTURA Y TECNOLOGIA DE COMPUTADORES Ros Bardisa, Alberto Cuesta Sáez, Blas Antonio Fernández-Pascual, Ricardo Gómez Requena, María Engracia Acacio Sánchez, Manuel E. Robles Martínez, Antonio García Carrasco, José Manuel Duato Marín, José Francisco Extending magny-cours cache coherence |
topic_facet |
High-performance computing Cache coherence Coherence extension Directory protocol Scalability Shared memory Traffic filtering ARQUITECTURA Y TECNOLOGIA DE COMPUTADORES |
description |
One cost-effective way to meet the increasing demand for larger high-performance shared-memory servers is to build clusters with off-the-shelf processors connected with low-latency point-to-point interconnections like HyperTransport. Unfortunately, HyperTransport addressing limitations prevent building systems with more than eight nodes. While the recent High-Node Count HyperTransport specification overcomes this limitation, recently launched twelve-core Magny-Cours processors have already inherited it and provide only 3 bits to encode the pointers used by the directory cache which they include to increase the scalability of their coherence protocol. In this work, we propose and develop an external device to extend the coherence domain of Magny-Cours processors beyond the 8-node limit while maintaining the advantages provided by the directory cache. Evaluation results for systems with up to 32 nodes show that the performance offered by our solution scales with the number of nodes, enhancing the directory cache effectiveness by filtering additional messages. Particularly, we reduce execution time by 47 percent in a 32-die system with respect to the 8-die Magny-Cours configuration. This work was supported by the Spanish MICINN, Consolider Programme and Plan E funds, as well as European Commission FEDER funds, under Grants CSD2006-00046 and TIN2009-14475-C04-01/03. It was also partly supported by (PROMETEO from Generalitat Valenciana (GVA) under Grant PROMETEO/2008/060). Ros Bardisa, A.; Cuesta Sáez, BA.; Fernández-Pascual, R.; Gómez Requena, ME.; Acacio Sánchez, ME.; Robles Martínez, A.; García Carrasco, JM. (2012). Extending magny-cours cache coherence. IEEE Transactions on Computers. 61(5):593-606. https://doi.org/10.1109/TC.2011.65 S 593 606 61 5 |
author2 |
Universitat Politècnica de València. Departamento de Informática de Sistemas y Computadores - Departament d'Informàtica de Sistemes i Computadors Ministerio de Educación y Ciencia Generalitat Valenciana Ministerio de Ciencia e Innovación |
format |
Article in Journal/Newspaper |
author |
Ros Bardisa, Alberto Cuesta Sáez, Blas Antonio Fernández-Pascual, Ricardo Gómez Requena, María Engracia Acacio Sánchez, Manuel E. Robles Martínez, Antonio García Carrasco, José Manuel Duato Marín, José Francisco |
author_facet |
Ros Bardisa, Alberto Cuesta Sáez, Blas Antonio Fernández-Pascual, Ricardo Gómez Requena, María Engracia Acacio Sánchez, Manuel E. Robles Martínez, Antonio García Carrasco, José Manuel Duato Marín, José Francisco |
author_sort |
Ros Bardisa, Alberto |
title |
Extending magny-cours cache coherence |
title_short |
Extending magny-cours cache coherence |
title_full |
Extending magny-cours cache coherence |
title_fullStr |
Extending magny-cours cache coherence |
title_full_unstemmed |
Extending magny-cours cache coherence |
title_sort |
extending magny-cours cache coherence |
publisher |
Institute of Electrical and Electronics Engineers (IEEE) |
publishDate |
2012 |
url |
http://hdl.handle.net/10251/36257 https://doi.org/10.1109/TC.2011.65 |
long_lat |
ENVELOPE(-63.367,-63.367,-64.800,-64.800) ENVELOPE(-62.183,-62.183,-64.650,-64.650) ENVELOPE(-61.450,-61.450,-64.367,-64.367) |
geographic |
Carrasco Martínez Robles |
geographic_facet |
Carrasco Martínez Robles |
genre |
The Pointers |
genre_facet |
The Pointers |
op_relation |
IEEE Transactions on Computers info:eu-repo/grantAgreement/MICINN//TIN2009-14475-C04-01/ES/Arquitecturas De Servidores, Aplicaciones Y Servicios/ info:eu-repo/grantAgreement/Generalitat Valenciana//PROMETEO08%2F2008%2F060/ES/Extensión de la tecnología de red hypertransport para la mejora de la escalabilidad de los servidores de internet/ info:eu-repo/grantAgreement/MICINN//TIN2009-14475-C04-03/ES/Arquitectura de servidores, aplicaciones y servicios/ info:eu-repo/grantAgreement/MEC//CSD2006-00046/ES/High-performance, reliable architectures for data centers and Internet servers/ http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5740853 urn:issn:0018-9340 http://hdl.handle.net/10251/36257 doi:10.1109/TC.2011.65 |
op_rights |
http://rightsstatements.org/vocab/InC/1.0/ info:eu-repo/semantics/openAccess |
op_doi |
https://doi.org/10.1109/TC.2011.65 |
container_title |
IEEE Transactions on Computers |
container_volume |
61 |
container_issue |
5 |
container_start_page |
593 |
op_container_end_page |
606 |
_version_ |
1766216914499010560 |