Mostra el registre d'ítem simple
Dynamic-vector execution on a general purpose EDGE chip multiprocessor
dc.contributor.author | Duric, Milovan |
dc.contributor.author | Palomar Pérez, Óscar |
dc.contributor.author | Smith, Aaron |
dc.contributor.author | Stanic, Milan |
dc.contributor.author | Unsal, Osman Sabri |
dc.contributor.author | Cristal Kestelman, Adrián |
dc.contributor.author | Valero Cortés, Mateo |
dc.contributor.author | Burger, Doug |
dc.contributor.author | Veidenbaum, Alexander V |
dc.contributor.other | Universitat Politècnica de Catalunya. Departament d'Arquitectura de Computadors |
dc.contributor.other | Barcelona Supercomputing Center |
dc.date.accessioned | 2015-05-06T12:37:41Z |
dc.date.created | 2014 |
dc.date.issued | 2014 |
dc.identifier.citation | Duric, M. [et al.]. Dynamic-vector execution on a general purpose EDGE chip multiprocessor. A: International Conference on Embedded Computer Systems: Architectures, Modeling and Simulation. "International Conference on Embedded Computer Systems: Architectures, Modeling and Simulation (SAMOS XIV): proceedings: July 14-17, 2014: Samos, Greece". Samos: Institute of Electrical and Electronics Engineers (IEEE), 2014, p. 18-25. |
dc.identifier.isbn | 978-1-4799-3770-7 |
dc.identifier.uri | http://hdl.handle.net/2117/27788 |
dc.description.abstract | This paper proposes a cost-effective technique that morphs the available cores of a low power chip multiprocessor (CMP) into an accelerator for data parallel (DLP) workloads. Instead of adding a special-purpose vector architecture as an accelerator, our technique leverages the resources of each CMP core to mimic the functionality of a vector processor. The morphing provides dynamic vector execution (DVX) on a general purpose CMP, by adding minimal hardware for vector control. DVX enhances the vector execution by dynamically configuring the allocation of compute and memory resources to match particular workload requirements. As an energy efficient substrate, we utilize modest dual issue cores based on an Explicit Data Graph Execution (EDGE) architecture. The results show that a DVX enabled 4-core EDGE CMP improves the energy-delay product over 14x, at the cost of only 1.1% of additional area. We compare DVX against a CMP that adds a dedicated DLP accelerator based on a conventional high performance vector design. The vector accelerator increases the area footprint over 74%, which greatly affects the cost of the modest processor. DVX avoids the additional costs and yet gains over 86% of the speedup obtained with the dedicated accelerator. |
dc.description.sponsorship | This work has been partially funded by the Spanish Government (TIN2012-34557), the European Research Council under the European Unions 7th FP (FP/2007-2013) / ERC GA n. 321253. and Microsoft Research |
dc.format.extent | 8 p. |
dc.language.iso | eng |
dc.publisher | Institute of Electrical and Electronics Engineers (IEEE) |
dc.subject | Àrees temàtiques de la UPC::Enginyeria electrònica::Microelectrònica |
dc.subject | Àrees temàtiques de la UPC::Informàtica::Arquitectura de computadors |
dc.subject.lcsh | Microprocessors |
dc.subject.lcsh | Computer architecture |
dc.subject.other | DLP accelerator |
dc.subject.other | DVX enabled 4-core EDGE CMP |
dc.subject.other | EDGE architecture |
dc.subject.other | Cost-effective technique |
dc.subject.other | Data parallel workloads |
dc.subject.other | Dedicated accelerator |
dc.subject.other | Dynamic vector execution |
dc.subject.other | Energy efficient substrate |
dc.subject.other | Energy-delay product |
dc.subject.other | Explicit data graph execution |
dc.subject.other | Functionality |
dc.subject.other | General purpose CMP |
dc.subject.other | General purpose EDGE chip |
dc.subject.other | Multiprocessor |
dc.subject.other | High performance vector design |
dc.subject.other | Low power chip multiprocessor |
dc.subject.other | Minimal hardware |
dc.subject.other | Modest processor |
dc.subject.other | Special-purpose vector architecture |
dc.subject.other | Vector accelerator |
dc.subject.other | Vector control |
dc.subject.other | Vector processor |
dc.subject.other | Microprocessor chips |
dc.subject.other | Multiprocessing systems |
dc.subject.other | Computational modeling |
dc.subject.other | Computer architecture |
dc.subject.other | Hardware |
dc.subject.other | Instruction sets |
dc.subject.other | Message systems |
dc.subject.other | Registers |
dc.subject.other | Vectors |
dc.title | Dynamic-vector execution on a general purpose EDGE chip multiprocessor |
dc.type | Conference report |
dc.subject.lemac | Microprocessadors |
dc.subject.lemac | Arquitectura d'ordinadors |
dc.contributor.group | Universitat Politècnica de Catalunya. CAP - Grup de Computació d'Altes Prestacions |
dc.identifier.doi | 10.1109/SAMOS.2014.6893190 |
dc.description.peerreviewed | Peer Reviewed |
dc.relation.publisherversion | http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6893190 |
dc.rights.access | Restricted access - publisher's policy |
local.identifier.drac | 15430860 |
dc.description.version | Postprint (published version) |
dc.relation.projectid | info:eu-repo/grantAgreement/EC/FP7/321253/EU/Riding on Moore's Law/ROMOL |
dc.date.lift | 10000-01-01 |
local.citation.author | Duric, M.; Palomar, O.; Smith, A.; Stanic, M.; Unsal, O.; Cristal, A.; Valero, M.; Burger, D.; Veidenbaum, A.V. |
local.citation.contributor | International Conference on Embedded Computer Systems: Architectures, Modeling and Simulation |
local.citation.pubplace | Samos |
local.citation.publicationName | International Conference on Embedded Computer Systems: Architectures, Modeling and Simulation (SAMOS XIV): proceedings: July 14-17, 2014: Samos, Greece |
local.citation.startingPage | 18 |
local.citation.endingPage | 25 |