Matches in SemOpenAlex for { <https://semopenalex.org/work/W2753100145> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W2753100145 endingPage "12:11" @default.
- W2753100145 startingPage "12:1" @default.
- W2753100145 abstract "Training deep learning models often occupies entire compute clusters, built solely for this purpose, for days or even weeks at a time. There exists a large body of work on approaches for improving training performance, ranging from novel algorithms to full custom hardware accelerators. Offering compute capabilities of multiple teraflops (trillion floating point operations per second), graphics processing units (GPUs) have established themselves as a de-facto standard for accelerating deep learning network training. As systems with up to 16 GPUs—each GPU consuming up to 300 W—become available, efficient usage of these resources becomes imperative. We conduct a detailed analysis of deep learning workloads to characterize their efficiency in making use of GPU acceleration. We found that many deep learning workloads consume only a fraction of GPU resources, and we demonstrate how sharing GPU resources can improve throughput by a factor of 3, effectively turning a 4-GPU commodity cloud system into a high-end 12-GPU supercomputer. Using Watson workloads from three major areas that incorporate deep learning technology—i.e., language classification, visual recognition, and speech recognition—we document the effectiveness and scalability of our approach. We are working toward enabling GPU virtualization not only to reduce cost, but also to accelerate new breakthroughs in deep learning by increasing compute capacity without making further hardware investments." @default.
- W2753100145 created "2017-09-15" @default.
- W2753100145 creator A5049413685 @default.
- W2753100145 creator A5058607529 @default.
- W2753100145 creator A5071385812 @default.
- W2753100145 date "2017-07-01" @default.
- W2753100145 modified "2023-09-30" @default.
- W2753100145 title "Optimizing the efficiency of deep learning through accelerator virtualization" @default.
- W2753100145 cites W1966838590 @default.
- W2753100145 cites W1972399950 @default.
- W2753100145 cites W2051598123 @default.
- W2753100145 cites W2076063813 @default.
- W2753100145 cites W2155893237 @default.
- W2753100145 cites W2235191389 @default.
- W2753100145 cites W2504282214 @default.
- W2753100145 cites W2753217274 @default.
- W2753100145 cites W2120920992 @default.
- W2753100145 doi "https://doi.org/10.1147/jrd.2017.2716598" @default.
- W2753100145 hasPublicationYear "2017" @default.
- W2753100145 type Work @default.
- W2753100145 sameAs 2753100145 @default.
- W2753100145 citedByCount "8" @default.
- W2753100145 countsByYear W27531001452018 @default.
- W2753100145 countsByYear W27531001452020 @default.
- W2753100145 countsByYear W27531001452021 @default.
- W2753100145 countsByYear W27531001452023 @default.
- W2753100145 crossrefType "journal-article" @default.
- W2753100145 hasAuthorship W2753100145A5049413685 @default.
- W2753100145 hasAuthorship W2753100145A5058607529 @default.
- W2753100145 hasAuthorship W2753100145A5071385812 @default.
- W2753100145 hasConcept C108583219 @default.
- W2753100145 hasConcept C111919701 @default.
- W2753100145 hasConcept C118524514 @default.
- W2753100145 hasConcept C154945302 @default.
- W2753100145 hasConcept C173608175 @default.
- W2753100145 hasConcept C21442007 @default.
- W2753100145 hasConcept C2778119891 @default.
- W2753100145 hasConcept C2779851693 @default.
- W2753100145 hasConcept C41008148 @default.
- W2753100145 hasConcept C48044578 @default.
- W2753100145 hasConcept C50630238 @default.
- W2753100145 hasConcept C513985346 @default.
- W2753100145 hasConcept C79974875 @default.
- W2753100145 hasConcept C83283714 @default.
- W2753100145 hasConceptScore W2753100145C108583219 @default.
- W2753100145 hasConceptScore W2753100145C111919701 @default.
- W2753100145 hasConceptScore W2753100145C118524514 @default.
- W2753100145 hasConceptScore W2753100145C154945302 @default.
- W2753100145 hasConceptScore W2753100145C173608175 @default.
- W2753100145 hasConceptScore W2753100145C21442007 @default.
- W2753100145 hasConceptScore W2753100145C2778119891 @default.
- W2753100145 hasConceptScore W2753100145C2779851693 @default.
- W2753100145 hasConceptScore W2753100145C41008148 @default.
- W2753100145 hasConceptScore W2753100145C48044578 @default.
- W2753100145 hasConceptScore W2753100145C50630238 @default.
- W2753100145 hasConceptScore W2753100145C513985346 @default.
- W2753100145 hasConceptScore W2753100145C79974875 @default.
- W2753100145 hasConceptScore W2753100145C83283714 @default.
- W2753100145 hasIssue "4/5" @default.
- W2753100145 hasLocation W27531001451 @default.
- W2753100145 hasOpenAccess W2753100145 @default.
- W2753100145 hasPrimaryLocation W27531001451 @default.
- W2753100145 hasRelatedWork W183542552 @default.
- W2753100145 hasRelatedWork W189420351 @default.
- W2753100145 hasRelatedWork W2153645222 @default.
- W2753100145 hasRelatedWork W2319727539 @default.
- W2753100145 hasRelatedWork W2755264124 @default.
- W2753100145 hasRelatedWork W2806898911 @default.
- W2753100145 hasRelatedWork W3041026627 @default.
- W2753100145 hasRelatedWork W3111284907 @default.
- W2753100145 hasRelatedWork W67367039 @default.
- W2753100145 hasRelatedWork W2187537688 @default.
- W2753100145 hasVolume "61" @default.
- W2753100145 isParatext "false" @default.
- W2753100145 isRetracted "false" @default.
- W2753100145 magId "2753100145" @default.
- W2753100145 workType "article" @default.