Matches in SemOpenAlex for { <https://semopenalex.org/work/W2972585271> ?p ?o ?g. }
- W2972585271 abstract "To amortize cost, cloud vendors providing DNN acceleration as a service to end-users employ consolidation and virtualization to share the underlying resources among multiple DNN service requests. This paper makes a case for a neural processing unit (NPU) and a predictive multi-task scheduler to meet the latency demands of high-priority inference while maintaining high throughput. We evaluate both the mechanisms that enable NPUs to be preemptible and the policies that utilize them to meet scheduling objectives. We show that preemptive NPU multi-tasking can achieve an average 7.8x, 1.4x, and 4.8x improvement in latency, throughput, and SLA satisfaction, respectively." @default.
- W2972585271 created "2019-09-19" @default.
- W2972585271 creator A5070087229 @default.
- W2972585271 creator A5091648103 @default.
- W2972585271 date "2019-09-06" @default.
- W2972585271 modified "2023-09-27" @default.
- W2972585271 title "PREMA: A Predictive Multi-task Scheduling Algorithm For Preemptible Neural Processing Units" @default.
- W2972585271 cites W1667652561 @default.
- W2972585271 cites W1686810756 @default.
- W2972585271 cites W1855892484 @default.
- W2972585271 cites W1922655562 @default.
- W2972585271 cites W1951216520 @default.
- W2972585271 cites W1968047430 @default.
- W2972585271 cites W2000967104 @default.
- W2972585271 cites W2034861439 @default.
- W2972585271 cites W2048266589 @default.
- W2972585271 cites W2067523571 @default.
- W2972585271 cites W2097117768 @default.
- W2972585271 cites W2097643185 @default.
- W2972585271 cites W2098274770 @default.
- W2972585271 cites W2107916517 @default.
- W2972585271 cites W2125551452 @default.
- W2972585271 cites W2126118553 @default.
- W2972585271 cites W2129381159 @default.
- W2972585271 cites W2130942839 @default.
- W2972585271 cites W2152839228 @default.
- W2972585271 cites W2162639668 @default.
- W2972585271 cites W2163605009 @default.
- W2972585271 cites W2237922334 @default.
- W2972585271 cites W2261808795 @default.
- W2972585271 cites W2285660444 @default.
- W2972585271 cites W2289252105 @default.
- W2972585271 cites W2323909431 @default.
- W2972585271 cites W2395693197 @default.
- W2972585271 cites W2442974303 @default.
- W2972585271 cites W2508602506 @default.
- W2972585271 cites W2511743527 @default.
- W2972585271 cites W2515287984 @default.
- W2972585271 cites W2518281301 @default.
- W2972585271 cites W2518511512 @default.
- W2972585271 cites W2529546376 @default.
- W2972585271 cites W2541839172 @default.
- W2972585271 cites W2562773490 @default.
- W2972585271 cites W2563587242 @default.
- W2972585271 cites W2565851976 @default.
- W2972585271 cites W2566334102 @default.
- W2972585271 cites W2583383421 @default.
- W2972585271 cites W2592389822 @default.
- W2972585271 cites W2594990650 @default.
- W2972585271 cites W2604514113 @default.
- W2972585271 cites W2605347906 @default.
- W2972585271 cites W2606722458 @default.
- W2972585271 cites W2612445135 @default.
- W2972585271 cites W2625457103 @default.
- W2972585271 cites W2740550900 @default.
- W2972585271 cites W2764337919 @default.
- W2972585271 cites W2788007484 @default.
- W2972585271 cites W2796360442 @default.
- W2972585271 cites W2901839763 @default.
- W2972585271 cites W2903754802 @default.
- W2972585271 cites W2911491685 @default.
- W2972585271 cites W2914554687 @default.
- W2972585271 cites W2962821792 @default.
- W2972585271 cites W2964174152 @default.
- W2972585271 cites W2979719709 @default.
- W2972585271 cites W2565954965 @default.
- W2972585271 cites W2594575824 @default.
- W2972585271 cites W2612092431 @default.
- W2972585271 hasPublicationYear "2019" @default.
- W2972585271 type Work @default.
- W2972585271 sameAs 2972585271 @default.
- W2972585271 citedByCount "1" @default.
- W2972585271 countsByYear W29725852712020 @default.
- W2972585271 crossrefType "posted-content" @default.
- W2972585271 hasAuthorship W2972585271A5070087229 @default.
- W2972585271 hasAuthorship W2972585271A5091648103 @default.
- W2972585271 hasConcept C111919701 @default.
- W2972585271 hasConcept C120314980 @default.
- W2972585271 hasConcept C127413603 @default.
- W2972585271 hasConcept C154945302 @default.
- W2972585271 hasConcept C206729178 @default.
- W2972585271 hasConcept C21547014 @default.
- W2972585271 hasConcept C2776214188 @default.
- W2972585271 hasConcept C41008148 @default.
- W2972585271 hasConcept C513985346 @default.
- W2972585271 hasConcept C76155785 @default.
- W2972585271 hasConcept C79974875 @default.
- W2972585271 hasConcept C82876162 @default.
- W2972585271 hasConceptScore W2972585271C111919701 @default.
- W2972585271 hasConceptScore W2972585271C120314980 @default.
- W2972585271 hasConceptScore W2972585271C127413603 @default.
- W2972585271 hasConceptScore W2972585271C154945302 @default.
- W2972585271 hasConceptScore W2972585271C206729178 @default.
- W2972585271 hasConceptScore W2972585271C21547014 @default.
- W2972585271 hasConceptScore W2972585271C2776214188 @default.
- W2972585271 hasConceptScore W2972585271C41008148 @default.
- W2972585271 hasConceptScore W2972585271C513985346 @default.
- W2972585271 hasConceptScore W2972585271C76155785 @default.
- W2972585271 hasConceptScore W2972585271C79974875 @default.
- W2972585271 hasConceptScore W2972585271C82876162 @default.