Matches in SemOpenAlex for { <https://semopenalex.org/work/W3043302551> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W3043302551 abstract "Recently, Deep Learning (DL) is widely applied to intelligent systems equipped with resource-constraint hardware accelerators. With multiple DL applications sharing the resource, the execution model can be divided into two stages: (i) batching independent inference tasks initiated by each application, and (ii) scheduling batches to run in a time-sharing manner. The state-of-the-art DL serving systems employ the execution model by organizing sequential tasks into batches and then scheduling batches concerning their targeting deep neural network (DNN) models in a round-robin manner. However, we demonstrated that these practices fail to alleviate the slowdown of tasks, and there is a need to re-visit batching and scheduling in terms of efficiency and fairness. To this end, we formulated batching as a resource allocation problem and investigated scheduling in terms of each application’s utilization on the device. Then, we proposed the fine-grained batching scheme and fairness-driven scheduling scheme for DL serving and implemented a prototype system called DyBatch. To be exact, DyBatch accomplishes efficient batching by taking into account Pareto efficiency of and envy between batches. Besides, DyBatch’s fair scheduler monitors the resource utilization of all applications and assigns a batch from the application with the lowest utilization for execution first. Evaluation under various benchmarks with comparison to the baseline system Tensorflow Serving (TFS) shows the superiority of DyBatch, which achieves up to 55% reduction of slowdown, and up to 12% improvement of throughput." @default.
- W3043302551 created "2020-07-23" @default.
- W3043302551 creator A5003905836 @default.
- W3043302551 creator A5015993565 @default.
- W3043302551 creator A5036601222 @default.
- W3043302551 creator A5042742679 @default.
- W3043302551 creator A5054836950 @default.
- W3043302551 date "2020-05-01" @default.
- W3043302551 modified "2023-10-07" @default.
- W3043302551 title "DyBatch: Efficient Batching and Fair Scheduling for Deep Learning Inference on Time-sharing Devices" @default.
- W3043302551 cites W1686810756 @default.
- W3043302551 cites W2068666958 @default.
- W3043302551 cites W2105947650 @default.
- W3043302551 cites W2113952909 @default.
- W3043302551 cites W2117539524 @default.
- W3043302551 cites W2158296221 @default.
- W3043302551 cites W2163961697 @default.
- W3043302551 cites W2194775991 @default.
- W3043302551 cites W2340897893 @default.
- W3043302551 cites W2412782625 @default.
- W3043302551 cites W2531409750 @default.
- W3043302551 cites W2572666084 @default.
- W3043302551 cites W2779888504 @default.
- W3043302551 cites W2798515322 @default.
- W3043302551 cites W2809420642 @default.
- W3043302551 cites W2893114493 @default.
- W3043302551 cites W2893890695 @default.
- W3043302551 cites W2899071864 @default.
- W3043302551 cites W2906773779 @default.
- W3043302551 cites W2963163009 @default.
- W3043302551 cites W2963697717 @default.
- W3043302551 cites W2964108773 @default.
- W3043302551 doi "https://doi.org/10.1109/ccgrid49817.2020.00-32" @default.
- W3043302551 hasPublicationYear "2020" @default.
- W3043302551 type Work @default.
- W3043302551 sameAs 3043302551 @default.
- W3043302551 citedByCount "1" @default.
- W3043302551 countsByYear W30433025512022 @default.
- W3043302551 crossrefType "proceedings-article" @default.
- W3043302551 hasAuthorship W3043302551A5003905836 @default.
- W3043302551 hasAuthorship W3043302551A5015993565 @default.
- W3043302551 hasAuthorship W3043302551A5036601222 @default.
- W3043302551 hasAuthorship W3043302551A5042742679 @default.
- W3043302551 hasAuthorship W3043302551A5054836950 @default.
- W3043302551 hasConcept C111919701 @default.
- W3043302551 hasConcept C120314980 @default.
- W3043302551 hasConcept C126255220 @default.
- W3043302551 hasConcept C154945302 @default.
- W3043302551 hasConcept C206729178 @default.
- W3043302551 hasConcept C2776214188 @default.
- W3043302551 hasConcept C31258907 @default.
- W3043302551 hasConcept C33923547 @default.
- W3043302551 hasConcept C38725249 @default.
- W3043302551 hasConcept C41008148 @default.
- W3043302551 hasConcept C51332947 @default.
- W3043302551 hasConcept C79403827 @default.
- W3043302551 hasConceptScore W3043302551C111919701 @default.
- W3043302551 hasConceptScore W3043302551C120314980 @default.
- W3043302551 hasConceptScore W3043302551C126255220 @default.
- W3043302551 hasConceptScore W3043302551C154945302 @default.
- W3043302551 hasConceptScore W3043302551C206729178 @default.
- W3043302551 hasConceptScore W3043302551C2776214188 @default.
- W3043302551 hasConceptScore W3043302551C31258907 @default.
- W3043302551 hasConceptScore W3043302551C33923547 @default.
- W3043302551 hasConceptScore W3043302551C38725249 @default.
- W3043302551 hasConceptScore W3043302551C41008148 @default.
- W3043302551 hasConceptScore W3043302551C51332947 @default.
- W3043302551 hasConceptScore W3043302551C79403827 @default.
- W3043302551 hasLocation W30433025511 @default.
- W3043302551 hasOpenAccess W3043302551 @default.
- W3043302551 hasPrimaryLocation W30433025511 @default.
- W3043302551 hasRelatedWork W1142101 @default.
- W3043302551 hasRelatedWork W12224968 @default.
- W3043302551 hasRelatedWork W14730544 @default.
- W3043302551 hasRelatedWork W14862478 @default.
- W3043302551 hasRelatedWork W3052431 @default.
- W3043302551 hasRelatedWork W3955301 @default.
- W3043302551 hasRelatedWork W6056896 @default.
- W3043302551 hasRelatedWork W8939044 @default.
- W3043302551 hasRelatedWork W9251591 @default.
- W3043302551 hasRelatedWork W14551109 @default.
- W3043302551 isParatext "false" @default.
- W3043302551 isRetracted "false" @default.
- W3043302551 magId "3043302551" @default.
- W3043302551 workType "article" @default.