Matches in SemOpenAlex for { <https://semopenalex.org/work/W2977088082> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W2977088082 endingPage "288" @default.
- W2977088082 startingPage "277" @default.
- W2977088082 abstract "Deep neural networks (DNNs) have been considered to be the state-of-the-art artificial intelligence methods in a very broad range of applications. However, DNNs are compute intensive and memory intensive which are difficult to be employed in practical scenarios. Due to their favorable parallel computing ability, a series of DNN accelerators have been proposed. However, the improvement of on-chip computing capacity and the increasing number of parameters in the neural networks make access to memory a bottleneck. In this paper, we analyze the existing DNN algorithms. We observe that the special structure of neural networks makes it have two useful characteristics, which are unilateral directivity and local independence. Based on these characteristics, we propose a general software scheduling method to reduce memory access cost. Based on the experimental results, our method can reduce 32% memory access cost and achieve a speedup of 1.6x in average on our experiment platform and the best result is in ResNet-50, which is up to 56% and 2.62x." @default.
- W2977088082 created "2019-10-03" @default.
- W2977088082 creator A5000432967 @default.
- W2977088082 creator A5033237869 @default.
- W2977088082 creator A5063268423 @default.
- W2977088082 creator A5069888696 @default.
- W2977088082 creator A5079809798 @default.
- W2977088082 creator A5081255430 @default.
- W2977088082 creator A5086219184 @default.
- W2977088082 date "2019-01-01" @default.
- W2977088082 modified "2023-09-29" @default.
- W2977088082 title "Deep Fusion: A Software Scheduling Method for Memory Access Optimization" @default.
- W2977088082 cites W1745334888 @default.
- W2977088082 cites W2048266589 @default.
- W2977088082 cites W2097117768 @default.
- W2977088082 cites W2194775991 @default.
- W2977088082 cites W2285660444 @default.
- W2977088082 cites W2289252105 @default.
- W2977088082 cites W2603836393 @default.
- W2977088082 cites W2963122961 @default.
- W2977088082 cites W4212788319 @default.
- W2977088082 cites W4230989867 @default.
- W2977088082 cites W4249932213 @default.
- W2977088082 cites W639708223 @default.
- W2977088082 doi "https://doi.org/10.1007/978-3-030-30709-7_22" @default.
- W2977088082 hasPublicationYear "2019" @default.
- W2977088082 type Work @default.
- W2977088082 sameAs 2977088082 @default.
- W2977088082 citedByCount "2" @default.
- W2977088082 countsByYear W29770880822020 @default.
- W2977088082 countsByYear W29770880822022 @default.
- W2977088082 crossrefType "book-chapter" @default.
- W2977088082 hasAuthorship W2977088082A5000432967 @default.
- W2977088082 hasAuthorship W2977088082A5033237869 @default.
- W2977088082 hasAuthorship W2977088082A5063268423 @default.
- W2977088082 hasAuthorship W2977088082A5069888696 @default.
- W2977088082 hasAuthorship W2977088082A5079809798 @default.
- W2977088082 hasAuthorship W2977088082A5081255430 @default.
- W2977088082 hasAuthorship W2977088082A5086219184 @default.
- W2977088082 hasBestOaLocation W29770880822 @default.
- W2977088082 hasConcept C111919701 @default.
- W2977088082 hasConcept C113775141 @default.
- W2977088082 hasConcept C120314980 @default.
- W2977088082 hasConcept C126255220 @default.
- W2977088082 hasConcept C149635348 @default.
- W2977088082 hasConcept C154945302 @default.
- W2977088082 hasConcept C159985019 @default.
- W2977088082 hasConcept C173608175 @default.
- W2977088082 hasConcept C192562407 @default.
- W2977088082 hasConcept C204323151 @default.
- W2977088082 hasConcept C206729178 @default.
- W2977088082 hasConcept C2777904410 @default.
- W2977088082 hasConcept C2780513914 @default.
- W2977088082 hasConcept C2984842247 @default.
- W2977088082 hasConcept C33923547 @default.
- W2977088082 hasConcept C41008148 @default.
- W2977088082 hasConcept C50644808 @default.
- W2977088082 hasConcept C68339613 @default.
- W2977088082 hasConceptScore W2977088082C111919701 @default.
- W2977088082 hasConceptScore W2977088082C113775141 @default.
- W2977088082 hasConceptScore W2977088082C120314980 @default.
- W2977088082 hasConceptScore W2977088082C126255220 @default.
- W2977088082 hasConceptScore W2977088082C149635348 @default.
- W2977088082 hasConceptScore W2977088082C154945302 @default.
- W2977088082 hasConceptScore W2977088082C159985019 @default.
- W2977088082 hasConceptScore W2977088082C173608175 @default.
- W2977088082 hasConceptScore W2977088082C192562407 @default.
- W2977088082 hasConceptScore W2977088082C204323151 @default.
- W2977088082 hasConceptScore W2977088082C206729178 @default.
- W2977088082 hasConceptScore W2977088082C2777904410 @default.
- W2977088082 hasConceptScore W2977088082C2780513914 @default.
- W2977088082 hasConceptScore W2977088082C2984842247 @default.
- W2977088082 hasConceptScore W2977088082C33923547 @default.
- W2977088082 hasConceptScore W2977088082C41008148 @default.
- W2977088082 hasConceptScore W2977088082C50644808 @default.
- W2977088082 hasConceptScore W2977088082C68339613 @default.
- W2977088082 hasLocation W29770880821 @default.
- W2977088082 hasLocation W29770880822 @default.
- W2977088082 hasLocation W29770880823 @default.
- W2977088082 hasLocation W29770880824 @default.
- W2977088082 hasLocation W29770880825 @default.
- W2977088082 hasOpenAccess W2977088082 @default.
- W2977088082 hasPrimaryLocation W29770880821 @default.
- W2977088082 hasRelatedWork W1509211761 @default.
- W2977088082 hasRelatedWork W1531488649 @default.
- W2977088082 hasRelatedWork W1585350690 @default.
- W2977088082 hasRelatedWork W2133693067 @default.
- W2977088082 hasRelatedWork W2366027386 @default.
- W2977088082 hasRelatedWork W2391299576 @default.
- W2977088082 hasRelatedWork W2540591044 @default.
- W2977088082 hasRelatedWork W2582456645 @default.
- W2977088082 hasRelatedWork W3037767301 @default.
- W2977088082 hasRelatedWork W2479014312 @default.
- W2977088082 isParatext "false" @default.
- W2977088082 isRetracted "false" @default.
- W2977088082 magId "2977088082" @default.
- W2977088082 workType "book-chapter" @default.