Matches in SemOpenAlex for { <https://semopenalex.org/work/W2736230459> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W2736230459 abstract "Thanks to modern deep learning frameworks that exploit GPUs, convolutional neural networks (CNNs) have been greatly successful in visual recognition tasks. In this paper, we analyze the GPU performance characteristics of five popular deep learning frameworks: Caffe, CNTK, TensorFlow, Theano, and Torch in the perspective of a representative CNN model, AlexNet. Based on the characteristics obtained, we suggest possible optimization methods to increase the efficiency of CNN models built by the frameworks. We also show the GPU performance characteristics of different convolution algorithms each of which uses one of GEMM, direct convolution, FFT, and the Winograd method. We also suggest criteria to choose convolution algorithms for GPUs and methods to build efficient CNN models on GPUs. Since scaling DNNs in a multi-GPU context becomes increasingly important, we also analyze the scalability of the CNN models built by the deep learning frameworks in the multi-GPU context and their overhead. The result indicates that we can increase the speed of training the AlexNet model up to 2X by just changing options provided by the frameworks." @default.
- W2736230459 created "2017-07-21" @default.
- W2736230459 creator A5025937126 @default.
- W2736230459 creator A5037054528 @default.
- W2736230459 creator A5057343217 @default.
- W2736230459 creator A5082375033 @default.
- W2736230459 date "2017-04-01" @default.
- W2736230459 modified "2023-10-02" @default.
- W2736230459 title "Performance analysis of CNN frameworks for GPUs" @default.
- W2736230459 cites W1487564550 @default.
- W2736230459 cites W1995562189 @default.
- W2736230459 cites W2031489346 @default.
- W2736230459 cites W2102605133 @default.
- W2736230459 cites W2112796928 @default.
- W2736230459 cites W2117539524 @default.
- W2736230459 cites W2145339207 @default.
- W2736230459 cites W2155893237 @default.
- W2736230459 cites W2407022425 @default.
- W2736230459 cites W2963674387 @default.
- W2736230459 doi "https://doi.org/10.1109/ispass.2017.7975270" @default.
- W2736230459 hasPublicationYear "2017" @default.
- W2736230459 type Work @default.
- W2736230459 sameAs 2736230459 @default.
- W2736230459 citedByCount "65" @default.
- W2736230459 countsByYear W27362304592017 @default.
- W2736230459 countsByYear W27362304592018 @default.
- W2736230459 countsByYear W27362304592019 @default.
- W2736230459 countsByYear W27362304592020 @default.
- W2736230459 countsByYear W27362304592021 @default.
- W2736230459 countsByYear W27362304592022 @default.
- W2736230459 countsByYear W27362304592023 @default.
- W2736230459 crossrefType "proceedings-article" @default.
- W2736230459 hasAuthorship W2736230459A5025937126 @default.
- W2736230459 hasAuthorship W2736230459A5037054528 @default.
- W2736230459 hasAuthorship W2736230459A5057343217 @default.
- W2736230459 hasAuthorship W2736230459A5082375033 @default.
- W2736230459 hasConcept C108583219 @default.
- W2736230459 hasConcept C113775141 @default.
- W2736230459 hasConcept C114614502 @default.
- W2736230459 hasConcept C119857082 @default.
- W2736230459 hasConcept C151730666 @default.
- W2736230459 hasConcept C154945302 @default.
- W2736230459 hasConcept C165696696 @default.
- W2736230459 hasConcept C173608175 @default.
- W2736230459 hasConcept C199360897 @default.
- W2736230459 hasConcept C2778119891 @default.
- W2736230459 hasConcept C2779343474 @default.
- W2736230459 hasConcept C2779960059 @default.
- W2736230459 hasConcept C33923547 @default.
- W2736230459 hasConcept C38652104 @default.
- W2736230459 hasConcept C41008148 @default.
- W2736230459 hasConcept C45347329 @default.
- W2736230459 hasConcept C48044578 @default.
- W2736230459 hasConcept C50644808 @default.
- W2736230459 hasConcept C68339613 @default.
- W2736230459 hasConcept C74193536 @default.
- W2736230459 hasConcept C77088390 @default.
- W2736230459 hasConcept C81363708 @default.
- W2736230459 hasConcept C86803240 @default.
- W2736230459 hasConceptScore W2736230459C108583219 @default.
- W2736230459 hasConceptScore W2736230459C113775141 @default.
- W2736230459 hasConceptScore W2736230459C114614502 @default.
- W2736230459 hasConceptScore W2736230459C119857082 @default.
- W2736230459 hasConceptScore W2736230459C151730666 @default.
- W2736230459 hasConceptScore W2736230459C154945302 @default.
- W2736230459 hasConceptScore W2736230459C165696696 @default.
- W2736230459 hasConceptScore W2736230459C173608175 @default.
- W2736230459 hasConceptScore W2736230459C199360897 @default.
- W2736230459 hasConceptScore W2736230459C2778119891 @default.
- W2736230459 hasConceptScore W2736230459C2779343474 @default.
- W2736230459 hasConceptScore W2736230459C2779960059 @default.
- W2736230459 hasConceptScore W2736230459C33923547 @default.
- W2736230459 hasConceptScore W2736230459C38652104 @default.
- W2736230459 hasConceptScore W2736230459C41008148 @default.
- W2736230459 hasConceptScore W2736230459C45347329 @default.
- W2736230459 hasConceptScore W2736230459C48044578 @default.
- W2736230459 hasConceptScore W2736230459C50644808 @default.
- W2736230459 hasConceptScore W2736230459C68339613 @default.
- W2736230459 hasConceptScore W2736230459C74193536 @default.
- W2736230459 hasConceptScore W2736230459C77088390 @default.
- W2736230459 hasConceptScore W2736230459C81363708 @default.
- W2736230459 hasConceptScore W2736230459C86803240 @default.
- W2736230459 hasLocation W27362304591 @default.
- W2736230459 hasOpenAccess W2736230459 @default.
- W2736230459 hasPrimaryLocation W27362304591 @default.
- W2736230459 hasRelatedWork W2256144436 @default.
- W2736230459 hasRelatedWork W2606416966 @default.
- W2736230459 hasRelatedWork W2794923745 @default.
- W2736230459 hasRelatedWork W2907215606 @default.
- W2736230459 hasRelatedWork W2923029190 @default.
- W2736230459 hasRelatedWork W2983282793 @default.
- W2736230459 hasRelatedWork W2989991345 @default.
- W2736230459 hasRelatedWork W3045877795 @default.
- W2736230459 hasRelatedWork W3159529979 @default.
- W2736230459 hasRelatedWork W4312417841 @default.
- W2736230459 isParatext "false" @default.
- W2736230459 isRetracted "false" @default.
- W2736230459 magId "2736230459" @default.
- W2736230459 workType "article" @default.