Matches in SemOpenAlex for { <https://semopenalex.org/work/W2786374423> ?p ?o ?g. }
Showing items 1 to 92 of
92
with 100 items per page.
- W2786374423 abstract "Recent work on Winograd-based convolution allows for a great reduction of computational complexity, but existing implementations are limited to 2D data and a single kernel size of 3 by 3. They can achieve only slightly better, and often worse performance than better optimized, direct convolution implementations. We propose and implement an algorithm for N-dimensional Winograd-based convolution that allows arbitrary kernel sizes and is optimized for manycore CPUs. Our algorithm achieves high hardware utilization through a series of optimizations. Our experiments show that on modern ConvNets, our optimized implementation, is on average more than 3 x, and sometimes 8 x faster than other state-of-the-art CPU implementations on an Intel Xeon Phi manycore processors. Moreover, our implementation on the Xeon Phi achieves competitive performance for 2D ConvNets and superior performance for 3D ConvNets, compared with the best GPU implementations." @default.
- W2786374423 created "2018-02-23" @default.
- W2786374423 creator A5023640358 @default.
- W2786374423 creator A5029097412 @default.
- W2786374423 creator A5051363890 @default.
- W2786374423 creator A5075155488 @default.
- W2786374423 date "2018-02-10" @default.
- W2786374423 modified "2023-10-01" @default.
- W2786374423 title "Optimizing N-dimensional, winograd-based convolution for manycore CPUs" @default.
- W2786374423 cites W1493004075 @default.
- W2786374423 cites W1745334888 @default.
- W2786374423 cites W1884620995 @default.
- W2786374423 cites W1901129140 @default.
- W2786374423 cites W2004642969 @default.
- W2786374423 cites W2022740893 @default.
- W2786374423 cites W2024165284 @default.
- W2786374423 cites W2029171059 @default.
- W2786374423 cites W2065710348 @default.
- W2786374423 cites W2097117768 @default.
- W2786374423 cites W2108598243 @default.
- W2786374423 cites W2117539524 @default.
- W2786374423 cites W2172654076 @default.
- W2786374423 cites W2211722331 @default.
- W2786374423 cites W2340076492 @default.
- W2786374423 cites W2485412435 @default.
- W2786374423 cites W2529546376 @default.
- W2786374423 cites W2617189289 @default.
- W2786374423 cites W2737305288 @default.
- W2786374423 cites W4247913514 @default.
- W2786374423 doi "https://doi.org/10.1145/3178487.3178496" @default.
- W2786374423 hasPublicationYear "2018" @default.
- W2786374423 type Work @default.
- W2786374423 sameAs 2786374423 @default.
- W2786374423 citedByCount "27" @default.
- W2786374423 countsByYear W27863744232018 @default.
- W2786374423 countsByYear W27863744232019 @default.
- W2786374423 countsByYear W27863744232020 @default.
- W2786374423 countsByYear W27863744232021 @default.
- W2786374423 countsByYear W27863744232022 @default.
- W2786374423 countsByYear W27863744232023 @default.
- W2786374423 crossrefType "proceedings-article" @default.
- W2786374423 hasAuthorship W2786374423A5023640358 @default.
- W2786374423 hasAuthorship W2786374423A5029097412 @default.
- W2786374423 hasAuthorship W2786374423A5051363890 @default.
- W2786374423 hasAuthorship W2786374423A5075155488 @default.
- W2786374423 hasConcept C111335779 @default.
- W2786374423 hasConcept C11413529 @default.
- W2786374423 hasConcept C114614502 @default.
- W2786374423 hasConcept C145108525 @default.
- W2786374423 hasConcept C154945302 @default.
- W2786374423 hasConcept C173608175 @default.
- W2786374423 hasConcept C199360897 @default.
- W2786374423 hasConcept C2524010 @default.
- W2786374423 hasConcept C26713055 @default.
- W2786374423 hasConcept C33923547 @default.
- W2786374423 hasConcept C41008148 @default.
- W2786374423 hasConcept C45347329 @default.
- W2786374423 hasConcept C50644808 @default.
- W2786374423 hasConcept C74193536 @default.
- W2786374423 hasConcept C96972482 @default.
- W2786374423 hasConceptScore W2786374423C111335779 @default.
- W2786374423 hasConceptScore W2786374423C11413529 @default.
- W2786374423 hasConceptScore W2786374423C114614502 @default.
- W2786374423 hasConceptScore W2786374423C145108525 @default.
- W2786374423 hasConceptScore W2786374423C154945302 @default.
- W2786374423 hasConceptScore W2786374423C173608175 @default.
- W2786374423 hasConceptScore W2786374423C199360897 @default.
- W2786374423 hasConceptScore W2786374423C2524010 @default.
- W2786374423 hasConceptScore W2786374423C26713055 @default.
- W2786374423 hasConceptScore W2786374423C33923547 @default.
- W2786374423 hasConceptScore W2786374423C41008148 @default.
- W2786374423 hasConceptScore W2786374423C45347329 @default.
- W2786374423 hasConceptScore W2786374423C50644808 @default.
- W2786374423 hasConceptScore W2786374423C74193536 @default.
- W2786374423 hasConceptScore W2786374423C96972482 @default.
- W2786374423 hasLocation W27863744231 @default.
- W2786374423 hasOpenAccess W2786374423 @default.
- W2786374423 hasPrimaryLocation W27863744231 @default.
- W2786374423 hasRelatedWork W2170268965 @default.
- W2786374423 hasRelatedWork W2209489587 @default.
- W2786374423 hasRelatedWork W2214459866 @default.
- W2786374423 hasRelatedWork W2475524688 @default.
- W2786374423 hasRelatedWork W2526069705 @default.
- W2786374423 hasRelatedWork W2753965283 @default.
- W2786374423 hasRelatedWork W2953672018 @default.
- W2786374423 hasRelatedWork W2981664121 @default.
- W2786374423 hasRelatedWork W3004176791 @default.
- W2786374423 hasRelatedWork W3022446491 @default.
- W2786374423 isParatext "false" @default.
- W2786374423 isRetracted "false" @default.
- W2786374423 magId "2786374423" @default.
- W2786374423 workType "article" @default.