FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2913221350> ?p ?o ?g. }

• W2913221350 endingPage "870" @default.
• W2913221350 startingPage "857" @default.
• W2913221350 abstract "In recent years, convolutional neural networks (CNNs) have become widely adopted for computer vision tasks. Field-programmable gate arrays (FPGAs) have been adequately explored as a promising hardware accelerator for CNNs due to its high performance, energy efficiency, and reconfigurability. However, prior FPGA solutions based on the conventional convolutional algorithm is often bounded by the computational capability of FPGAs (e.g., the number of DSPs). To address this problem, the feature maps are transformed to a special domain using fast algorithms to reduce the arithmetic complexity. Winograd and fast Fourier transformation (FFT), as fast algorithm representatives, first transform input data and filter to Winograd or frequency domain, then perform element-wise multiplication, and apply inverse transformation to get the final output. In this paper, we propose a novel architecture for implementing fast algorithms on FPGAs. Our design employs line buffer structure to effectively reuse the feature map data among different tiles. We also effectively pipeline the Winograd/FFT processing element (PE) engine and initiate multiple PEs through parallelization. Meanwhile, there exists a complex design space to explore. We propose an analytical model to predict the resource usage and the performance. Then, we use the model to guide a fast design space exploration. Experiments using the state-of-the-art CNNs demonstrate the best performance and energy efficiency on FPGAs. We achieve 854.6 and 2479.6 GOP/s for AlexNet and VGG16 on Xilinx ZCU102 platform using Winograd. We achieve 130.4 GOP/s for Resnet using Winograd and 201.1 GOP/s for YOLO using FFT on Xilinx ZC706 platform." @default.
• W2913221350 created "2019-02-21" @default.
• W2913221350 creator A5002704866 @default.
• W2913221350 creator A5049910854 @default.
• W2913221350 creator A5057484285 @default.
• W2913221350 creator A5072265593 @default.
• W2913221350 date "2020-04-01" @default.
• W2913221350 modified "2023-10-17" @default.
• W2913221350 title "Evaluating Fast Algorithms for Convolutional Neural Networks on FPGAs" @default.
• W2913221350 cites W1487564550 @default.
• W2913221350 cites W1677182931 @default.
• W2913221350 cites W1902041153 @default.
• W2913221350 cites W1969057818 @default.
• W2913221350 cites W1970608452 @default.
• W2913221350 cites W2009832130 @default.
• W2913221350 cites W2015332951 @default.
• W2913221350 cites W2018055497 @default.
• W2913221350 cites W2097117768 @default.
• W2913221350 cites W2102605133 @default.
• W2913221350 cites W2152839228 @default.
• W2913221350 cites W2155893237 @default.
• W2913221350 cites W2166029537 @default.
• W2913221350 cites W2172654076 @default.
• W2913221350 cites W2194775991 @default.
• W2913221350 cites W2241336042 @default.
• W2913221350 cites W2276486856 @default.
• W2913221350 cites W2294282016 @default.
• W2913221350 cites W2398354748 @default.
• W2913221350 cites W2403646140 @default.
• W2913221350 cites W2442974303 @default.
• W2913221350 cites W2475840367 @default.
• W2913221350 cites W2515287984 @default.
• W2913221350 cites W2574797063 @default.
• W2913221350 cites W2584311934 @default.
• W2913221350 cites W2584616277 @default.
• W2913221350 cites W2585774018 @default.
• W2913221350 cites W2605487586 @default.
• W2913221350 cites W2625954420 @default.
• W2913221350 cites W2626112510 @default.
• W2913221350 cites W2627042741 @default.
• W2913221350 cites W2727238169 @default.
• W2913221350 cites W2729080111 @default.
• W2913221350 cites W2742152118 @default.
• W2913221350 cites W2765235648 @default.
• W2913221350 cites W2773339846 @default.
• W2913221350 cites W2789246071 @default.
• W2913221350 cites W2792503273 @default.
• W2913221350 cites W2794754997 @default.
• W2913221350 cites W2795915628 @default.
• W2913221350 cites W2804541029 @default.
• W2913221350 cites W2808917878 @default.
• W2913221350 cites W2889758810 @default.
• W2913221350 cites W2900082550 @default.
• W2913221350 cites W2962820060 @default.
• W2913221350 cites W2963037989 @default.
• W2913221350 cites W4243682116 @default.
• W2913221350 cites W4245205197 @default.
• W2913221350 doi "https://doi.org/10.1109/tcad.2019.2897701" @default.
• W2913221350 hasPublicationYear "2020" @default.
• W2913221350 type Work @default.
• W2913221350 sameAs 2913221350 @default.
• W2913221350 citedByCount "66" @default.
• W2913221350 countsByYear W29132213502019 @default.
• W2913221350 countsByYear W29132213502020 @default.
• W2913221350 countsByYear W29132213502021 @default.
• W2913221350 countsByYear W29132213502022 @default.
• W2913221350 countsByYear W29132213502023 @default.
• W2913221350 crossrefType "journal-article" @default.
• W2913221350 hasAuthorship W2913221350A5002704866 @default.
• W2913221350 hasAuthorship W2913221350A5049910854 @default.
• W2913221350 hasAuthorship W2913221350A5057484285 @default.
• W2913221350 hasAuthorship W2913221350A5072265593 @default.
• W2913221350 hasConcept C11413529 @default.
• W2913221350 hasConcept C119857082 @default.
• W2913221350 hasConcept C149635348 @default.
• W2913221350 hasConcept C154945302 @default.
• W2913221350 hasConcept C173608175 @default.
• W2913221350 hasConcept C41008148 @default.
• W2913221350 hasConcept C42935608 @default.
• W2913221350 hasConcept C81363708 @default.
• W2913221350 hasConceptScore W2913221350C11413529 @default.
• W2913221350 hasConceptScore W2913221350C119857082 @default.
• W2913221350 hasConceptScore W2913221350C149635348 @default.
• W2913221350 hasConceptScore W2913221350C154945302 @default.
• W2913221350 hasConceptScore W2913221350C173608175 @default.
• W2913221350 hasConceptScore W2913221350C41008148 @default.
• W2913221350 hasConceptScore W2913221350C42935608 @default.
• W2913221350 hasConceptScore W2913221350C81363708 @default.
• W2913221350 hasFunder F4320322919 @default.
• W2913221350 hasIssue "4" @default.
• W2913221350 hasLocation W29132213501 @default.
• W2913221350 hasOpenAccess W2913221350 @default.
• W2913221350 hasPrimaryLocation W29132213501 @default.
• W2913221350 hasRelatedWork W2337926734 @default.
• W2913221350 hasRelatedWork W2799614062 @default.
• W2913221350 hasRelatedWork W2978290780 @default.
• W2913221350 hasRelatedWork W3021430260 @default.
• W2913221350 hasRelatedWork W3027997911 @default.

• next