Matches in SemOpenAlex for { <https://semopenalex.org/work/W2466675884> ?p ?o ?g. }
Showing items 1 to 95 of
95
with 100 items per page.
- W2466675884 abstract "Convolutional neural networks (CNNs) are revolutionizing machine learning, but they present significant computational challenges. Recently, many FPGA-based accelerators have been proposed to improve the performance and efficiency of CNNs. Current approaches construct a single processor that computes the CNN layers one at a time; the processor is optimized to maximize the throughput at which the collection of layers is computed. However, this approach leads to inefficient designs because the same processor structure is used to compute CNN layers of radically varying dimensions. We present a new CNN accelerator paradigm and an accompanying automated design methodology that partitions the available FPGA resources into multiple processors, each of which is tailored for a different subset of the CNN convolutional layers. Using the same FPGA resources as a single large processor, multiple smaller specialized processors increase computational efficiency and lead to a higher overall throughput. Our design methodology achieves 3.8x higher throughput than the state-of-the-art approach on evaluating the popular AlexNet CNN on a Xilinx Virtex-7 FPGA. For the more recent SqueezeNet and GoogLeNet, the speedups are 2.2x and 2.0x." @default.
- W2466675884 created "2016-07-22" @default.
- W2466675884 creator A5010051312 @default.
- W2466675884 creator A5010989227 @default.
- W2466675884 creator A5061612911 @default.
- W2466675884 date "2017-06-24" @default.
- W2466675884 modified "2023-10-17" @default.
- W2466675884 title "Maximizing CNN Accelerator Efficiency Through Resource Partitioning" @default.
- W2466675884 cites W1968422655 @default.
- W2466675884 cites W1990315422 @default.
- W2466675884 cites W2009832130 @default.
- W2466675884 cites W2048266589 @default.
- W2466675884 cites W2094756095 @default.
- W2466675884 cites W2096645269 @default.
- W2466675884 cites W2097117768 @default.
- W2466675884 cites W2117130368 @default.
- W2466675884 cites W2117696986 @default.
- W2466675884 cites W2122989035 @default.
- W2466675884 cites W2125203716 @default.
- W2466675884 cites W2152839228 @default.
- W2466675884 cites W2276486856 @default.
- W2466675884 cites W2294282016 @default.
- W2466675884 cites W2330958039 @default.
- W2466675884 cites W2396572963 @default.
- W2466675884 cites W2403646140 @default.
- W2466675884 cites W2562773490 @default.
- W2466675884 cites W2603836393 @default.
- W2466675884 cites W2765815218 @default.
- W2466675884 cites W2906043559 @default.
- W2466675884 cites W4243519499 @default.
- W2466675884 cites W4251575795 @default.
- W2466675884 cites W4254672563 @default.
- W2466675884 doi "https://doi.org/10.1145/3079856.3080221" @default.
- W2466675884 hasPublicationYear "2017" @default.
- W2466675884 type Work @default.
- W2466675884 sameAs 2466675884 @default.
- W2466675884 citedByCount "190" @default.
- W2466675884 countsByYear W24666758842017 @default.
- W2466675884 countsByYear W24666758842018 @default.
- W2466675884 countsByYear W24666758842019 @default.
- W2466675884 countsByYear W24666758842020 @default.
- W2466675884 countsByYear W24666758842021 @default.
- W2466675884 countsByYear W24666758842022 @default.
- W2466675884 countsByYear W24666758842023 @default.
- W2466675884 crossrefType "proceedings-article" @default.
- W2466675884 hasAuthorship W2466675884A5010051312 @default.
- W2466675884 hasAuthorship W2466675884A5010989227 @default.
- W2466675884 hasAuthorship W2466675884A5061612911 @default.
- W2466675884 hasBestOaLocation W24666758841 @default.
- W2466675884 hasConcept C108583219 @default.
- W2466675884 hasConcept C111919701 @default.
- W2466675884 hasConcept C113775141 @default.
- W2466675884 hasConcept C118524514 @default.
- W2466675884 hasConcept C149635348 @default.
- W2466675884 hasConcept C154945302 @default.
- W2466675884 hasConcept C157764524 @default.
- W2466675884 hasConcept C173608175 @default.
- W2466675884 hasConcept C199360897 @default.
- W2466675884 hasConcept C2780801425 @default.
- W2466675884 hasConcept C41008148 @default.
- W2466675884 hasConcept C42935608 @default.
- W2466675884 hasConcept C555944384 @default.
- W2466675884 hasConcept C81363708 @default.
- W2466675884 hasConceptScore W2466675884C108583219 @default.
- W2466675884 hasConceptScore W2466675884C111919701 @default.
- W2466675884 hasConceptScore W2466675884C113775141 @default.
- W2466675884 hasConceptScore W2466675884C118524514 @default.
- W2466675884 hasConceptScore W2466675884C149635348 @default.
- W2466675884 hasConceptScore W2466675884C154945302 @default.
- W2466675884 hasConceptScore W2466675884C157764524 @default.
- W2466675884 hasConceptScore W2466675884C173608175 @default.
- W2466675884 hasConceptScore W2466675884C199360897 @default.
- W2466675884 hasConceptScore W2466675884C2780801425 @default.
- W2466675884 hasConceptScore W2466675884C41008148 @default.
- W2466675884 hasConceptScore W2466675884C42935608 @default.
- W2466675884 hasConceptScore W2466675884C555944384 @default.
- W2466675884 hasConceptScore W2466675884C81363708 @default.
- W2466675884 hasLocation W24666758841 @default.
- W2466675884 hasLocation W24666758842 @default.
- W2466675884 hasOpenAccess W2466675884 @default.
- W2466675884 hasPrimaryLocation W24666758841 @default.
- W2466675884 hasRelatedWork W2320205417 @default.
- W2466675884 hasRelatedWork W2400714260 @default.
- W2466675884 hasRelatedWork W2466675884 @default.
- W2466675884 hasRelatedWork W2524802307 @default.
- W2466675884 hasRelatedWork W2612090114 @default.
- W2466675884 hasRelatedWork W2904058793 @default.
- W2466675884 hasRelatedWork W2951390974 @default.
- W2466675884 hasRelatedWork W2954307240 @default.
- W2466675884 hasRelatedWork W2997828269 @default.
- W2466675884 hasRelatedWork W4220896354 @default.
- W2466675884 isParatext "false" @default.
- W2466675884 isRetracted "false" @default.
- W2466675884 magId "2466675884" @default.
- W2466675884 workType "article" @default.