FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3207050768> ?p ?o ?g. }

• W3207050768 abstract "Existing deep convolutional neural networks (CNNs) generate massive interlayer feature data during network inference. To maintain real-time processing in embedded systems, large on-chip memory is required to buffer the interlayer feature maps. In this paper, we propose an efficient hardware accelerator with an interlayer feature compression technique to significantly reduce the required on-chip memory size and off-chip memory access bandwidth. The accelerator compresses interlayer feature maps through transforming the stored data into frequency domain using hardware-implemented 8x8 discrete cosine transform (DCT). The high-frequency components are removed after the DCT through quantization. Sparse matrix compression is utilized to further compress the interlayer feature maps. The on-chip memory allocation scheme is designed to support dynamic configuration of the feature map buffer size and scratch pad size according to different network-layer requirements. The hardware accelerator combines compression, decompression, and CNN acceleration into one computing stream, achieving minimal compressing and processing delay. A prototype accelerator is implemented on an FPGA platform and also synthesized in TSMC 28-nm COMS technology. It achieves 403GOPS peak throughput and 1.4x~3.3x interlayer feature map reduction by adding light hardware area overhead, making it a promising hardware accelerator for intelligent IoT devices." @default.
• W3207050768 created "2021-10-25" @default.
• W3207050768 creator A5001374364 @default.
• W3207050768 creator A5010738468 @default.
• W3207050768 creator A5020801023 @default.
• W3207050768 creator A5025473895 @default.
• W3207050768 creator A5039665206 @default.
• W3207050768 creator A5042006645 @default.
• W3207050768 creator A5047445405 @default.
• W3207050768 creator A5084816122 @default.
• W3207050768 creator A5085147583 @default.
• W3207050768 date "2021-10-12" @default.
• W3207050768 modified "2023-09-30" @default.
• W3207050768 title "Memory-Efficient CNN Accelerator Based on Interlayer Feature Map Compression" @default.
• W3207050768 cites W1964846093 @default.
• W3207050768 cites W1995562189 @default.
• W3207050768 cites W1999085092 @default.
• W3207050768 cites W2031614119 @default.
• W3207050768 cites W2112796928 @default.
• W3207050768 cites W2117696986 @default.
• W3207050768 cites W2140196014 @default.
• W3207050768 cites W2152839228 @default.
• W3207050768 cites W2165931732 @default.
• W3207050768 cites W2194775991 @default.
• W3207050768 cites W2285660444 @default.
• W3207050768 cites W2289252105 @default.
• W3207050768 cites W2346952186 @default.
• W3207050768 cites W2516141709 @default.
• W3207050768 cites W2593564159 @default.
• W3207050768 cites W2605487586 @default.
• W3207050768 cites W2625457103 @default.
• W3207050768 cites W2734572653 @default.
• W3207050768 cites W2740685955 @default.
• W3207050768 cites W2766143712 @default.
• W3207050768 cites W2773339846 @default.
• W3207050768 cites W2792447253 @default.
• W3207050768 cites W2809624076 @default.
• W3207050768 cites W2896718327 @default.
• W3207050768 cites W2912168260 @default.
• W3207050768 cites W2919115771 @default.
• W3207050768 cites W2963037989 @default.
• W3207050768 cites W2963122961 @default.
• W3207050768 cites W2963163009 @default.
• W3207050768 cites W2963374099 @default.
• W3207050768 cites W2973613048 @default.
• W3207050768 cites W2987129023 @default.
• W3207050768 cites W3018757597 @default.
• W3207050768 cites W3034739212 @default.
• W3207050768 cites W3091421838 @default.
• W3207050768 cites W3091762720 @default.
• W3207050768 cites W3092502001 @default.
• W3207050768 cites W3106250896 @default.
• W3207050768 cites W639708223 @default.
• W3207050768 doi "https://doi.org/10.48550/arxiv.2110.06155" @default.
• W3207050768 hasPublicationYear "2021" @default.
• W3207050768 type Work @default.
• W3207050768 sameAs 3207050768 @default.
• W3207050768 citedByCount "0" @default.
• W3207050768 crossrefType "posted-content" @default.
• W3207050768 hasAuthorship W3207050768A5001374364 @default.
• W3207050768 hasAuthorship W3207050768A5010738468 @default.
• W3207050768 hasAuthorship W3207050768A5020801023 @default.
• W3207050768 hasAuthorship W3207050768A5025473895 @default.
• W3207050768 hasAuthorship W3207050768A5039665206 @default.
• W3207050768 hasAuthorship W3207050768A5042006645 @default.
• W3207050768 hasAuthorship W3207050768A5047445405 @default.
• W3207050768 hasAuthorship W3207050768A5084816122 @default.
• W3207050768 hasAuthorship W3207050768A5085147583 @default.
• W3207050768 hasBestOaLocation W32070507681 @default.
• W3207050768 hasConcept C115961682 @default.
• W3207050768 hasConcept C13164978 @default.
• W3207050768 hasConcept C138885662 @default.
• W3207050768 hasConcept C149635348 @default.
• W3207050768 hasConcept C154945302 @default.
• W3207050768 hasConcept C165005293 @default.
• W3207050768 hasConcept C2221639 @default.
• W3207050768 hasConcept C2776401178 @default.
• W3207050768 hasConcept C41008148 @default.
• W3207050768 hasConcept C41895202 @default.
• W3207050768 hasConcept C42935608 @default.
• W3207050768 hasConcept C76155785 @default.
• W3207050768 hasConcept C78548338 @default.
• W3207050768 hasConcept C81363708 @default.
• W3207050768 hasConcept C9390403 @default.
• W3207050768 hasConceptScore W3207050768C115961682 @default.
• W3207050768 hasConceptScore W3207050768C13164978 @default.
• W3207050768 hasConceptScore W3207050768C138885662 @default.
• W3207050768 hasConceptScore W3207050768C149635348 @default.
• W3207050768 hasConceptScore W3207050768C154945302 @default.
• W3207050768 hasConceptScore W3207050768C165005293 @default.
• W3207050768 hasConceptScore W3207050768C2221639 @default.
• W3207050768 hasConceptScore W3207050768C2776401178 @default.
• W3207050768 hasConceptScore W3207050768C41008148 @default.
• W3207050768 hasConceptScore W3207050768C41895202 @default.
• W3207050768 hasConceptScore W3207050768C42935608 @default.
• W3207050768 hasConceptScore W3207050768C76155785 @default.
• W3207050768 hasConceptScore W3207050768C78548338 @default.
• W3207050768 hasConceptScore W3207050768C81363708 @default.
• W3207050768 hasConceptScore W3207050768C9390403 @default.
• W3207050768 hasLocation W32070507681 @default.

• next