FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3122144106> ?p ?o ?g. }

• W3122144106 endingPage "169" @default.
• W3122144106 startingPage "156" @default.
• W3122144106 abstract "Convolutional Neural Networks (CNN) are widely employed in the contemporary artificial intelligence systems. However these models have millions of connections between the layers, that are both memory prohibitive and computationally expensive. Employing these models on an embedded mobile application is resource limited with high power consumption and significant bandwidth requirement to access the data from the off-chip DRAM. Reducing the data movement between the on-chip and off-chip DRAM is the main criteria to achieve high throughput and overall better energy efficiency. Our proposed multi-bit accelerator achieves these goals by employing the truncation of the partial sum (Psum) results of the preceding layer before feeding it into the next layer. We exhibit the architecture by inferencing 32-bits for the first convolution layers and sequentially truncate the bits on the MSB/LSB of integer and fractional part without any further training on the original network. At the last fully connected layer, the top-1 accuracy is maintained with the reduced bit width of 14 and top-5 accuracy upto 10-bit width. The computation engine consists of an systolic array of 1024 processing elements (PE). Large CNNs such as AlexNet, MobileNet, SqueezeNet and EfficientNet were used as benchmark CNN model and Virtex Ultrascale FPGA was used to test the architecture. The proposed truncation scheme has 49% power reduction and resource utilization was reduced by 73.25% for LUTs (Look-up tables), 68.76% for FFs (Flip-Flops), 74.60% for BRAMs (Block RAMs) and 79.425% for Digital Signal Processors (DSPs) when compared with the 32 bits architecture. The design achieves a performance of 223.69 GOPS on a Virtex Ultrascale FPGA, the design has a overall gain of 3.63 × throughput when compared to other prior FPGA accelerators. In addition, the overall power consumption is 4.5 × lower when compared to other prior architectures. The ASIC version of the accelerator was designed in 22nm FDSOI CMOS process to achieve a overall throughput of 2.03 TOPS/W with a total power consumption of 791 mW and with a area of 1 mm × 1.2 mm." @default.
• W3122144106 created "2021-02-01" @default.
• W3122144106 creator A5032302022 @default.
• W3122144106 creator A5054827934 @default.
• W3122144106 creator A5075540562 @default.
• W3122144106 creator A5081055268 @default.
• W3122144106 date "2021-01-01" @default.
• W3122144106 modified "2023-10-14" @default.
• W3122144106 title "A Power Efficiency Enhancements of a Multi-Bit Accelerator for Memory Prohibitive Deep Neural Networks" @default.
• W3122144106 cites W1498436455 @default.
• W3122144106 cites W2094756095 @default.
• W3122144106 cites W2097117768 @default.
• W3122144106 cites W2112796928 @default.
• W3122144106 cites W2117539524 @default.
• W3122144106 cites W2125203716 @default.
• W3122144106 cites W2136922672 @default.
• W3122144106 cites W2276486856 @default.
• W3122144106 cites W2289252105 @default.
• W3122144106 cites W2294282016 @default.
• W3122144106 cites W2560023338 @default.
• W3122144106 cites W2896983500 @default.
• W3122144106 cites W2907463061 @default.
• W3122144106 cites W2946355854 @default.
• W3122144106 cites W2989331028 @default.
• W3122144106 cites W3090896767 @default.
• W3122144106 cites W4242577057 @default.
• W3122144106 cites W4247198796 @default.
• W3122144106 doi "https://doi.org/10.1109/ojcas.2020.3047225" @default.
• W3122144106 hasPublicationYear "2021" @default.
• W3122144106 type Work @default.
• W3122144106 sameAs 3122144106 @default.
• W3122144106 citedByCount "8" @default.
• W3122144106 countsByYear W31221441062021 @default.
• W3122144106 countsByYear W31221441062022 @default.
• W3122144106 countsByYear W31221441062023 @default.
• W3122144106 crossrefType "journal-article" @default.
• W3122144106 hasAuthorship W3122144106A5032302022 @default.
• W3122144106 hasAuthorship W3122144106A5054827934 @default.
• W3122144106 hasAuthorship W3122144106A5075540562 @default.
• W3122144106 hasAuthorship W3122144106A5081055268 @default.
• W3122144106 hasBestOaLocation W31221441061 @default.
• W3122144106 hasConcept C111335779 @default.
• W3122144106 hasConcept C113775141 @default.
• W3122144106 hasConcept C119599485 @default.
• W3122144106 hasConcept C127413603 @default.
• W3122144106 hasConcept C13280743 @default.
• W3122144106 hasConcept C154945302 @default.
• W3122144106 hasConcept C157764524 @default.
• W3122144106 hasConcept C173608175 @default.
• W3122144106 hasConcept C185798385 @default.
• W3122144106 hasConcept C188045654 @default.
• W3122144106 hasConcept C205649164 @default.
• W3122144106 hasConcept C2524010 @default.
• W3122144106 hasConcept C2742236 @default.
• W3122144106 hasConcept C2776257435 @default.
• W3122144106 hasConcept C2777210771 @default.
• W3122144106 hasConcept C2777674469 @default.
• W3122144106 hasConcept C33923547 @default.
• W3122144106 hasConcept C41008148 @default.
• W3122144106 hasConcept C42935608 @default.
• W3122144106 hasConcept C555944384 @default.
• W3122144106 hasConcept C7366592 @default.
• W3122144106 hasConcept C76155785 @default.
• W3122144106 hasConcept C81363708 @default.
• W3122144106 hasConcept C9390403 @default.
• W3122144106 hasConceptScore W3122144106C111335779 @default.
• W3122144106 hasConceptScore W3122144106C113775141 @default.
• W3122144106 hasConceptScore W3122144106C119599485 @default.
• W3122144106 hasConceptScore W3122144106C127413603 @default.
• W3122144106 hasConceptScore W3122144106C13280743 @default.
• W3122144106 hasConceptScore W3122144106C154945302 @default.
• W3122144106 hasConceptScore W3122144106C157764524 @default.
• W3122144106 hasConceptScore W3122144106C173608175 @default.
• W3122144106 hasConceptScore W3122144106C185798385 @default.
• W3122144106 hasConceptScore W3122144106C188045654 @default.
• W3122144106 hasConceptScore W3122144106C205649164 @default.
• W3122144106 hasConceptScore W3122144106C2524010 @default.
• W3122144106 hasConceptScore W3122144106C2742236 @default.
• W3122144106 hasConceptScore W3122144106C2776257435 @default.
• W3122144106 hasConceptScore W3122144106C2777210771 @default.
• W3122144106 hasConceptScore W3122144106C2777674469 @default.
• W3122144106 hasConceptScore W3122144106C33923547 @default.
• W3122144106 hasConceptScore W3122144106C41008148 @default.
• W3122144106 hasConceptScore W3122144106C42935608 @default.
• W3122144106 hasConceptScore W3122144106C555944384 @default.
• W3122144106 hasConceptScore W3122144106C7366592 @default.
• W3122144106 hasConceptScore W3122144106C76155785 @default.
• W3122144106 hasConceptScore W3122144106C81363708 @default.
• W3122144106 hasConceptScore W3122144106C9390403 @default.
• W3122144106 hasLocation W31221441061 @default.
• W3122144106 hasOpenAccess W3122144106 @default.
• W3122144106 hasPrimaryLocation W31221441061 @default.
• W3122144106 hasRelatedWork W1997955449 @default.
• W3122144106 hasRelatedWork W2047588290 @default.
• W3122144106 hasRelatedWork W2091231956 @default.
• W3122144106 hasRelatedWork W2466675884 @default.
• W3122144106 hasRelatedWork W2524802307 @default.
• W3122144106 hasRelatedWork W2951390974 @default.

• next