Matches in SemOpenAlex for { <https://semopenalex.org/work/W3208788005> ?p ?o ?g. }
- W3208788005 endingPage "211" @default.
- W3208788005 startingPage "198" @default.
- W3208788005 abstract "This work demonstrates a programmable in-memory-computing (IMC) inference accelerator for scalable execution of neural network (NN) models, leveraging a high-signal-to-noise ratio (SNR) capacitor-based analog technology. IMC accelerates computations and reduces memory accessing for matrix-vector multiplies (MVMs), which dominate in NNs. The accelerator architecture focuses on scalable execution, addressing the overheads of state swapping and the challenges of maintaining high utilization across highly dense and parallel hardware. The architecture is based on a configurable on-chip network (OCN) and scalable array of cores, which integrate mixed-signal IMC with programmable near-memory single-instruction multiple-data (SIMD) digital computing, configurable buffering, and programmable control. The cores enable flexible NN execution mappings that exploit data- and pipeline-parallelism to address utilization and efficiency across models. A prototype is presented, incorporating a <inline-formula> <tex-math notation=LaTeX>$4 times 4$ </tex-math></inline-formula> array of cores demonstrated in 16 nm CMOS, achieving peak multiply-accumulate (MAC)-level throughput of 3 TOPS and peak MAC-level energy efficiency of 30 TOPS/W, both for 8-b operations. The measured results shows high accuracy of the analog computations, matching bit-true simulations. This enables the abstractions required for robust and scalable architectural and software integration. Developed software libraries and NN-mapping tools are used to demonstrate CIFAR-10 and ImageNet classification, with an 11-layer CNN and ResNet-50, respectively, achieving accuracy, throughput, and energy efficiency of 91.51% and 73.33%, 7815 and 581 image/s, 51.5 k and 3.0 k image/s/W, with 4-b weights and activations." @default.
- W3208788005 created "2021-11-08" @default.
- W3208788005 creator A5014075867 @default.
- W3208788005 creator A5014213149 @default.
- W3208788005 creator A5019057160 @default.
- W3208788005 creator A5028744078 @default.
- W3208788005 creator A5036267240 @default.
- W3208788005 creator A5079625063 @default.
- W3208788005 creator A5088075152 @default.
- W3208788005 date "2022-01-01" @default.
- W3208788005 modified "2023-09-30" @default.
- W3208788005 title "Scalable and Programmable Neural Network Inference Accelerator Based on In-Memory Computing" @default.
- W3208788005 cites W2046964635 @default.
- W3208788005 cites W2048266589 @default.
- W3208788005 cites W2064675550 @default.
- W3208788005 cites W2194775991 @default.
- W3208788005 cites W2285660444 @default.
- W3208788005 cites W2289252105 @default.
- W3208788005 cites W2538451693 @default.
- W3208788005 cites W2591601611 @default.
- W3208788005 cites W2604319603 @default.
- W3208788005 cites W2606722458 @default.
- W3208788005 cites W2894696827 @default.
- W3208788005 cites W2897997157 @default.
- W3208788005 cites W2899641901 @default.
- W3208788005 cites W2920326572 @default.
- W3208788005 cites W2921898934 @default.
- W3208788005 cites W2963037989 @default.
- W3208788005 cites W2969812992 @default.
- W3208788005 cites W2976137532 @default.
- W3208788005 cites W2990591126 @default.
- W3208788005 cites W2991232789 @default.
- W3208788005 cites W2999044305 @default.
- W3208788005 cites W3046305306 @default.
- W3208788005 cites W3097655915 @default.
- W3208788005 cites W3112860454 @default.
- W3208788005 cites W3128966520 @default.
- W3208788005 cites W4252370083 @default.
- W3208788005 doi "https://doi.org/10.1109/jssc.2021.3119018" @default.
- W3208788005 hasPublicationYear "2022" @default.
- W3208788005 type Work @default.
- W3208788005 sameAs 3208788005 @default.
- W3208788005 citedByCount "9" @default.
- W3208788005 countsByYear W32087880052022 @default.
- W3208788005 countsByYear W32087880052023 @default.
- W3208788005 crossrefType "journal-article" @default.
- W3208788005 hasAuthorship W3208788005A5014075867 @default.
- W3208788005 hasAuthorship W3208788005A5014213149 @default.
- W3208788005 hasAuthorship W3208788005A5019057160 @default.
- W3208788005 hasAuthorship W3208788005A5028744078 @default.
- W3208788005 hasAuthorship W3208788005A5036267240 @default.
- W3208788005 hasAuthorship W3208788005A5079625063 @default.
- W3208788005 hasAuthorship W3208788005A5088075152 @default.
- W3208788005 hasConcept C11413529 @default.
- W3208788005 hasConcept C118524514 @default.
- W3208788005 hasConcept C119599485 @default.
- W3208788005 hasConcept C127413603 @default.
- W3208788005 hasConcept C150552126 @default.
- W3208788005 hasConcept C157764524 @default.
- W3208788005 hasConcept C173608175 @default.
- W3208788005 hasConcept C2742236 @default.
- W3208788005 hasConcept C41008148 @default.
- W3208788005 hasConcept C45374587 @default.
- W3208788005 hasConcept C48044578 @default.
- W3208788005 hasConcept C555944384 @default.
- W3208788005 hasConcept C76155785 @default.
- W3208788005 hasConcept C77088390 @default.
- W3208788005 hasConcept C9390403 @default.
- W3208788005 hasConceptScore W3208788005C11413529 @default.
- W3208788005 hasConceptScore W3208788005C118524514 @default.
- W3208788005 hasConceptScore W3208788005C119599485 @default.
- W3208788005 hasConceptScore W3208788005C127413603 @default.
- W3208788005 hasConceptScore W3208788005C150552126 @default.
- W3208788005 hasConceptScore W3208788005C157764524 @default.
- W3208788005 hasConceptScore W3208788005C173608175 @default.
- W3208788005 hasConceptScore W3208788005C2742236 @default.
- W3208788005 hasConceptScore W3208788005C41008148 @default.
- W3208788005 hasConceptScore W3208788005C45374587 @default.
- W3208788005 hasConceptScore W3208788005C48044578 @default.
- W3208788005 hasConceptScore W3208788005C555944384 @default.
- W3208788005 hasConceptScore W3208788005C76155785 @default.
- W3208788005 hasConceptScore W3208788005C77088390 @default.
- W3208788005 hasConceptScore W3208788005C9390403 @default.
- W3208788005 hasFunder F4320309292 @default.
- W3208788005 hasIssue "1" @default.
- W3208788005 hasLocation W32087880051 @default.
- W3208788005 hasOpenAccess W3208788005 @default.
- W3208788005 hasPrimaryLocation W32087880051 @default.
- W3208788005 hasRelatedWork W1569389315 @default.
- W3208788005 hasRelatedWork W1585350690 @default.
- W3208788005 hasRelatedWork W1637649405 @default.
- W3208788005 hasRelatedWork W2000051442 @default.
- W3208788005 hasRelatedWork W2008876287 @default.
- W3208788005 hasRelatedWork W2062683875 @default.
- W3208788005 hasRelatedWork W2105469423 @default.
- W3208788005 hasRelatedWork W2502560717 @default.
- W3208788005 hasRelatedWork W4245302940 @default.
- W3208788005 hasRelatedWork W2503642292 @default.