High Energy Efficiency Neural Network Processor with Combined Digital and Computing-in-Memory Architecture [electronic resource] / by Jinshan Yue.

Por: Yue, Jinshan [author.]Colaborador(es): SpringerLink (Online service)Tipo de material: TextoTextoSeries Springer Theses, Recognizing Outstanding Ph.D. ResearchEditor: Singapore : Springer Nature Singapore : Imprint: Springer, 2024Edición: 1st ed. 2024Descripción: XVI, 118 p. 81 illus., 78 illus. in color. online resourceTipo de contenido: text Tipo de medio: computer Tipo de portador: online resourceISBN: 9789819734771Tema(s): Electronics | Electronic circuits | Microprocessors | Computer architecture | Computational intelligence | Electronics and Microelectronics, Instrumentation | Electronic Circuits and Systems | Processor Architectures | Computational IntelligenceFormatos físicos adicionales: Printed edition:: Sin título; Printed edition:: Sin título; Printed edition:: Sin títuloClasificación CDD: 621.381 Clasificación LoC:TK7800-8360Recursos en línea: Libro electrónicoTexto
Contenidos:
Introduction -- Basis and research status of neural network processor -- Neural network processor for specific kernel optimized data reuse -- Neural network processor with frequency domain compression algorithm optimization -- Neural network processor combining digital and computing in memory architecture -- Digital computing in memory neural network processor supporting large scale models -- Conclusion and prospect.
En: Springer Nature eBookResumen: Neural network (NN) algorithms are driving the rapid development of modern artificial intelligence (AI). The energy-efficient NN processor has become an urgent requirement for the practical NN applications on widespread low-power AI devices. To address this challenge, this dissertation investigates pure-digital and digital computing-in-memory (digital-CIM) solutions and carries out four major studies. For pure-digital NN processors, this book analyses the insufficient data reuse in conventional architectures and proposes a kernel-optimized NN processor. This dissertation adopts a structural frequency-domain compression algorithm, named CirCNN. The fabricated processor shows 8.1x/4.2x area/energy efficiency compared to the state-of-the-art NN processor. For digital-CIM NN processors, this dissertation combines the flexibility of digital circuits with the high energy efficiency of CIM. The fabricated CIM processor validates the sparsity improvement of the CIM architecture for the first time. This dissertation further designs a processor that considers the weight updating problem on the CIM architecture for the first time. This dissertation demonstrates that the combination of digital and CIM circuits is a promising technical route for an energy-efficient NN processor, which can promote the large-scale application of low-power AI devices. .
Star ratings
    Valoración media: 0.0 (0 votos)
Existencias
Tipo de ítem Biblioteca actual Colección Signatura Copia número Estado Fecha de vencimiento Código de barras
Libro Electrónico Biblioteca Electrónica
Colección de Libros Electrónicos 1 No para préstamo

Introduction -- Basis and research status of neural network processor -- Neural network processor for specific kernel optimized data reuse -- Neural network processor with frequency domain compression algorithm optimization -- Neural network processor combining digital and computing in memory architecture -- Digital computing in memory neural network processor supporting large scale models -- Conclusion and prospect.

Neural network (NN) algorithms are driving the rapid development of modern artificial intelligence (AI). The energy-efficient NN processor has become an urgent requirement for the practical NN applications on widespread low-power AI devices. To address this challenge, this dissertation investigates pure-digital and digital computing-in-memory (digital-CIM) solutions and carries out four major studies. For pure-digital NN processors, this book analyses the insufficient data reuse in conventional architectures and proposes a kernel-optimized NN processor. This dissertation adopts a structural frequency-domain compression algorithm, named CirCNN. The fabricated processor shows 8.1x/4.2x area/energy efficiency compared to the state-of-the-art NN processor. For digital-CIM NN processors, this dissertation combines the flexibility of digital circuits with the high energy efficiency of CIM. The fabricated CIM processor validates the sparsity improvement of the CIM architecture for the first time. This dissertation further designs a processor that considers the weight updating problem on the CIM architecture for the first time. This dissertation demonstrates that the combination of digital and CIM circuits is a promising technical route for an energy-efficient NN processor, which can promote the large-scale application of low-power AI devices. .

UABC ; Perpetuidad

Con tecnología Koha