INTERNATIONAL JOURNAL OF NOVEL RESEARCH AND DEVELOPMENT International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2456-4184 | Impact factor: 8.76 | ESTD Year: 2016
Scholarly open access journals, Peer-reviewed, and Refereed Journals, Impact factor 8.76 (Calculate by google scholar and Semantic Scholar | AI-Powered Research Tool) , Multidisciplinary, Monthly, Indexing in all major database & Metadata, Citation Generator, Digital Object Identifier(DOI)
To control strong neuron-to-neuron traffic, Many processing nodes form a high-level neural network (NN), which can be used as a spinal or multi-core tool and built using network-on-chip (NoC). Chip-to-chip communication networks bring together many NN-based processors based on NoC to improve complete neural acceleration. A large amount of multicast-based traffic passes through a chip or between chips, making it difficult to network communication design and NN system performance and power impact. We recommend NeuronLink for NN accelerators in this study, which combines intrachip and interchip communication methods. With virtual-channel intrachip communication, we provide cross-cutting solution, troubleshooting, and compatible route calculation methods, resulting in the best NoC with reduced hardware cost based on multiple streams. For interchip communications, we propose a chip-to-chip communication system with NoC-aware that allows for efficient communication of NN-based NN devices. We have also put the proposed methods to the test on DNN-based NoC chips with four flexible gate frames (FPGAs). Test results show that the proposed communication network can better manage data flow within DNNs with higher output and less overhead compared to modern connections.
Keywords:
Neural Network Accelerator, Deep Neural Network, Nodes
Cite Article:
"Neuronlink: An Effective Large Scale Neural Network Accelerators Based on Chip-To-Chip Interconnect ", International Journal of Novel Research and Development (www.ijnrd.org), ISSN:2456-4184, Vol.7, Issue 5, page no.348-352, May-2022, Available :http://www.ijnrd.org/papers/IJNRDA001063.pdf
Downloads:
000118753
ISSN:
2456-4184 | IMPACT FACTOR: 8.76 Calculated By Google Scholar| ESTD YEAR: 2016
An International Scholarly Open Access Journal, Peer-Reviewed, Refereed Journal Impact Factor 8.76 Calculate by Google Scholar and Semantic Scholar | AI-Powered Research Tool, Multidisciplinary, Monthly, Multilanguage Journal Indexing in All Major Database & Metadata, Citation Generator
Facebook Twitter Instagram LinkedIn