中文|CAS

logo
banner

Journals & Publications

Journals Publications Papers

Papers

DaDianNao: A Neural Network Supercomputer
Feb 08, 2017Author：
PrintText Size A A

Title: DaDianNao: A Neural Network Supercomputer
Authors: Luo, T; Liu, SL; Li, L; Wang, YQ; Zhang, SJ; Chen, TS; Xu, ZW; Temam, O; Chen, YJ
Author Full Names: Luo, Tao; Liu, Shaoli; Li, Ling; Wang, Yuqing; Zhang, Shijin; Chen, Tianshi; Xu, Zhiwei; Temam, Olivier; Chen, Yunji
Source: IEEE TRANSACTIONS ON COMPUTERS, 66 (1):73-88; 10.1109/TC.2016.2574353 JAN 1 2017
Language: English
Abstract: Many companies are deploying services largely based on machine-learning algorithms for sophisticated processing of large amounts of data, either for consumers or industry. The state-of-the-art and most popular such machine-learning algorithms are Convolutional and Deep Neural Networks (CNNs and DNNs), which are known to be computationally and memory intensive. A number of neural network accelerators have been recently proposed which can offer high computational capacity/area ratio, but which remain hampered by memory accesses. However, unlike the memory wall faced by processors on general-purpose workloads, the CNNs and DNNs memory footprint, while large, is not beyond the capability of the on-chip storage of a multi-chip system. This property, combined with the CNN/DNN algorithmic characteristics, can lead to high internal bandwidth and low external communications, which can in turn enable high-degree parallelism at a reasonable area cost. In this article, we introduce a custom multi-chip machine-learning architecture along those lines, and evaluate performance by integrating electrical and optical inter-chip interconnects separately. We show that, on a subset of the largest known neural network layers, it is possible to achieve a speedup of 656.63 x over a GPU, and reduce the energy by 184. 05 x on average for a 64-chip system. We implement the node down to the place and route at 28 nm, containing a combination of custom storage and computational units, with electrical inter-chip interconnects.
ISSN: 0018-9340
eISSN: 1557-9956
IDS Number: EF9RO
Unique ID: WOS:00039066760009
*Click Here to View Full Record