Work

A domain-specific architecture for deep neural networks

Year: 2018

Type: article

Abstract: Tensor processing units improve performance per watt of neural networks in Google datacenters by roughly 50x.

Source: Communications of the ACM

DOI: 10.1145/3154484

Authors: Norman P. Jouppi, Cliff Young, Nishant Patil, David Patterson

Institutions: Google (United States), University of California, Berkeley

Language: English

Topic: Parallel Computing and Optimization Techniques

Subfield: Hardware and Architecture

Field: Computer Science

Domain: Physical Sciences

Open Access status: bronze

OpenAlex ID: W2888727064

FWCI: 21.01

Cites: 32

Cited by: 176

Related to: 10