UC San Diego

Deep neural networks have become prominent in solving many real-life problems. However, they need to rely on learning patterns of data. As the demand for such services grows, merely scaling-out the number of accelerators is not economically cost-effective. Although multi-tenancy has propelled data center scalability, it has not been a primary factor in designing DNN accelerators due to the arms race for higher speed and efficiency. A new architecture is proposed which helps in spatially co-locating multiple DNN inference services on the same hardware, offering simultaneous multi-tenant DNN acceleration.