This paper sets out to explore this timely requirement of multi-tenancy through a new dimension: dynamic architecture fission. To that end, we define Planaria 1 ...

This paper introduces and explores the dimension of dynamic fission in DNN accelerators. This innovation enables simultaneous execution of multiple DNN.

This paper defines Planaria1, a microarchitectural capability that can dynamically fission (break) into multiple smaller yet full-fledged DNN engines at ...

This microarchitectural capability enables spatially co-locating multiple DNN inference services on the same hardware, offering simultaneous multi-tenant DNN ...

The objective is to enable multi-tenant execution of DNNs by spatially co-locating multiple DNN tasks on a single accelerator.

An improved algorithm for speeding up learning of neural networks is proposed. The improved algorithm combines fast cascade-correlation and dynamic optimization ...

Oct 17, 2020 · Dive into the research topics of 'Planaria: Dynamic architecture fission for spatial multi-tenant acceleration of deep neural networks'.

Duration: 14:41
Posted: Feb 10, 2021

Major contribution: This paper introduces and explores the dimension of dynamic fission in DNN accelerators. This innovation enables simultaneous execution of ...

This codebase provides a simulation framework for Planraia, which is a spatially multi-tenant DNN accelerator. The codebase includes a hardware simulator ...