Cerebras Wafer Scale Engine: An Introduction

August 21, 2019 //By Cerebras Systems Inc.
Cerebras Wafer Scale Engine: An Introduction

This white paper discusses the nature and scale of the workload for artificial intelligence applications as a background to the architecture of the Cerebras' waferscale neural network compute engine. The paper stresses the advantages of more cores, more memory close to cores and more bandwidth between cores. It also emphasizes the importance of archtectures to address sparse data and eliminate multiplication by zero.

AI, neural network, wafer-scale,

Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.