Answers are grounded in this paper and cite their sources.
Editor's note: Systolic array is embracing its renaissance after being accepted by Google TPU as the core computing architecture of machine learning acceleration. In this article, the authors propose two strategies to enhance fault tolerance of systolic array based deep neural network accelerators.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2019) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: