Hosted by Utah State University Libraries
Reclaiming Fault Resilience and Energy Efficiency With Enhanced Performance in Low Power Architectures
Abstract
dc:description.abstractRapid developments of the AI domain has revolutionized the computing industry by the introduction of state-of-art AI architectures. This growth is also accompanied by a massive increase in the power consumption. Near-Theshold Computing (NTC) has emerged as a viable solution by offering significant savings in power consumption paving the way for an energy efficient design paradigm. However, these benefits are accompanied by a deterioration in performance due to the severe process variation and slower transistor switching at Near-Threshold operation. These problems severely restrict the usage of Near-Threshold operation in commercial applications. In this work, a novel AI architecture, Tensor Processing Unit, operating at NTC is thoroughly investigated to tackle the issues hindering system performance. Research problems are demonstrated in a scientific manner and unique opportunities are explored to propose novel design methodologies.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy (PhD)
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Electrical and Computer Engineering
- Grantor dc:publisher
- Hosted by Utah State University Libraries
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Gundi, Noel Daniel
- Contributors dc:contributor
-
- Sanghamitra Roy
- Koushik Chakraborty
- Greg Droge