Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability

None, None; None, None; None, None; None, None; None, None

Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability

Conference Paper (2019)

Author(s)

Anderson Luiz Sartor (Carnegie Mellon University)

Pedro Henrique Exenberger Becker (Universidade Federal do Rio Grande do Sul)

Stephan Wong (TU Delft - Computer Engineering)

Radu Marculescu (Carnegie Mellon University)

Antonio Carlos Schneider Beck (Universidade Federal do Rio Grande do Sul)

Machine learning Fault tolerance Energy consumption Adaptive processor Runtime optimization

DOI related publication

https://doi.org/10.1109/ISVLSI.2019.00037 Final published version

To reference this document use

https://resolver.tudelft.nl/uuid:7faf83f9-e8ad-42e5-ab47-b8ccb9069124

More Info

expand_more

Publication Year

2019

Language

English

Article number

8839457

Pages (from-to)

158-163

ISBN (print)

978-1-7281-3392-8

ISBN (electronic)

978-1-7281-3391-1

Event

Downloads counter

182

Abstract

Adaptive processors can dynamically change their hardware configuration by tuning several knobs that optimize a given metric, according to the current application. However, the complexity of choosing the best setup at runtime increases exponentially as more adaptive resources become available. Therefore, we propose a polymorphic VLIW processor coupled to a machine learning-based decision mechanism that quickly and accurately delivers the best trade-off in terms of energy, performance, and reliability. The proposed system predicts the best processor configuration in 97.37% of the test cases and achieves an efficiency that is close to an oracle (more than 93.30% on all benchmarks).