Details
Presenter(s)
![Kris Nikov Headshot](https://confcats-catavault.s3.amazonaws.com/CATAVault/ieeecass/master/files/styles/cc_user_photo/s3/user-pictures/61761.jpg?h=e29c489f&itok=PxpchVFh)
Display Name
Kris Nikov
- Affiliation
-
AffiliationUniversity of Bristol
- Country
Abstract
This paper presents a novel approach to developing accurate models for the Arm Cortex-M0 CPU. The methodology uses a custom energy measurement harness on a commercially available platform and an Instruction Set Simulator to obtain statically predictable PMC events. The models are trained and validated on a much larger set of workloads compared to relevant work and achieve an average error of 5%.