Publication detail

Prediction of Inference Energy on CNN Accelerators Supporting Approximate Circuits

PIŇOS, M. MRÁZEK, V. SEKANINA, L.

Original Title

Prediction of Inference Energy on CNN Accelerators Supporting Approximate Circuits

Type

conference paper

Language

English

Original Abstract

Design methodologies developed for optimizing hardware implementations of convolutional neural networks (CNN) or searching for new hardware-aware neural architectures rely on the fast and reliable estimation of key hardware parameters, such as the energy needed for one inference. Utilizing approximate circuits in hardware accelerators of CNNs faces the designers with new problems during their simulation - commonly used tools (TimeLoop, Accelergy, Maestro) do not support approximate arithmetic operations. This work addresses the fast and efficient prediction of consumed energy in hardware accelerators of CNNs that utilize approximate circuits such as approximate multipliers. First, we extend the state-of-the-art software frameworks TimeLoop and Accelergy to predict the inference energy when exact multipliers are replaced with various approximate implementations. The energies obtained using the modified tools are then considered the ground truth (reference) values. Then, we propose and evaluate, using two accelerators (Eyeriss and Simba) and two types of networks (CNNs generated by EvoApproxNAS and standard ResNet CNNs), two predictors of inference energy. We conclude that a simple predictor based on summing the energies needed for all multiplications highly correlates with the reference values if the CNN's architecture is fixed. For complex CNNs with variable architectures typically generated by neural architecture search algorithms, a more sophisticated predictor based on a machine learning model has to be employed. The proposed predictors are 420-533× faster than reference solutions.

Keywords

approximate computing, prediction, hardware accelerator, convolutional neural networks

Authors

PIŇOS, M.; MRÁZEK, V.; SEKANINA, L.

Released

3. 5. 2023

Publisher

Institute of Electrical and Electronics Engineers

Location

Talinn

ISBN

979-8-3503-3277-3

Book

2023 26th International Symposium on Design and Diagnostics of Electronic Circuits and Systems

Pages from

45

Pages to

50

Pages count

6

URL

https://ieeexplore.ieee.org/document/10139724

BibTex

@inproceedings{BUT183569,
  author="Michal {Piňos} and Vojtěch {Mrázek} and Lukáš {Sekanina}",
  title="Prediction of Inference Energy on CNN Accelerators Supporting Approximate Circuits",
  booktitle="2023 26th International Symposium on Design and Diagnostics of Electronic Circuits and Systems",
  year="2023",
  pages="45--50",
  publisher="Institute of Electrical and Electronics Engineers",
  address="Talinn",
  doi="10.1109/DDECS57882.2023.10139724",
  isbn="979-8-3503-3277-3",
  url="https://ieeexplore.ieee.org/document/10139724"
}