Abstract

Accurate and reliable measurement of energy consumption is critical for making well-informed design choices when choosing and training large scale NLP models. In this work, we show that existing software-based energy estimations are not accurate because they do not take into account hardware differences and how resource utilization affects energy consumption. We conduct energy measurement experiments with four different models for a question answering task. We quantify the error of existing software-based energy estimations by using a hardware power meter that provides highly accurate energy measurements. Our key takeaway is the need for a more accurate energy estimation model that takes into account hardware variabilities and the non-linear relationship between resource utilization and energy consumption. We release the code and data at https://github.com/csarron/sustainlp2020-energy.

Highlights

  • State-of-the-art NLP models of today (Devlin et al, 2019; Liu et al, 2019; Raffel et al, 2020) consume large amounts of energy

  • The total energy consumption is computed as the sum of the of the CPU, GPU, and memory, which is adjusted by a compensation constant (Henderson et al, 2020; Strubell et al, 2019)

  • We focus on energy consumption of inference for a question answering (QA) task using a hardware power meter

Read more

Summary

Introduction

State-of-the-art NLP models of today (Devlin et al, 2019; Liu et al, 2019; Raffel et al, 2020) consume large amounts of energy Such high-levels of energy consumption adds to the worsening global warming and can cause significant social health and safety impacts (Glo; Rolnick et al, 2019). Recent studies have raised awareness of the carbon footprints and potential energy impacts and suggest ways to estimate and reduce consumption (Strubell et al, 2019; Schwartz et al, 2019). The total energy consumption is computed as the sum of the (utilization × power counter) of the CPU, GPU, and memory, which is adjusted by a compensation constant (Henderson et al, 2020; Strubell et al, 2019) We call this technique software-based power measurement

Objectives
Methods
Results
Conclusion

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.