Towards Accurate and Reliable Energy Measurement of NLP Models
This work addresses the need for reliable energy measurement to inform design choices in large-scale NLP, though it is incremental as it highlights existing inaccuracies without proposing a new solution.
The paper tackled the problem of inaccurate software-based energy measurements for NLP models by quantifying errors using a hardware power meter, revealing that existing methods fail to account for hardware differences and non-linear resource utilization effects.
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 measurements 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 measurements 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.