There is a Python package, CodeCarbon which can wrap code and track the power usage while it's running to provide an estimate of the carbon impact/power usage of a given task. There are other packages too such as Cumulator and Carbon Tracker. Cumulator seems the most robust from what I've seen but also hasn't been updated in a while.
This would be interesting data to provide (wrapped in the necessary caveats, because such calculations are imprecise by definition) with results, given the increased attention for climate impact of large scale computing, etc
A deep dive into various estimation methods can be found here: https://aclanthology.org/2021.sustainlp-1.2.pdf though it is from 2021 so newer, better tools may exist
There is a Python package, CodeCarbon which can wrap code and track the power usage while it's running to provide an estimate of the carbon impact/power usage of a given task. There are other packages too such as Cumulator and Carbon Tracker. Cumulator seems the most robust from what I've seen but also hasn't been updated in a while.
This would be interesting data to provide (wrapped in the necessary caveats, because such calculations are imprecise by definition) with results, given the increased attention for climate impact of large scale computing, etc
A deep dive into various estimation methods can be found here: https://aclanthology.org/2021.sustainlp-1.2.pdf though it is from 2021 so newer, better tools may exist