Green Algorithms: Quantifying the Carbon Footprint of Computation

[Arne Tarara]

I have read the paper some time ago and can give some summary to the mailing list.

The formula is sadly very basic and also does often not even measure even its basic metrics.

The CPU energy for instance is calculates as Thermal Design Power (TDP) * Usage.

Usage is defined as follows: The energy used by the processor was the power draw multiplied by processing time, scaled by the usage factor. However, processing time could not be known a priori and, on some platforms, tracking could be impractical at scale. Modeling exact processing time of past projects may also necessitate re-running jobs, which would generate unnecessary emissions. Therefore, when this processing time is unknown, the simplifying assumption was made that core usage is 100% of run time (uc =  1 in Equation (1 <https://onlinelibrary.wiley.com/doi/10.1002/advs.202100707#advs2603-disp-0001>)).

Even if they were using the Load*Runtime of the processor it would maybe be a resolution problem involved. They are not too specific on which tool they use with which resolution to get the load.

To be fair: The deal with runtimes of 79h+ for their algorithms, so it probably is less important if you have a 1s or 100ms or less resolution. Given the KDEs scope of sometimes minimal load desktop apps this would pose a problem though I guess.


Besten Gruß / Best regards,

Arne Tarara

Green Coding Berlin
Jablonskistr. 24
10405 Berlin

Tel.: +49 160 947 930 24
Web: https://www.green-coding.org <https://www.green-coding.org/>
Github: https://github.com/green-coding-berlin <https://github.com/green-coding-berlin>
Linkedin: https://www.linkedin.com/in/arne-tarara/ <https://www.linkedin.com/in/arne-tarara/>

> On 25. Apr 2022, at 12:58 AM, Julien Moutinho <julm+kde at sourcephile.fr> wrote:
> 
> Hi energy-efficiency@,
> 
> While reading this article https://physicsworld.com/a/the-huge-carbon-footprint-of-large-scale-computing/
> I've just learnt about this site : https://green-algorithms.org
> the associated paper :
>> Green Algorithms: Quantifying the Carbon Footprint of Computation
> https://onlinelibrary.wiley.com/doi/10.1002/advs.202100707
> and the associated source code :
> https://github.com/GreenAlgorithms/green-algorithms-tool
> 
> It's focusing on high performance computing and cloud computing
> but there's likely bits of methodology and data
> that may be of interest to you in all that.
> 
> Thanks for your work,
> Julien

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-eco-discuss/attachments/20220425/c2d0c8a6/attachment.htm>