Energy-Languages icon indicating copy to clipboard operation
Energy-Languages copied to clipboard

Python examples

Open xmnlab opened this issue 2 years ago • 1 comments

very interesting initiative!

about the python benchmark, maybe it would be great to check the result using numpy and/or numba (also pypy). because nowadays, no one would use pure python (when possible) when performance is important.

thanks!

xmnlab avatar Sep 15 '22 00:09 xmnlab

Completely agree.

Generally, it would be interesting to allow a fairer comparison between compiled and interpreted languages by extending the data collection script to measure also during the Makescript (compilation) and probably sum it with the results for the first run (to hopefully give a relevant comparison even in the case of languages with virtual machine or interpreter). The energy consumption, time and memory for this should be shown in separate columns of the result tables.

At the level of interpretation and discussion of the results, one could then consider different trade-offs for different use-cases. I.e. for a program that will handle huge datasets but seldom need to be updated, compilation consumption isn't important. But for interactive use on small datasets where the program is tweaked until it works or until the diagrams are annotated as desired, the tradeoff could be completely different. And the second situation is the kind of use where many students and scientists may use Python (typically with numpy for the actual array operations).

erik-mansson avatar Jul 12 '23 07:07 erik-mansson