Benchmark suggestion: deepcopy

[eendebakpt]

At the python language summit @markshannon invited ideas for benchmarking. One python method that is currently not in the pyperformance benchmarks is deepcopy (at least not in the benchmarks that execute on my system).

In #91610 (reimplement deepcopy in C) there is a simple benchmarking script for deepcopy. It is a simple execution of deepcopy on a dict and a dataclass. The benchmark is really simple (not a program), but in larger projects the cost of deepcopy was sometimes a significant part of the total execution time.

[ericsnowcurrently]

Would you mind opening an issue on pyperformance?

[eendebakpt]

Opened an issue at pyperformance: pyperformance #199

[AlexWaygood]

Would you mind opening an issue on pyperformance?

Ah, should I update my blog post on Mark's talk? I believe at the moment I have a sentence in there inviting people to open issues here if they have benchmarking suggestions, but maybe pyperformance would be better?

[ericsnowcurrently]

Would you mind opening an issue on pyperformance?

Ah, should I update my blog post on Mark's talk? I believe at the moment I have a sentence in there inviting people to open issues here if they have benchmarking suggestions, but maybe pyperformance would be better?

Good catch! It would probably be better to point them to the pyperformance tracker. Perhaps @markshannon has a good reason for opening the issues here? Otherwise, pyperformance would be better.

[AlexWaygood]

Perhaps @markshannon has a good reason for opening the issues here? Otherwise, pyperformance would be better.

I've just listened back to the recording of the talk, and I don't think Mark actually specified which issue tracker would be best to suggest new benchmarks (he just said "please give us benchmark suggestions"), so I think this is entirely on me :)

I'll go ahead and update my article.

[AlexWaygood]

Fixed: https://pyfound.blogspot.com/2022/05/the-2022-python-language-summit_2.html

[ericsnowcurrently]

You're the best!