PGO benchmark results for the library

[zamazan4ik]

Hi! Thank you for the library!

As far as I see, PGO is already integrated into the build scripts. However, after a quick search in the repository I didn't find PGO benchmarks for the library.

Did anyone test performance improvements for the library by enabling PGO? It would be awesome if you could share these numbers with the auditory in some way (at least here, in Issues) - it can convince more people to enable PGO for their builds. Also, I would be happy to add the results to my https://github.com/zamazan4ik/awesome-pgo repository.

Thank you.

[davidhewitt]

We did this mostly following the process we'd already undergone in pydantic-core.

IIRC the PGO did show improvements, we can test the shipped artefacts against locally-built to see the difference. I think @samuelcolvin did this.

Btw see also PyO3/maturin#1840 - I would love for some help on this.