-
Notifications
You must be signed in to change notification settings - Fork 149
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add nalgrebra-0.33.0
benchmark
#1977
Conversation
Thank you! I agree that given the recent hang, it seems useful to add this crate to the suite, it is quite popular and does some crazy things with the trait and const system. Based on local testing, this should make each benchmark run longer by ~3 minutes. Btw, the |
Dang I had done this locally and may have just forgotten to add it before pushing :/ I’ll fix it in a few minutes after rustconf ends |
(Done) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Maybe we could delete the tests
directory, it has 10k lines, but no problem with leaving it as-is. Otherwise LGTM.
I can do that. Edit: done. |
The 2022 upgrade docs have the following:
But if the difference between nalgebra and bitmaps has increased with the new solver, seems reasonable to add it. |
I’m not at a computer where I could open a revert if something goes wrong, but we should merge this soon so that the new PR enabling the new solver can make use of it |
clean build times - cargo check: 10.2s - cargo build: 10.9s - cargo build --release: 13.6s leaf crate build times: - check full: 4.9s - debug full: 5.4s - opt full: 7.6s
Agree. I just pushed a change that inlined the commit that removed tests, so that we don't have them in git history, hope you don't mind. I'll merge it once CI is green. |
This adds
nalgebra
as a benchmark, because it exercises the new trait solver differently than the old solver. We saw that with the hang it caused while trying to enable it for coherence.Tracking its behavior over time will be interesting for enabling the solver for coherence, but also in the future when the trait solver is enabled entirely.
It's not a very costly benchmark to add, here are clean build times (with dependencies)
And here are the times for the actual leaf crates benchmarks:
I'm opening this as draft for discussion purposes and to see if there are concerns.