-
-
Notifications
You must be signed in to change notification settings - Fork 290
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
Improve inferability of lineaxis and axis #1968
Conversation
Compile Times benchmarkNote, that these numbers may fluctuate on the CI servers, so take them with a grain of salt. using time
This PR does not change the using time. ttfp time
This PR does not change the ttfp time. |
@IanButterworth, you may want to check this PR out, as I turned quite a lot of observables into "change observables" while at it... |
Does |
What benchmarks? We don't have run time benchmarks and both using + ttfp times are an improvement
Depends on the code, fully typed + inferable code shouldn't change. Only code which calls functions, where the arguments aren't fully inferable may get slower runtime dispatch. |
Iirc the first compile time benchmark I saw here had around -15% on using and maybe 1% more or less on ttfp. With all the other changes now I don't know what's actually improving things. If the ttfp improvements are mostly from change observables it might not be worth having the experimental stuff. Especially since "experimental" hints that it may break soon. |
It was regarded as a bad move to name the module Some more detailed benchmarks:
I think this shows nicely, that |
Oh I thought ttfp would include using time as well. So then this is really just a straight up improvement |
usemoved to use max_methods and optlevel 0 where appropriate #2000@max_methods
for performance insensitive modules