-
Notifications
You must be signed in to change notification settings - Fork 793
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
Enable PGO #17286
Enable PGO #17286
Conversation
✅ No release notes required |
/azp run |
Azure Pipelines successfully started running 2 pipeline(s). |
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.
We shouldn't be producing our own profiles, but consume ones which are produced for us, and which we consume via darc
@EgorBo we don't really want to do the profile generation ourselves, can't we just consume what you're publishing already for fsi and contribute some tests for the fsc? |
I guess it's up to you, I'd also recommend consuming the MIBC from nuget and use the same workflow that we use in dotnet/runtime where training is done in a separate repo. I thought that you decided to do the training here in this repo, if you don't then I'd also vote for just consuming the MIBC from the nuget |
Yeah, I'd say it would be easier for us to consume ready to use profiles, in case if something changes in generating them, we don't have to adjust ours |
Closing in favor of #17317 |
Description
Fixes #12636
Fixes #13328
Benchmarks
I published FSI and FSC locally for net 8.0 and win-64 runtime, then ran them against an empty F# script.
FSI results
Before:
First run: ~2.9 s
Later runs: ~1.4 s
After:
First run: ~1.7 s
Later runs: ~0.9 s
FSC results
Before:
First run: ~2.9 s
Later runs: ~1.9 s
After:
First run: ~1.8 s
Later runs: ~1.0 s
The improvements are about 40% in all cases!
Checklist