-
Notifications
You must be signed in to change notification settings - Fork 4.9k
Remove phi nodes during rationalize. #53269
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
Conversation
/azp run runtime-coreclr outerloop |
Azure Pipelines successfully started running 1 pipeline(s). |
PTAL @AndyAyersMS @BruceForstall @dotnet/jit-contrib |
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.
LGTM too. Thanks for doing this Sergey.
pin shows 0.1-0.2% jit compilation time improvement (x64 spmi replay on release without pgo), sounds like what I expected. |
finally found the reason for the diff, when we create runtime/src/coreclr/jit/lsrabuild.cpp Line 1891 in bc0fda4
then, when we assign weights we check runtime/src/coreclr/jit/lsra.cpp Line 189 in bc0fda4
so not a "real" diff, just not the best mechanism for these zero-inits. |
Get rid of PHI/PHI_ARG nodes during rationalize phase, these nodes are not used after optimization phases.
This change allows us to
I did not expect any diffs but I have one on x86 windows libraries tests (out of all spmi collections that we have)
it looks like LSRA noise but I am analyzing it now.