Skip to content
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

Fix optIsLoopEntry to skip removed loops #61527

Merged
merged 1 commit into from
Nov 12, 2021

Conversation

BruceForstall
Copy link
Member

This was preventing block compaction with loop entry blocks in loops
that had been previously optimized away (and thus removed from the
loop table).

There are a few cases where we now delete dead code that was previously
left in the function. There are a number of spurious local weighting
and IG textual asm diffs changes, possibly due to how PerfScore is implemented
(there are some surprisingly large PerfScore changes in a few cases,
despite no change in (most) generated code).

This was preventing block compaction with loop entry blocks in loops
that had been previously optimized away (and thus removed from the
loop table).

There are a few cases where we now delete dead code that was previously
left in the function. There are a number of spurious local weighting
and IG textual asm diffs changes, possibly due to how PerfScore is implemented
(there are some surprisingly large PerfScore changes in a few cases,
despite no change in (most) generated code).
@dotnet-issue-labeler dotnet-issue-labeler bot added the area-CodeGen-coreclr CLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMI label Nov 12, 2021
@ghost
Copy link

ghost commented Nov 12, 2021

Tagging subscribers to this area: @JulieLeeMSFT
See info in area-owners.md if you want to be subscribed.

Issue Details

This was preventing block compaction with loop entry blocks in loops
that had been previously optimized away (and thus removed from the
loop table).

There are a few cases where we now delete dead code that was previously
left in the function. There are a number of spurious local weighting
and IG textual asm diffs changes, possibly due to how PerfScore is implemented
(there are some surprisingly large PerfScore changes in a few cases,
despite no change in (most) generated code).

Author: BruceForstall
Assignees: -
Labels:

area-CodeGen-coreclr

Milestone: -

@BruceForstall
Copy link
Member Author

@BruceForstall
Copy link
Member Author

@dotnet/jit-contrib PTAL

Copy link
Member

@AndyAyersMS AndyAyersMS left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder how many other places get fooled by removed loops.

I assume you are (or have) been looking for this.

@BruceForstall
Copy link
Member Author

I assume you are (or have) been looking for this.

I was prototyping some loop table iterators, including ones that would automatically skip "removed" loops. This involved looking at the various places that currently loop over the table. However, in this case, I found the problem because I had a bug in my iterator that led to unexpected asm diffs!

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
area-CodeGen-coreclr CLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMI
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants