8369646: Detection of redundant conversion patterns in add_users_of_use_to_worklist is too restrictive #27900

benoitmaillard · 2025-10-20T14:08:59Z

This PR addresses a missed optimization opportunity in PhaseIterGVN. The missed optimization is the simplification of redundant conversion patterns of the shape ConvX2Y->ConvY2X->ConvX2Y.

This optimization pattern is implemented as an ideal optimization on ConvX2Y nodes. Because it depends on the input of the input of the node in question, we need to have an appropriate notification mechanism in PhaseIterGVN::add_users_of_use_to_worklist. The notification for this pattern was added in JDK-8359603.

However, that fix was based on the wrong assumption that in PhaseIterGVN::add_users_of_use_to_worklist, argument n is already the optimized node. However in some cases this argument is actually the node that is about to get replaced.

This happens for example in PhaseIterGVN::transform_old. If we find that node k returned by Ideal actually already exists by calling hash_find_insert(k), we call add_users_to_worklist(k).
As we replace node k with i, and i as a different opcode than k, then we cannot use the opcode of k to detect the redundant conversion pattern.

...
  // Global Value Numbering
  i = hash_find_insert(k);      // Check for pre-existing node
  if (i && (i != k)) {
    // Return the pre-existing node if it isn't dead
    NOT_PRODUCT(set_progress();)
    add_users_to_worklist(k);
    subsume_node(k, i);       // Everybody using k now uses i
    return i;
  }
...

The bug was quite intermittent and only showed up in some cases with -XX:+StressIGVN.

Proposed Fix

We make the detection of the pattern less specific by only looking at the opcode of the user of n, and not directly the opcode of n. This is consistent with the detection of other patterns in PhaseIterGVN::add_users_of_use_to_worklist.

Testing

GitHub Actions
tier1-3, plus some internal testing
Added a second run for the existing test with -XX:+StressIGVN and a fixed stress seed

Thank you for reviewing!

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8369646: Detection of redundant conversion patterns in add_users_of_use_to_worklist is too restrictive (Bug - P4)

Reviewers

Christian Hagedorn (@chhagedorn - Reviewer)
Emanuel Peter (@eme64 - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/27900/head:pull/27900
$ git checkout pull/27900

Update a local copy of the PR:
$ git checkout pull/27900
$ git pull https://git.openjdk.org/jdk.git pull/27900/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 27900

View PR using the GUI difftool:
$ git pr show -t 27900

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/27900.diff

Using Webrev

Link to Webrev Comment

…orklist

bridgekeeper · 2025-10-20T14:10:18Z

👋 Welcome back bmaillard! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

openjdk · 2025-10-20T14:11:13Z

@benoitmaillard This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8369646: Detection of redundant conversion patterns in add_users_of_use_to_worklist is too restrictive

Reviewed-by: chagedorn, epeter

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 235 new commits pushed to the master branch:

795bf9f: 8351159: Remaining cleanups in cpu/x86 after 32-bit x86 removal
80fcfaf: 8369435: C2: transform (LShiftX (SubX con0 a), con1) into (SubX con0<<con1 (LShiftX a con1))
87e5341: 8370804: G1: Make G1HeapRegionAttr::remset_is_tracked() conform to coding style
... and 232 more: https://git.openjdk.org/jdk/compare/9d11787799718a82087f3c7b0c70abe7dfcead72...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

➡️ To integrate this PR with the above commit message to the master branch, type /integrate in a new comment.

openjdk · 2025-10-20T14:12:02Z

@benoitmaillard The following label will be automatically applied to this pull request:

hotspot-compiler

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

mlbridge · 2025-10-20T14:23:40Z

Webrevs

chhagedorn

Otherwise, looks good, thanks!

chhagedorn · 2025-10-27T08:08:41Z

test/hotspot/jtreg/compiler/c2/TestEliminateRedundantConversionSequences.java

 * @run main/othervm -XX:+IgnoreUnrecognizedVMOptions
 *      -XX:CompileCommand=compileonly,compiler.c2.TestEliminateRedundantConversionSequences::test*
- *      -XX:-TieredCompilation -Xbatch -XX:VerifyIterativeGVN=1110 compiler.c2.TestEliminateRedundantConversionSequences
+ *      -XX:-TieredCompilation -Xbatch -XX:VerifyIterativeGVN=1110


You could either add a separate run with -XX:+StressIGVN without a fixed seed or just add -XX:+StressIGVN here. I guess the latter is good enough.

I agree, and I just added -XX:+StressIGVN to the existing run.

Looks good, thanks!

chhagedorn · 2025-10-27T10:32:41Z

test/hotspot/jtreg/compiler/c2/TestEliminateRedundantConversionSequences.java

 * @run main/othervm -XX:+IgnoreUnrecognizedVMOptions
 *      -XX:CompileCommand=compileonly,compiler.c2.TestEliminateRedundantConversionSequences::test*
- *      -XX:-TieredCompilation -Xbatch -XX:VerifyIterativeGVN=1110 compiler.c2.TestEliminateRedundantConversionSequences
+ *      -XX:-TieredCompilation -Xbatch -XX:VerifyIterativeGVN=1110


Looks good, thanks!

eme64

Looks good to me :)

eme64 · 2025-10-30T12:23:18Z

src/hotspot/share/opto/phaseX.cpp

  // ConvF2L->ConvL2F->ConvF2L
  // ConvI2F->ConvF2I->ConvI2F
  // Note: there may be other 3-nodes conversion chains that would require to be added here, but these
  // are the only ones that are known to trigger missed optimizations otherwise


You may want to update the description, and give a bit of extra information. Because you are saying n does not have to be a conversion, but it may be that n is about to be replaced with a conversion, right?

Yes, in some cases add_users_of_use_to_worklist is called with the node about to be replaced as argument n. The point is that we might replace n with a node that already has other uses, and we only want to notify the uses for which there is a potential change.
But this is in no way specific to this one optimization, so I think adding something here would cause more confusion than anything else. Perhaps we should update the description of add_users_of_use_to_worklist then?

Right, this is not specific to this optimization here. Why not add something at the level of add_users_of_use_to_worklist.

benoitmaillard added 2 commits October 20, 2025 15:51

8369646: Add second test run with -XX:+StressIGVN and stress seed

c63e308

8369646: Fix excissively restrictive pattern in add_users_of_use_to_w…

402def2

…orklist

openjdk bot changed the title ~~8369646~~ 8369646: Test compiler/c2/TestEliminateRedundantConversionSequences.java fails with "Missed optimization opportunity in PhaseIterGVN" Oct 20, 2025

openjdk bot added the hotspot-compiler [email protected] label Oct 20, 2025

benoitmaillard changed the title ~~8369646: Test compiler/c2/TestEliminateRedundantConversionSequences.java fails with "Missed optimization opportunity in PhaseIterGVN"~~ 8369646: Detection of redundant conversion patterns in add_users_of_use_to_worklist is too restrictive Oct 20, 2025

benoitmaillard marked this pull request as ready for review October 20, 2025 14:18

openjdk bot added the rfr Pull request is ready for review label Oct 20, 2025

benoitmaillard added 2 commits October 20, 2025 16:37

Add new issue name to jtreg headers

86ed466

Missing -XX:+UnlockDiagnosticVMOptions

1270663

chhagedorn approved these changes Oct 27, 2025

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Oct 27, 2025

Add -XX:+StressIGVN to run without fixed seed

16842d0

openjdk bot removed the ready Pull request is ready to be integrated label Oct 27, 2025

chhagedorn approved these changes Oct 27, 2025

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Oct 27, 2025

eme64 approved these changes Oct 30, 2025

View reviewed changes

8369646: Detection of redundant conversion patterns in add_users_of_use_to_worklist is too restrictive #27900

Are you sure you want to change the base?

8369646: Detection of redundant conversion patterns in add_users_of_use_to_worklist is too restrictive #27900

Conversation

benoitmaillard commented Oct 20, 2025 • edited by openjdk bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Proposed Fix

Testing

Progress

Issue

Reviewers

Reviewing

Uh oh!

bridgekeeper bot commented Oct 20, 2025

Uh oh!

openjdk bot commented Oct 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openjdk bot commented Oct 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mlbridge bot commented Oct 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Webrevs

Uh oh!

chhagedorn left a comment

Choose a reason for hiding this comment

Uh oh!

chhagedorn Oct 27, 2025

Choose a reason for hiding this comment

Uh oh!

benoitmaillard Oct 27, 2025

Choose a reason for hiding this comment

Uh oh!

chhagedorn Oct 27, 2025

Choose a reason for hiding this comment

Uh oh!

chhagedorn Oct 27, 2025

Choose a reason for hiding this comment

Uh oh!

eme64 left a comment

Choose a reason for hiding this comment

Uh oh!

eme64 Oct 30, 2025

Choose a reason for hiding this comment

Uh oh!

benoitmaillard Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

eme64 Nov 4, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

3 participants

benoitmaillard commented Oct 20, 2025 •

edited by openjdk bot

Loading

openjdk bot commented Oct 20, 2025 •

edited

Loading

openjdk bot commented Oct 20, 2025 •

edited

Loading

mlbridge bot commented Oct 20, 2025 •

edited

Loading