8354348: Enable Extended EVEX to REX2/REX demotion for commutative operations with same dst and src2 by vamsi-parasa · Pull Request #26997 · openjdk/jdk

vamsi-parasa · 2025-08-28T21:09:03Z

This change extends Extended EVEX (EEVEX) to REX2/REX demotion for Intel APX NDD instructions to handle commutative operations when the destination register and the second source register (src2) are the same.

Currently, EEVEX to REX2/REX demotion is only enabled when the first source (src1) and the destination are the same. This enhancement allows additional cases of valid demotion for commutative instructions (add, imul, and, or, xor).

For example:
eaddl r18, r25, r18 can be encoded as addl r18, r25 using APX REX2 encoding
eaddl r2, r7, r2 can be encoded as addl r2, r7 using non-APX legacy encoding

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-8354348: Enable Extended EVEX to REX2/REX demotion for commutative operations with same dst and src2 (Sub-task - P4)

Reviewers

Sandhya Viswanathan (@sviswa7 - Reviewer) Review applies to 9714a9b1
Jatin Bhateja (@jatin-bhateja - Reviewer)
Emanuel Peter (@eme64 - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/26997/head:pull/26997
$ git checkout pull/26997

Update a local copy of the PR:
$ git checkout pull/26997
$ git pull https://git.openjdk.org/jdk.git pull/26997/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 26997

View PR using the GUI difftool:
$ git pr show -t 26997

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/26997.diff

Using Webrev

Link to Webrev Comment

…erations with same dst and src2

bridgekeeper · 2025-08-28T21:09:36Z

👋 Welcome back sparasa! 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 · 2025-08-28T21:09:54Z

@vamsi-parasa 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:

8354348: Enable Extended EVEX to REX2/REX demotion for commutative operations with same dst and src2

Reviewed-by: jbhateja, epeter, sviswanathan

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 192 new commits pushed to the master branch:

76e464b: 8367150: Add a header line to improve VMErrorCallback printing
0fbae80: 8252582: HotSpot Style Guide should permit variable templates
90e81c2: 8367616: RISC-V: Auto-enable Zicboz extension for debug builds
... and 189 more: https://git.openjdk.org/jdk/compare/0d85f076cc32494c1162baea3ea6b0db67136d41...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 · 2025-08-28T21:10:30Z

@vamsi-parasa 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 · 2025-08-28T21:50:28Z

Webrevs

jatin-bhateja · 2025-09-01T12:59:20Z

Hi @vamsi-parasa , thanks for working on this, I am process of validating #26283 and find that additional RA biasing will enable demotion for more cases, with a minimal test case I see following results

Test point:-

Baseline:-

With this patch:-

With additional RA biasing

src/hotspot/cpu/x86/assembler_x86.cpp

jatin-bhateja · 2025-09-02T02:40:59Z

src/hotspot/cpu/x86/x86_64.ad

+instruct addI_rReg_mem_rReg_ndd(rRegI dst, memory src1, rRegI src2, rFlagsReg cr)
+%{
+  predicate(UseAPX);
+  match(Set dst (AddI (LoadI src1) src2));


Will this not be covered by the pattern at line 7103, since ADLC automatically generates a DFA to handle both cases?

Will run experiments to make sure that the RegRegMem pattern also applies to RegMemReg case and remove the newly added match rules if they're redundant. Will update you soon.

Hi @vamsi-parasa, your latest patch does not address this.

Hi Jatin (@jatin-bhateja), please see the latest update which removed the unnecessary match rules for RegMemReg case.

vamsi-parasa · 2025-09-04T20:11:30Z

Hi @vamsi-parasa , thanks for working on this, I am process of validating #26283 and find that additional RA biasing will enable demotion for more cases, with a minimal test case I see following results

Hi Jatin (@jatin-bhateja), thank you for sharing the information about the register allocation biasing PR you're working on that will improve demotion.

src/hotspot/cpu/x86/assembler_x86.hpp

src/hotspot/cpu/x86/assembler_x86.cpp

… passing attribute

sviswa7

Looks good to me.

vamsi-parasa · 2025-09-11T16:25:32Z

Hi Emanuel (@eme64),

Could you please run the tests for this PR?

Thanks,
Vamsi

jatin-bhateja

Hi @vamsi-parasa , Thanks for addressing my comments.

eme64 · 2025-09-12T12:17:23Z

@vamsi-parasa Quickly scanned the patch, looks reasonable. Launching tests 🚀

vamsi-parasa · 2025-09-15T17:51:59Z

@vamsi-parasa Quickly scanned the patch, looks reasonable. Launching tests 🚀

Hi Emanuel (@eme64), could you please let me know if the tests passed?

eme64

Not reviewed in detail, but looks reasonable.
Tests pass :)

vamsi-parasa · 2025-09-16T17:39:12Z

Not reviewed in detail, but looks reasonable. Tests pass :)

Thank you Emanuel! :)

vamsi-parasa · 2025-09-16T18:11:31Z

/integrate

openjdk · 2025-09-16T18:13:40Z

Going to push as commit c41add8.
Since your change was applied there have been 203 commits pushed to the master branch:

075ebb4: 8366229: runtime/Thread/TooSmallStackSize.java runs with all collectors
15d42c6: 8366978: dead code in SunCertPathBuilder
58007c0: 8367619: String.format in outOfRangeException uses wrong format specifier for String argument
... and 200 more: https://git.openjdk.org/jdk/compare/0d85f076cc32494c1162baea3ea6b0db67136d41...master

Your commit was automatically rebased without conflicts.

openjdk · 2025-09-16T18:14:07Z

@vamsi-parasa Pushed as commit c41add8.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

vamsi-parasa · 2025-10-20T16:39:51Z

/backport openjdk/jdk25u

openjdk · 2025-10-20T16:42:03Z

@vamsi-parasa the backport was successfully created on the branch backport-vamsi-parasa-c41add8d-master in my personal fork of openjdk/jdk25u. To create a pull request with this backport targeting openjdk/jdk25u:master, just click the following link:

➡️ Create pull request

The title of the pull request is automatically filled in correctly and below you find a suggestion for the pull request body:

Hi all,

This pull request contains a backport of commit c41add8d from the openjdk/jdk repository.

The commit being backported was authored by Srinivas Vamsi Parasa on 16 Sep 2025 and was reviewed by Jatin Bhateja, Emanuel Peter and Sandhya Viswanathan.

Thanks!

If you need to update the source branch of the pull then run the following commands in a local clone of your personal fork of openjdk/jdk25u:

$ git fetch https://github.com/openjdk-bots/jdk25u.git backport-vamsi-parasa-c41add8d-master:backport-vamsi-parasa-c41add8d-master
$ git checkout backport-vamsi-parasa-c41add8d-master
# make changes
$ git add paths/to/changed/files
$ git commit --message 'Describe additional changes made'
$ git push https://github.com/openjdk-bots/jdk25u.git backport-vamsi-parasa-c41add8d-master

8354348: Enable Extended EVEX to REX2/REX demotion for commutative op…

51df85a

…erations with same dst and src2

openjdk bot added the hotspot-compiler hotspot-compiler-dev@openjdk.org label Aug 28, 2025

vamsi-parasa added 2 commits August 28, 2025 14:17

remove unused instructions

80d8e40

remove trailing whitespaces

bd14470

openjdk bot added the rfr Pull request is ready for review label Aug 28, 2025

anirudhsengar mentioned this pull request Aug 29, 2025

Test GlitchWitcher Workflow adoptium/aqa-test-tools#1096

Open

jatin-bhateja reviewed Sep 2, 2025

View reviewed changes

vamsi-parasa added 2 commits September 2, 2025 11:35

Merge branch 'master' of https://git.openjdk.java.net/jdk into cdemotion

089e72b

nomenclature change

91962f4

sviswa7 reviewed Sep 5, 2025

View reviewed changes

src/hotspot/cpu/x86/assembler_x86.hpp Show resolved Hide resolved

src/hotspot/cpu/x86/assembler_x86.cpp Show resolved Hide resolved

vamsi-parasa added 3 commits September 8, 2025 12:49

undo swap in emit_arith and refactor accordinly

9d3705c

refactor emit_eevex_prefix_or_demote_arith_ndd to use size instead of…

83a22e1

… passing attribute

undo the passing of demotable flag

9714a9b

sviswa7 approved these changes Sep 9, 2025

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Sep 9, 2025

undo new match rules for RegMemReg for commutative operations

012511a

openjdk bot removed the ready Pull request is ready to be integrated label Sep 11, 2025

jatin-bhateja approved these changes Sep 11, 2025

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Sep 11, 2025

eme64 approved these changes Sep 16, 2025

View reviewed changes

openjdk bot added the integrated Pull request has been integrated label Sep 16, 2025

openjdk bot closed this Sep 16, 2025

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Sep 16, 2025

Conversation

vamsi-parasa commented Aug 28, 2025 • edited by openjdk bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Progress

Issue

Reviewers

Reviewing

Uh oh!

bridgekeeper bot commented Aug 28, 2025

Uh oh!

openjdk bot commented Aug 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openjdk bot commented Aug 28, 2025

Uh oh!

mlbridge bot commented Aug 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Webrevs

Uh oh!

jatin-bhateja commented Sep 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jatin-bhateja Sep 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

vamsi-parasa Sep 4, 2025

Choose a reason for hiding this comment

Uh oh!

jatin-bhateja Sep 9, 2025

Choose a reason for hiding this comment

Uh oh!

vamsi-parasa Sep 11, 2025

Choose a reason for hiding this comment

Uh oh!

vamsi-parasa commented Sep 4, 2025

Uh oh!

Uh oh!

Uh oh!

sviswa7 left a comment

Choose a reason for hiding this comment

Uh oh!

vamsi-parasa commented Sep 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jatin-bhateja left a comment

Choose a reason for hiding this comment

Uh oh!

eme64 commented Sep 12, 2025

Uh oh!

vamsi-parasa commented Sep 15, 2025

Uh oh!

eme64 left a comment

Choose a reason for hiding this comment

Uh oh!

vamsi-parasa commented Sep 16, 2025

Uh oh!

vamsi-parasa commented Sep 16, 2025

Uh oh!

openjdk bot commented Sep 16, 2025

Uh oh!

openjdk bot commented Sep 16, 2025

Uh oh!

vamsi-parasa commented Oct 20, 2025

Uh oh!

openjdk bot commented Oct 20, 2025

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

4 participants

vamsi-parasa commented Aug 28, 2025 •

edited by openjdk bot

Loading

openjdk bot commented Aug 28, 2025 •

edited

Loading

mlbridge bot commented Aug 28, 2025 •

edited

Loading

jatin-bhateja commented Sep 1, 2025 •

edited

Loading

jatin-bhateja Sep 2, 2025 •

edited

Loading

vamsi-parasa commented Sep 11, 2025 •

edited

Loading