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

8266798: C1: More types of instruction can also apply LoopInvariantCodeMotion #3965

Closed
wants to merge 2 commits into from

Conversation

kelthuzadx
Copy link
Member

@kelthuzadx kelthuzadx commented May 11, 2021

C1 only applies LoopInvariantCodeMotion for instructions whose types are Constant/ArithmeticOp/LoadField/ArrayLength/LoadIndexed. We are possible to apply this optimization for more types of instruction.

Candidates: NegateOp,Convert.

Due to the lack of verification at IR level, it is difficult to write jtreg to check if it transformed, so I can only demonstrate it with a simple program:

// run with -XX:+PrintValueNumbering
static int foo10(int t){
    int sum=12;
    for(int i=0;i<100;i++){
        sum += 12;
        sum += -t;
        sum += (long)t;
    }
    return sum;
}

Before:

[...]
* loop invariant code motion for short loop B1
processing block B1
Value Numbering: insert Constant i10  (size 11, entries 3, nesting 2)
Instruction i10 is loop invariant
processing block B2
Value Numbering: Constant i12 equal to i5  (size 11, entries 3, nesting-diff 1)
substitution for 12 set to 5
Instruction i12 is loop invariant   // only 12 is recongized
Value Numbering: insert Constant i20  (size 11, entries 4, nesting 2)
Instruction i20 is loop invariant
** loop successfully optimized
[...]

After:

[...]
** loop invariant code motion for short loop B1
processing block B1
Value Numbering: insert Constant i10  (size 11, entries 3, nesting 2)
Instruction i10 is loop invariant
  6    0    i10    100
processing block B2
Value Numbering: Constant i12 equal to i5  (size 11, entries 3, nesting-diff 1)
substitution for i12 set to 105
Instruction i12 is loop invariant
  11   0    i12    12
Instruction i14 is loop invariant
. 16   0    i14    -i4
Value Numbering: insert Convert l17  (size 11, entries 4, nesting 2)
Instruction l17 is loop invariant
. 22   0    l17    i2l(i4)
Value Numbering: insert Constant i20  (size 11, entries 5, nesting 2)
Instruction i20 is loop invariant
  26   0    i20    1
[...]

Progress

  • Change must not contain extraneous whitespace
  • Commit message must refer to an issue
  • Change must be properly reviewed

Issue

  • JDK-8266798: C1: More types of instruction can also apply LoopInvariantCodeMotion

Reviewers

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.java.net/jdk pull/3965/head:pull/3965
$ git checkout pull/3965

Update a local copy of the PR:
$ git checkout pull/3965
$ git pull https://git.openjdk.java.net/jdk pull/3965/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 3965

View PR using the GUI difftool:
$ git pr show -t 3965

Using diff file

Download this PR as a diff file:
https://git.openjdk.java.net/jdk/pull/3965.diff

@bridgekeeper
Copy link

@bridgekeeper bridgekeeper bot commented May 11, 2021

👋 Welcome back yyang! 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 openjdk bot added the rfr label May 11, 2021
@openjdk
Copy link

@openjdk openjdk bot commented May 11, 2021

@kelthuzadx 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.

@openjdk openjdk bot added the hotspot-compiler label May 11, 2021
@mlbridge
Copy link

@mlbridge mlbridge bot commented May 11, 2021

Webrevs

@@ -585,7 +589,7 @@ void GlobalValueNumbering::substitute(Instruction* instr) {
if (subst != instr) {
assert(!subst->has_subst(), "can't have a substitution");

TRACE_VALUE_NUMBERING(tty->print_cr("substitution for %d set to %d", instr->id(), subst->id()));
TRACE_VALUE_NUMBERING(tty->print_cr("substitution for %c%d set to %c%d", instr->type()->tchar(), instr->id(), subst->id(), subst->type()->tchar()));
Copy link
Member

@TobiHartmann TobiHartmann May 11, 2021

Choose a reason for hiding this comment

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

Shouldn't the third argument be a char and the fourth an int?

Copy link
Member Author

@kelthuzadx kelthuzadx May 11, 2021

Choose a reason for hiding this comment

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

Good catch, fixed.

Copy link
Member

@TobiHartmann TobiHartmann left a comment

Looks good to me.

@openjdk
Copy link

@openjdk openjdk bot commented May 11, 2021

@kelthuzadx 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:

8266798: C1: More types of instruction can also apply LoopInvariantCodeMotion

Reviewed-by: thartmann, neliasso

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 56 new commits pushed to the master branch:

  • f3b510b: 8266923: [JVMCI] expose StackOverflow::_stack_overflow_limit to JVMCI
  • 548899d: 8266189: Remove C1 "IfInstanceOf" instruction
  • b46086d: 8266874: Clean up C1 canonicalizer for TableSwitch/LookupSwitch
  • 97367c0: 8266808: Search label still uses old search field id
  • 06d7602: 8261158: JVMState should not be shared between SafePointNodes
  • 476994a: 8266531: ZAddress::address() should be removed from SA
  • e828a93: 8261395: C1 crash "cannot make java calls from the native compiler"
  • 3c47cab: 8261034: improve jcmd GC.class_histogram to support parallel
  • ed32e02: 8241187: ToolBox::grep should allow for negative filtering
  • cc03734: 8266925: Add a test to verify that hidden class's members are not statically invocable
  • ... and 46 more: https://git.openjdk.java.net/jdk/compare/04fad70437a43c0f38fd53414b8eace2eac93509...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.

As you do not have Committer status in this project an existing Committer must agree to sponsor your change. Possible candidates are the reviewers of this PR (@TobiHartmann, @neliasso) but any other Committer may sponsor as well.

➡️ To flag this PR as ready for integration with the above commit message, type /integrate in a new comment. (Afterwards, your sponsor types /sponsor in a new comment to perform the integration).

@openjdk openjdk bot added the ready label May 11, 2021
@kelthuzadx
Copy link
Member Author

@kelthuzadx kelthuzadx commented May 11, 2021

Thank you Tobias for the review!

/integrate

@openjdk openjdk bot added the sponsor label May 11, 2021
@openjdk
Copy link

@openjdk openjdk bot commented May 11, 2021

@kelthuzadx
Your change (at version 447dc3b) is now ready to be sponsored by a Committer.

@TobiHartmann
Copy link
Member

@TobiHartmann TobiHartmann commented May 12, 2021

This is non-trivial and requires a second review.

Copy link

@neliasso neliasso left a comment

Looks good.

@TobiHartmann
Copy link
Member

@TobiHartmann TobiHartmann commented May 12, 2021

/sponsor

@openjdk openjdk bot closed this May 12, 2021
@openjdk openjdk bot added integrated and removed sponsor ready rfr labels May 12, 2021
@openjdk
Copy link

@openjdk openjdk bot commented May 12, 2021

@TobiHartmann @kelthuzadx Since your change was applied there have been 57 commits pushed to the master branch:

Your commit was automatically rebased without conflicts.

Pushed as commit 11759bf.

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

@kelthuzadx
Copy link
Member Author

@kelthuzadx kelthuzadx commented May 12, 2021

Thank you Tobias and Nils!

@kelthuzadx kelthuzadx deleted the codemotion_opt branch May 12, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
hotspot-compiler integrated
3 participants