8311932: Suboptimal compiled code of nested loop over memory segment #16650

rwestrel · 2023-11-14T09:03:56Z

To enable the elimination of long range checks, the loop that contains
the range checks is transformed into a loop nest and the range checks
are changed to operate on int values computed before the loop is
entered. This causes extra overhead out of loop and once, the range
checks are eliminated, can only pay off if the loop is executed for
long enough. This change disable the transformation if the trip count
computed from profile data is too low. This came up with a
MemorySegment API micro benchmarks and improves performance
significantly.

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-8311932: Suboptimal compiled code of nested loop over memory segment (Bug - P2)

Reviewers

Tobias Hartmann (@TobiHartmann - Reviewer) ⚠️ Review applies to a221320e
Christian Hagedorn (@chhagedorn - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/16650/head:pull/16650
$ git checkout pull/16650

Update a local copy of the PR:
$ git checkout pull/16650
$ git pull https://git.openjdk.org/jdk.git pull/16650/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 16650

View PR using the GUI difftool:
$ git pr show -t 16650

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/16650.diff

Webrev

Link to Webrev Comment

bridgekeeper · 2023-11-14T09:04:42Z

👋 Welcome back roland! 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 · 2023-11-14T09:05:57Z

@rwestrel 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 · 2023-11-14T09:08:40Z

Webrevs

TobiHartmann · 2023-11-14T09:58:56Z

TestLongRangeChecks.java fails with -ea -esa -XX:CompileThreshold=100 -XX:+UnlockExperimentalVMOptions -server -XX:-TieredCompilation:

1) Method "public static void compiler.c2.irTests.TestLongRangeChecks.testStridePosScalePosInIntLoopShortLoop1(int,int,long,long)" - [Failed IR rules: 2]:
   * @IR rule 1: "@compiler.lib.ir_framework.IR(phase={DEFAULT}, applyIfPlatformAnd={}, applyIfCPUFeatureOr={}, counts={"_#COUNTED_LOOP#_", "1"}, applyIfPlatform={}, applyIfPlatformOr={}, failOn={}, applyIfOr={}, applyIfCPUFeatureAnd={}, applyIf={}, applyIfCPUFeature={}, applyIfAnd={}, applyIfNot={})"
     > Phase "PrintIdeal":
       - counts: Graph contains wrong number of nodes:
         * Constraint 1: "(\\d+(\\s){2}(CountedLoop\\b.*)+(\\s){2}===.*)"
           - Failed comparison: [found] 0 = 1 [given]
           - No nodes matched!
   * @IR rule 2: "@compiler.lib.ir_framework.IR(phase={DEFAULT}, applyIfPlatformAnd={}, applyIfCPUFeatureOr={}, counts={}, applyIfPlatform={}, applyIfPlatformOr={}, failOn={"_#LOOP#_"}, applyIfOr={}, applyIfCPUFeatureAnd={}, applyIf={}, applyIfCPUFeature={}, applyIfAnd={}, applyIfNot={})"
     > Phase "PrintIdeal":
       - failOn: Graph contains forbidden nodes:
         * Constraint 1: "(\\d+(\\s){2}(Loop.*)+(\\s){2}===.*)"
           - Matched forbidden node:
             * 173  Loop  === 173 167 170  [[ 173 420 174 ]] inner 

2) Method "public static void compiler.c2.irTests.TestLongRangeChecks.testStridePosScalePosShortLoop(long,long,long,long)" - [Failed IR rules: 2]:
   * @IR rule 1: "@compiler.lib.ir_framework.IR(phase={DEFAULT}, applyIfPlatformAnd={}, applyIfCPUFeatureOr={}, counts={"_#LONG_COUNTED_LOOP#_", "1"}, applyIfPlatform={}, applyIfPlatformOr={}, failOn={}, applyIfOr={}, applyIfCPUFeatureAnd={}, applyIf={}, applyIfCPUFeature={}, applyIfAnd={}, applyIfNot={})"
     > Phase "PrintIdeal":
       - counts: Graph contains wrong number of nodes:
         * Constraint 1: "(\\d+(\\s){2}(LongCountedLoop\\b.*)+(\\s){2}===.*)"
           - Failed comparison: [found] 0 = 1 [given]
           - No nodes matched!
   * @IR rule 2: "@compiler.lib.ir_framework.IR(phase={DEFAULT}, applyIfPlatformAnd={}, applyIfCPUFeatureOr={}, counts={}, applyIfPlatform={}, applyIfPlatformOr={}, failOn={"_#COUNTED_LOOP#_", "_#LOOP#_"}, applyIfOr={}, applyIfCPUFeatureAnd={}, applyIf={}, applyIfCPUFeature={}, applyIfAnd={}, applyIfNot={})"
     > Phase "PrintIdeal":
       - failOn: Graph contains forbidden nodes:
         * Constraint 2: "(\\d+(\\s){2}(Loop.*)+(\\s){2}===.*)"
           - Matched forbidden node:
             * 170  Loop  === 170 165 167  [[ 170 403 171 ]] inner

rwestrel · 2023-11-14T12:35:53Z

TestLongRangeChecks.java fails with -ea -esa -XX:CompileThreshold=100 -XX:+UnlockExperimentalVMOptions -server -XX:-TieredCompilation:

Thanks for running testing. Should be fixed now.

TobiHartmann

Looks good to me. I've re-submitted testing and report back once it passed.

openjdk · 2023-11-15T08:21:03Z

@rwestrel 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:

8311932: Suboptimal compiled code of nested loop over memory segment

Reviewed-by: thartmann, chagedorn

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 96 new commits pushed to the master branch:

9faead1: 8319927: Log that IEEE rounding mode was corrupted by loading a library
f33c874: 8319764: C2 compilation asserts during incremental inlining because Phi input is out of bounds
6868b37: 8318826: C2: "Bad graph detected in build_loop_late" with incremental inlining
1d96886: 8319531: FileServerHandler::discardRequestBody could be improved
73e19f6: 8319825: jdk.net/jdk.net.ExtendedSocketOptions::IP_DONTFRAGMENT is missing @SInCE 19
faeea07: 8319747: galoisCounterMode_AESCrypt stack walking broken
b4c2d1c: 8319542: Fix boundaries of region to be tested with os::is_readable_range
c36ec2c: 8316653: Large NMethodSizeLimit triggers assert during C1 code buffer allocation
2db9ea9: 8317723: C2: CountedLoopEndNodes and Zero Trip Guards are wrongly treated as Runtime Predicate
eaa4417: 8319301: Static analysis warnings after JDK-8318016
... and 86 more: https://git.openjdk.org/jdk/compare/7c7f8ea30da7fe552bcd4f2b593fa9aad27dcdb4...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.

chhagedorn

Looks good!

src/hotspot/share/opto/loopnode.cpp

chhagedorn · 2023-11-15T12:24:21Z

test/hotspot/jtreg/compiler/c2/irTests/TestLongRangeChecks.java

-        TestFramework.runWithFlags("-XX:-UseCountedLoopSafepoints");
-        TestFramework.runWithFlags("-XX:+UseCountedLoopSafepoints", "-XX:LoopStripMiningIter=1");
-        TestFramework.runWithFlags("-XX:+UseCountedLoopSafepoints", "-XX:LoopStripMiningIter=1000");
+        TestFramework.runWithFlags("-XX:+TieredCompilation", "-XX:-UseCountedLoopSafepoints", "-XX:LoopUnrollLimit=0");


You could add 8311932 to the @bug numbers above.

Done in new commit.

test/hotspot/jtreg/compiler/c2/irTests/TestLongRangeChecks.java

Co-authored-by: Christian Hagedorn <christian.hagedorn@oracle.com>

rwestrel · 2023-11-15T15:50:21Z

Looks good!

Thanks for reviewing and for the suggestions.

TobiHartmann

All tests passed.

chhagedorn · 2023-11-16T07:11:45Z

src/hotspot/share/opto/loopnode.cpp

+
+  if (range_checks.size() > 0) {
+    // This transformation requires peeling one iteration. Also, if it has range checks and they are eliminated by
+    // Loop Predication, then 2 Hoisted Check Predicates are added for one range check. Finally, transforming a long range check requires


Just noticed that the line wraps are off now - maybe you want to re-wrap these lines. But that's just a minor detail.

chhagedorn

That looks good, thanks for the update.

rwestrel · 2023-11-17T07:52:43Z

@TobiHartmann @chhagedorn thanks for the reviews

rwestrel · 2023-11-17T07:52:49Z

/integrate

openjdk · 2023-11-17T07:54:17Z

Going to push as commit 129c470.
Since your change was applied there have been 105 commits pushed to the master branch:

369bbec: 8319896: Remove monitor deflation from final audit
1588dd9: 8319567: Update java/lang/invoke tests to support vm flags
9727f4b: 8320199: Fix HTML 5 errors in java.math.BigInteger
d6aa7c8: 8314621: ClassNotFoundException due to lambda reference to elided anonymous inner class
52e2878: 8319987: compilation of sealed classes leads to infinite recursion
b05e69f: 8320209: VectorMaskGen clobbers rflags on x86_64
f3ed275: 8319103: Popups that request focus are not shown on Linux with Wayland
9e7a3ae: 8319630: Monitor final audit log lacks separator
87be6b6: 8318757: VM_ThreadDump asserts in interleaved ObjectMonitor::deflate_monitor calls
9faead1: 8319927: Log that IEEE rounding mode was corrupted by loading a library
... and 95 more: https://git.openjdk.org/jdk/compare/7c7f8ea30da7fe552bcd4f2b593fa9aad27dcdb4...master

Your commit was automatically rebased without conflicts.

openjdk · 2023-11-17T07:54:35Z

@rwestrel Pushed as commit 129c470.

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

disable loop nest with small number of iterations

f355e6c

openjdk bot added the rfr Pull request is ready for review label Nov 14, 2023

openjdk bot added the hotspot-compiler hotspot-compiler-dev@openjdk.org label Nov 14, 2023

-XX:+TieredCompilation

8a0dd86

TobiHartmann approved these changes Nov 15, 2023

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Nov 15, 2023

chhagedorn approved these changes Nov 15, 2023

View reviewed changes

rwestrel and others added 4 commits November 15, 2023 16:21

Update src/hotspot/share/opto/loopnode.cpp

784b920

Co-authored-by: Christian Hagedorn <christian.hagedorn@oracle.com>

Update test/hotspot/jtreg/compiler/c2/irTests/TestLongRangeChecks.java

b8f81d4

Co-authored-by: Christian Hagedorn <christian.hagedorn@oracle.com>

Update test/hotspot/jtreg/compiler/c2/irTests/TestLongRangeChecks.java

192e2f4

Co-authored-by: Christian Hagedorn <christian.hagedorn@oracle.com>

review

a221320

TobiHartmann approved these changes Nov 16, 2023

View reviewed changes

chhagedorn reviewed Nov 16, 2023

View reviewed changes

comment

1d1020e

chhagedorn approved these changes Nov 16, 2023

View reviewed changes

openjdk bot added the integrated Pull request has been integrated label Nov 17, 2023

openjdk bot closed this Nov 17, 2023

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Nov 17, 2023

8311932: Suboptimal compiled code of nested loop over memory segment #16650

8311932: Suboptimal compiled code of nested loop over memory segment #16650

Uh oh!

Conversation

rwestrel commented Nov 14, 2023 • edited by openjdk bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Progress

Issue

Reviewers

Reviewing

Webrev

Uh oh!

bridgekeeper bot commented Nov 14, 2023

Uh oh!

openjdk bot commented Nov 14, 2023

Uh oh!

mlbridge bot commented Nov 14, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Webrevs

Uh oh!

TobiHartmann commented Nov 14, 2023

Uh oh!

rwestrel commented Nov 14, 2023

Uh oh!

TobiHartmann left a comment

Choose a reason for hiding this comment

Uh oh!

openjdk bot commented Nov 15, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

chhagedorn left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

chhagedorn Nov 15, 2023

Choose a reason for hiding this comment

Uh oh!

rwestrel Nov 15, 2023

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

rwestrel commented Nov 15, 2023

Uh oh!

TobiHartmann left a comment

Choose a reason for hiding this comment

Uh oh!

chhagedorn Nov 16, 2023

Choose a reason for hiding this comment

Uh oh!

chhagedorn left a comment

Choose a reason for hiding this comment

Uh oh!

rwestrel commented Nov 17, 2023

Uh oh!

rwestrel commented Nov 17, 2023

Uh oh!

openjdk bot commented Nov 17, 2023

Uh oh!

openjdk bot commented Nov 17, 2023

Uh oh!

Uh oh!

rwestrel commented Nov 14, 2023 •

edited by openjdk bot

Loading

mlbridge bot commented Nov 14, 2023 •

edited

Loading

openjdk bot commented Nov 15, 2023 •

edited

Loading