RFR: 8332537: C2: High memory usage reported for compiler/loopopts/superword/TestAlignVectorFuzzer.java

[Tobias Hartmann]

On Thu, 6 Jun 2024 14:22:48 GMT, Emanuel Peter <epeter at openjdk.org> wrote:

> This bug was a regression of:
> [JDK-8324517](https://bugs.openjdk.org/browse/JDK-8324517); C2: crash in compiled code because of dependency on removed range check CastIIs
> 
> That change was backed out with:
> [JDK-8332829](https://bugs.openjdk.org/browse/JDK-8332829): [BACKOUT] C2: crash in compiled code because of dependency on removed range check CastIIs
> 
> I `git revert` ed the BACKOUT change:
> [JDK-8332829](https://bugs.openjdk.org/browse/JDK-8332829): [BACKOUT] C2: crash in compiled code because of dependency on removed range check CastIIs
> 
> And only then this reproduced (not on master, only with Roland's CastII regression code):
> `~/Documents/jtreg/bin/jtreg -va -s -jdk:/oracle-work/jdk-fork2/build/linux-x64-debug/jdk -javaoptions:"-Djdk.test.lib.random.seed=3249981201344669190" -J-Djavatest.maxOutputSize=10000000 /oracle-work/jdk-fork2/open/test/hotspot/jtreg/compiler/loopopts/superword/TestAlignVectorFuzzer.java`
> 
> Now that we know that it is a regression of a backed-out change, we can enable the MemLimit check again.
> 
> Also the stack-traces were related to CastII code, and looked very similar to those in another duplicate:
> [JDK-8332765](https://bugs.openjdk.org/browse/JDK-8332765): Test compiler/loopopts/superword/TestAlignVectorFuzzer.java still times out after JDK-8327978

Marked as reviewed by thartmann (Reviewer).

-------------

PR Review: https://git.openjdk.org/jdk/pull/19580#pullrequestreview-2102219215