RFR: 8257561: Some code is not vectorized after 8251925 and 8250607

Vladimir Kozlov kvn at openjdk.java.net
Thu Dec 3 17:05:56 UTC 2020


On Thu, 3 Dec 2020 10:01:20 GMT, Christian Hagedorn <chagedorn at openjdk.org> wrote:

>> I noticed that some our tests are not vectorized since JDK 16 b24. I found it was caused by small RFE change [8250607](https://bugs.openjdk.java.net/browse/JDK-8250607) and incorrect code move in changes for [8251925](https://bugs.openjdk.java.net/browse/JDK-8251925).
>> I reverted those changes and got back all vectorized tests.
>> 
>> Tested: tier1-4, precheckin-comp, 100 runs of renaissance test (fixed by 8251925).
>
> Looks good to me! Thanks for fixing this.
> 
> Maybe you could add a comment at JDK-8250607 with the test that was not vectorized anymore in case we want to further investigate and redo the RFE.

Thank you @chhagedorn, @iwanowww, @TobiHartmann for reviews.

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

PR: https://git.openjdk.java.net/jdk/pull/1584


More information about the hotspot-compiler-dev mailing list