RFR: 8316594: C2 SuperWord: wrong result with hand unrolled loops

Emanuel Peter epeter at openjdk.org
Wed Oct 4 13:07:39 UTC 2023


On Tue, 3 Oct 2023 07:17:12 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

>> Please read description in https://github.com/openjdk/jdk/pull/15864.
>> 
>> Currently, we only compare the first element of the load-pack with the StoreVector. This is not correct, we need to compare all elements of the load-pack with the StoreVector. Otherwise it may be that the first element lays below/before the memory region of the StoreVector (Less), but the last element of the load-pack for example could have overlap with the StoreVector (Equal).
>> 
>> Tier1-6 + stress-testing.
>> Performance testing.
>
> Okay, makes sense. Thanks for the clarification.

@TobiHartmann @vnkozlov merged with master after https://github.com/openjdk/jdk/pull/15864.
And added one more test with unsafe (test3).
I did not change the VM code, only the tests. Quick re-review would be appreciated.

Rerunning tier1-6 and stress testing.

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

PR Comment: https://git.openjdk.org/jdk/pull/15866#issuecomment-1746839449


More information about the hotspot-compiler-dev mailing list