RFR: 8323220: Reassociate loop invariants involved in Cmps and Add/Subs [v5]
Emanuel Peter
epeter at openjdk.org
Thu Feb 29 07:29:53 UTC 2024
On Wed, 28 Feb 2024 23:59:07 GMT, Joshua Cao <duke at openjdk.org> wrote:
>> @caojoshua sorry, I hope this is not too annoying. I think you are doing good work. We will get there :)
>>
>> About the add/sub matching fragility: do you know where the other add/sub are coming from? I wonder if there is anything that we could improve in the IR framework to help with that in the future.
>>
>> I hope that in the future we can drop the `@Check` methods completely, see https://bugs.openjdk.org/browse/JDK-8310533
>> But that will take a while, since we also plan to refactor the IR framework a little.
>
> @eme64 can you take another look?
@caojoshua it looks really good now!
I'm running our internal testing again, will report back.
One more concern I just had: do we have tests for the pre-existing Add/Sub reassociations? Because you now touched the logic around there we should make sure there are at least correctness tests. IR tests are probably basically impossible because it is the same number of Add/Sub nodes before and after the optimization.
I'd like to have another Reviewer look over this as well, therefore:
-------------
PR Comment: https://git.openjdk.org/jdk/pull/17375#issuecomment-1970556420
More information about the hotspot-compiler-dev
mailing list