RFR: 8343685: C2 SuperWord: refactor VPointer with MemPointer [v13]

[Christian Hagedorn]

On Thu, 16 Jan 2025 15:05:19 GMT, Christian Hagedorn <chagedorn at openjdk.org> wrote:

>> Emanuel Peter has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   refactor to has_invar_summands and print invar and fix non-product guards
>
> src/hotspot/share/opto/vectorization.hpp line 728:
> 
>> 726: //       to memory align a pointer using the pre-loop limit.
>> 727: //
>> 728: // A VPointer can be marked "invalid", if some of these conditions are not met, or
> 
> You mention here "some of these conditions" which one could guess that you mean that it does not conform to this form. But further down you present the actual cases where a VPointer becomes invalid. Should we maybe first add these invalid cases and only then follow up with your explanation what this means for the aliasing queries? (just swap the two sections)

Maybe add "provably" since you mention "or unknown" in the next sentence.
Suggestion:

// A VPointer can be marked "invalid", if some of these conditions are provably met, or

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

PR Review Comment: https://git.openjdk.org/jdk/pull/21926#discussion_r1918723047