RFR: 8324890: C2 SuperWord: refactor out VLoop, make unrolling_analysis static, remove init/reset mechanism [v8]

[Emanuel Peter]

On Thu, 8 Feb 2024 18:21:52 GMT, Vladimir Kozlov <kvn at openjdk.org> wrote:

>> Emanuel Peter has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 18 commits:
>> 
>>  - manual merge
>>  - Merge branch 'master' into JDK-8324890
>>  - add VSharedData class
>>  - manual merge
>>  - timing code from JDK-8325159
>>  - handle AutoVectorizeStatus::TriedAndFailed outside autovectorize
>>  - Merge branch 'master' into JDK-8324890
>>  - _vtrace is moved to VLoop
>>  - comment update
>>  - cosmetics
>>  - ... and 8 more: https://git.openjdk.org/jdk/compare/b02599d2...dcade5f3
>
> src/hotspot/share/opto/loopnode.cpp line 4872:
> 
>> 4870:     // Shared data structures for all AutoVectorizations, to reduce allocations
>> 4871:     // of large arrays.
>> 4872:     VSharedData vshared;
> 
> So it is local for each `build_and_optimize()` call and space will be freed by destructor.

Yes, exactly.
Before my change, we had a SuperWord object per `build_and_optimize()` that allocated all the data structures.
So the scope is now still the same.
Except that before it all went to the `comp_arena`.

So before, we might have allocated those data structures multiple times (once per `build_and_optimize`), and grown `comp_arena` each time. Now we put it to a dedicated arena, which is freed in the destructor. So the memory usage should be a little lower that way.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17624#discussion_r1483881585