RFR: 8325144: C1: Optimize CriticalEdgeFinder [v2]

[Denghui Dong]

On Tue, 13 Feb 2024 02:55:35 GMT, Dean Long <dlong at openjdk.org> wrote:

>> Denghui Dong has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   update
>
> src/hotspot/share/c1/c1_Instruction.hpp line 2427:
> 
>> 2425:   BlockBegin* _from;
>> 2426:   BlockBegin* _to;
>> 2427:   int _index;
> 
> We don't really need `_to` anymore, because sux_at(_index) == _to if `insert_block_between` hasn't already processed this edge.  We can detect if sux_at(_index)  has been replaced in a couple different ways:
> 1. set a block flag on the newly inserted block
> 2. check the block_id on the newly inserted block, assume block ids increase monotonically.
> But maybe the added complexity isn't worth it?  What's the performance improvement with this change vs the footprint increase for this new field?

Inspired by your comment.
Only the method `BlockBegin::insert_block_between` sets `critical_edge_split_flag`, so we can check if that flag is set or not.
Then there is no footprint problem.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17674#discussion_r1487141801