RFR: 8353290: C2: Refactor PhaseIdealLoop::is_counted_loop() [v2]
Cesar Soares Lucas
cslucas at openjdk.org
Sat Apr 12 04:15:29 UTC 2025
On Thu, 10 Apr 2025 15:30:24 GMT, Kangcheng Xu <kxu at openjdk.org> wrote:
>> This PR refactors `PhaseIdealLoop::is_counted_loop()` into (mostly) `CountedLoopConverter::is_counted_loop()` and `CountedLoopConverter::convert()` to decouple the detection and conversion code. This enables us to try different loop configurations easily and finally convert once a counted loop is found.
>>
>> A nested `PhaseIdealLoop::CountedLoopConverter` class is created to handle the context, but I'm not if this is the best name or place for it. Please let me know what you think.
>>
>> Blocks [JDK-8336759](https://bugs.openjdk.org/browse/JDK-8336759).
>
> Kangcheng Xu has updated the pull request incrementally with one additional commit since the last revision:
>
> reviewer suggested changes
src/hotspot/share/opto/loopnode.hpp line 267:
> 265: const TypeInteger* trunc_type = nullptr;
> 266: };
> 267: static TruncatedIncrement match_incr_with_optional_truncation(Node* expr, BasicType bt);
Just a drive-by comment: this function and some others that you created are returning a value by copy, for performance reasons it may be better to return a reference or even a pointer as usually is the case in HotSpot.
src/hotspot/share/opto/loopnode.hpp line 1276:
> 1274:
> 1275: struct LoopExitTest {
> 1276: Node* cmp = nullptr;
I think it might be a good idea to use a more specific node type if you know before hand what kind of node the field will point to - in this case CmpNode*
src/hotspot/share/opto/loopnode.hpp line 1343:
> 1341: _loop(loop),
> 1342: _iv_bt(iv_bt) {
> 1343: assert(phase != nullptr, "");
Would be nice to have an explanation of why the expression should be true.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/24458#discussion_r2040546046
PR Review Comment: https://git.openjdk.org/jdk/pull/24458#discussion_r2040546249
PR Review Comment: https://git.openjdk.org/jdk/pull/24458#discussion_r2040546536
More information about the hotspot-compiler-dev
mailing list