On Thu, 3 Sep 2026 07:02:32 GMT, Quan Anh Mai <[email protected]> wrote:

>> On second thought, if the values are sufficiently large, and not just 
>> 128-bit versions of 32- and 64-bit numbers, then I guess overflow could be 
>> unpredictable.  But maybe we shouldn't assume that but instead let profiling 
>> decide.
>
> That's true, but this is only the fall back implementation, so correctness 
> and ease of implementation are more important. We will implement back-end 
> support for these nodes later.

In my experiments, CMove is not often as good as than a diamond from If + a 
Phi, and rarely better. Even if cases where it looks like it should be good! 
But it's not huge... so I agree to favor easiness, at least for now.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/31008#discussion_r3924808806

Reply via email to