RFR: 8323220: Reassociate loop invariants involved in Cmps and Add/Subs [v9]

[Joshua Cao]

On Fri, 16 Feb 2024 08:26:12 GMT, Joshua Cao <duke at openjdk.org> wrote:

>> Might it be that sometimes the size was too small, so that the loop was simply fully unrolled?
>
> My mistake. I was reading from a class member variable `size`. Probably cannot be a counted loop because the `size` can change by iteration.
> 
> Should be fine if I just pass a `size` variable.

If I pass `size` variable, I get some weird matching patterns after strip mining transformations. I avoid this by moving `blackhole()` to before the if condition.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17375#discussion_r1492998201