On Wed, 9 Sep 2026 20:44:14 GMT, Chen Liang <[email protected]> wrote:
> The return value of `convert0('X', vspecies().withLanes(laneType))` is
> polymorphic, making the subsequent call `swapIfNeeded` megamorphic and thus
> causing allocations. We have the constant species available and should
> execute code on the species instead.
>
> ---------
> - [x] I confirm that I make this contribution in accordance with the [OpenJDK
> Interim AI Policy](https://openjdk.org/legal/ai).
I noticed my initial version might have accidentally changed the behavior for
BE `Vector::reinterpretAsBytes()` and have reverted that part's behavior. Can
you help test again?
Instead of introducing a new benchmark, I introduced a new IR test from the
benchmark in 8391913 and verified it has no potential allocations seen as in
master. I think such an IR test is more resilient. @merykitty can you help
check it out?
-------------
PR Comment: https://git.openjdk.org/jdk/pull/32803#issuecomment-5627432062