[vectorIntrinsics] RFR: VectorMask.intoArray intrinsics

[Mai Đặng Quân Anh]

On Thu, 4 Nov 2021 06:29:59 GMT, Mai Đặng Quân Anh <duke at openjdk.java.net> wrote:

> Hi,
> 
> This small patch applies intrinsics for `VectorMask.intoArray`. This uses the existing `VectorSupport.store` with the signature changes similar to `VectorSupport.load`.
> `VectorMask.intoArray` implementation is moved to each concrete class, `VectorMask.toArray` implementation is changed similar to `Vector.toArray`, avoiding referring to `getBits`.
> In the hotspot, a small change is made to wrap the node in a `VectorStoreMask` node before passing it to `StoreVectorNode`, similar to how `VectorMask.fromArray` is handled there.
> 
> Thank you very much.

I will close the pull request now. A small point here, do you think `intoArray` should reside in each concrete implementation like other operations to ensure constant information being passed to the compiler. Furthermore, I see that `VectorMask.toArray` is still implemented through `getBits().clone()`, I think we can use `intoArray` for this.

Thank you very much.

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

PR: https://git.openjdk.java.net/panama-vector/pull/160