JDK-8346236[2][3] added the auto-vectorization support for Float16.add/sub/mul/div/sqrt/fma/max/min APIs.

Enable inference of the Float16 vector IR for HalffloatVector APIs by leveraging common backend implementation for various targets (x86, aarch64, etc) as per panama-dev proposal[1].

[1] https://mail.openjdk.org/pipermail/panama-dev/2025-August/021100.html
[2] openjdk/jdk#22755
[3] openjdk/jdk#22754