_mm256_round_ps operation

[Paul Sandoz]

Not yet. There is a prototype here:

  https://github.com/openjdk/valhalla/tree/lworld%2Bfp16

Support for operations is currently very limited so I doubt it is ready for your use cases. We anticipate making more progress especially when Valhalla previews, since that will allow us to add a float16 scalar value class.

Paul.

> On Jun 3, 2024, at 7:48 AM, Andrii Lomakin <andrii0lomakin at gmail.com> wrote:
> 
> Hi Paul.
> Thank you!
> 
>> Presumably this is more beneficial for lower precision floating point numbers? Like float32 or even float16?
> 
> Does Vector API allow it to work with F16? I would be curious to try.