On Mon, 16 Nov 2026 22:57:51 GMT, Joe Darcy <[email protected]> wrote:
>> The existing noDoubleRoundingToFloat16 check doesn't take into account the >> exponent / significand interaction of subnormals so only normal-range values >> are eligible to go through the existing fast-path. >> >> >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > Joe Darcy has updated the pull request incrementally with one additional > commit since the last revision: > > Add test coverage. test/jdk/jdk/incubator/vector/BasicFloat16ArithTests.java line 649: > 647: // if the starting value is even; otherwise, it should round > up. > 648: float roundedBack = valueOf( halfWay).floatValue(); > 649: float roundedBackNeg = valueOf(-halfWay).floatValue(); What are these locals good for? They don't seem to be used. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/33392#discussion_r4600686267
