missing family of floating point methods for rounding values

[royaltm]

Currently none of the fp rounding methods are implemented: .round() .trunc() .floor() .ceil(), .fract() nor signum(). Is it possible to add them?

[gnzlbg]

Yes, this shouldn't be too hard.

[burrbull]

You can take them from my port of sleef:
https://github.com/burrbull/sleef-rs/blob/e4995e1c66837a136c17538439e76d6054c95bff/src/f32x.rs#L651
https://github.com/burrbull/sleef-rs/blob/e4995e1c66837a136c17538439e76d6054c95bff/src/f64x.rs#L747

[CarterLi]

Any progress?

How can I cast f32x8 to __m256 for core::arch::x86_64::_mm256_floor_ps?