Difference between max/min_element for simd and normal? #372
-
assert_eq!(glam::Vec3::new(0.32, f32::NAN, f32::NAN).max_element(), 0.32);
assert_eq!(glam::Vec3A::new(0.32, f32::NAN, f32::NAN).max_element(), 0.32);Top one is correct, bottom one is wrong. Is this intended? Wrong with Vec4 too, and correct with vec2, so is consistent. Can't find in documentation. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
I'm not sure if this should be considered a bug or more just don't have any expectations when dealing with NaNs. The difference will be due to differences in implementation. |
Beta Was this translation helpful? Give feedback.
-
|
If you are asking, I had expectations created by seeing consistent behaviour, and found nothing telling me it was a bad assumption. Thank you for answer. |
Beta Was this translation helpful? Give feedback.
I'm not sure if this should be considered a bug or more just don't have any expectations when dealing with NaNs. The difference will be due to differences in implementation.