Make Vec3A implement NoUninit or Possibly Pod

[vrtgs]

Implement NoUnint and possibly Pod on Vec3A since sometimes its useful to have a byte view of the array for example, vec3 in wgsl has an align of 16 so it has the same layout as vec3a so it would be nice to have simpler interop between the two. its also defined that its 16 byte aligned, and simd vector types can't have uninit fields, and the bytes in padding don't really matter, and it would be trivial to make the scalar just put 0 for padding, so it seems pretty straight forward and simple (?)

[bitshifter]

It Vec3A currently implements AnyBitPattern, What functionality does NoUninit offer over AnyBitPattern?

The oldest open glam bug #122 is to add padding to Vec3A, I didn't want to for a while until I'd added neon, but most major platforms should have a simd implementation now so less concerned about adding overhead to the scalar Vec3A these days.

[bitshifter]

I'm going to close this since it's covered by #122 already.