test the vld1* functions

[folkertdev]

This adds tests for the vld1 functions using a roundtrip.

It also changes the generated code for unsigned vld1 for big-endian targets. The unsigned version called the signed version, but then applied an additional shuffle. I see no reason that signedness would require a shuffle, so this is probably a quirk in the generator.

cc @adamgemmell if I'm missing anything

[rustbot]

r? @Amanieu

rustbot has assigned @Amanieu.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[Amanieu]

I'm generally happy with the changes, but the large amount of repetition is unfortunate. Is there a way we could reduce this using a macro?

[folkertdev]

I cleaned it up a bunch

[adamgemmell]

Thank you very much for working on this, looks good to me. There are still tests missing for all f16 loads but this is a significant chunk of work and I don't see a reason to block either this patch or the other one on that

[adamgemmell]

Just checked and we should be able to get tests in for any intrinsics missing them by the end of this quarter.