x86: implemented _mm{,256}_maskload_epi{32,64}

[mrowqa]

Hi,

just wanted to ask if I'm doing this in the right way. I will add the documentation and tests soon. Then I will also implement the mask stores.

[alexcrichton]

Looks great, thanks! r=me w/ tests and docs :)

[mrowqa]

@alexcrichton tests and docs are ready. I will create new PR for mask stores, so you can merge this if everything is okay :)

[mrowqa]

Hm, I have just noticed that I forgot about mask loads for f32 and f64 (there was no TODO comment in avx2.rs). I will add them soon.

[BurntSushi]

@mrowqa Do you want to do that in a separate PR? That way we can just merge this now.

[mrowqa]

It is fine to merge this now. I will implement the rest of masked loads in a new PR. I have changed the title to something more accurate.
Edit: actually, _mm{,256}_maskload_{ps,pd} are implemented.

[alexcrichton]

@mrowqa I think this may need a rebase now?

[mrowqa]

As I wrote in #155, it is based on this PR (#152) - that is why I asked you to merge this one before the second one. You have already merged commits from this PR, so it can be safely closed without merging.

[alexcrichton]

Ah sorry about that!

[mrowqa]

No worries. I wanted to contribute and you merged the changes anyway :)