Complete the remaining neon instructions

[SparrowLii]

This PR completes the remaining neon instructions, except for the following instructions:

1. The following instructions are only available in aarch64 now, because the corresponding target_feature cannot be found in the available features of arm:
   vcadd_rot、vcmla、vdot

2. The feature i8mm is not valid:
   vmmla、vusmmla: https://rust.godbolt.org/z/8GbKW5ef4

3. LLVM ERROR(Can be reproduced in godbolt):
   vsm4e: https://rust.godbolt.org/z/xhT1xvGTP

4. LLVM ERROR(Normal in gotbolt, but LLVM ERROR: Cannot select: intrinsic raises at runtime)
   vsudot、vusdot: https://rust.godbolt.org/z/aMnEvab3n
   vqshlu: https://rust.godbolt.org/z/hvGhrhdMT

5. Not implmented in LLVM and cannot be implemented manually:
   vmull_p64（for arm）、vsm3、vrax1q_u64、vxarq_u64、vrnd32、vrnd64、vsha512

Update #148

[rust-highfive]

r? @Amanieu

(rust-highfive has picked a reviewer for you, use r? to override)

[SparrowLii]

There seems to be some problems with the CI of aarch64-apple-darwin. Could get some help?

[hkratz]

There seems to be some problems with the CI of aarch64-apple-darwin. Could get some help?

#1231.

[Amanieu]

#1231 is now merged, can you rebase?

[Amanieu]

Does this mean that we now have 100% coverage of the neon intrinsics? Is there anything still missing?

[pthariensflame]

Does this mean that we now have 100% coverage of the neon intrinsics? Is there anything still missing?

f16 and bf16 stuff are missing, I assume waiting on Rust core support for it. Not sure what else might be.

[SparrowLii]

Does this mean that we now have 100% coverage of the neon intrinsics? Is there anything still missing?

Except for the following instructions that I cannot complete the implement by myself:

1. The following instructions are only available in aarch64 now, because the corresponding target_feature cannot be found in the available features of arm:
   vcadd_rot、vcmla、vdot

2. The feature i8mm is not valid:
   vmmla、vusmmla: https://rust.godbolt.org/z/8GbKW5ef4

3. LLVM ERROR(Can be reproduced in godbolt):
   vsm4e: https://rust.godbolt.org/z/xhT1xvGTP

4. LLVM ERROR(Normal in gotbolt, but LLVM ERROR: Cannot select: intrinsic raises at runtime)
   vsudot、vusdot: https://rust.godbolt.org/z/aMnEvab3n
   vqshlu: https://rust.godbolt.org/z/hvGhrhdMT

5. Not implmented in LLVM and cannot be implemented manually:
   vmull_p64（for arm）、vsm3、vrax1q_u64、vxarq_u64、vrnd32、vrnd64、vsha512

I hope to get some help, or decide whether to not implment them at current.

[SparrowLii]

@Amanieu vcadd_rot, vcmla, vdot need to use fcma, dorprod target features on aarch64. Do you know what corresponding target features should be used on ARM?

[Amanieu]

LGTM! Could you update the intrinsic tests (which compare with the C version) to enable all the newly added intrinsics. You just need to edit the first column of crates/intrinsic-test/neon-intrinsics.csv.

[SparrowLii]

It seems that the intrinsics test tool needs some updates to support multi-vector (such as float32x2x2_t).
@JamieCunliffe Could you help update it? I will try to see if I can fix it, too.

[JamieCunliffe]

I have actually made that change for the types, and a few other changes to the tool so that it works with the intrinsic's that have constraints on them. I plan to open a pull request by the end of the week for that.

I ran this pull request the other day with those changes and there wasn't any differences.

@Amanieu To do the constraints I have had to change the data source, so if you want to merge this when I open the pull request for my changes I can add these to that list.

[Amanieu]

OK, let's revert the intrinsic checker changes in this PR for now so we can merge this.