ARMv8 neon support for studio util

[MoeMod]

No description provided.

[a1batross]

Does it make any difference?

Of course, compared with release build.

[MoeMod]

Does it make any difference?

Of course, compared with release build.

I found that AngleQuaternion and QuaternionSlerp takes 30%+ cpu time in multiplayer CS, so in such scenario it may save some performance.

[a1batross]

It would make sense to add such changes to engine's libpublic as well. And write tests, so the results are accurate. :).

[nekonomicon]

BTW may be use SSE version(useful for x86 and e2k) + SIMD everywhere?
It can give performance boost on many architectures.

[a1batross]

@nekonomicon it would be nice to at least start with ARM.

[MoeMod]

BTW may be use SSE version(useful for x86 and e2k) + SIMD everywhere?
It can give performance boost on many architectures.

I have implemented some function with SSE, but didn't get significant performance boost on PC. That's too much redundant work to work on double platforms, so I am working on only arm64 now.

[a1batross]

I mostly approve. Maybe it only needs slight reformatting to be in line with the other SDK code.

Adding a separate SSE variant is out of the scope of this PR.