This is a reland of baecb7d5

Directly use NEON instructions on ARM/ARM64 on MSVC.

Original change's description:
> Refactor BitCount
>
> POPCNT intrinsics cannot be used without hardware support, so a CPUID
> check and polyfills are required for some CPUs when using MSVC to
> avoid runtime failure.
>
> Other changes include:
> - Clang: use builtins on all platforms to provide exact intent to the
> compiler;
> - MSVC on ARM: use dedicated intrinsics;
> - x86/x64 fallback is now branchless and works in constant time.
>
> Bug: angleproject:4462
> Change-Id: I00fcabda1c842677d8cb4bfd280d932d0d10c0a5
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2102811
> Reviewed-by: Jamie Madill <jmadill@chromium.org>
> Reviewed-by: Geoff Lang <geofflang@chromium.org>
> Commit-Queue: Geoff Lang <geofflang@chromium.org>

Bug: angleproject:4462
Change-Id: Ia58ae00fa4230f77d981d4ba393fa3481806c5b0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2113570Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>