Provide x86-64-v2 optimizations

[lemire]

Windows 11 version 24H2, Rocky Linux 9, RedHat Linux 9, all defaults on x86-64-v2 which means that we have SSSE3, SSE4.1, SSE4.2 and POPCNT. We might be able to include more optimizations for these cases.

https://gist.github.com/CKingX/0d555071fa5555c4a982b20e8273077f

[IRainman]

Also since Windows 8 it's not boot without popcnt.

[lemire]

@IRainman We may not be able to support x86-64-v2 optimizations under Visual Studio because it does not set macros.

Having to do runtime dispatching might be too much for a header-only library.