Add LZCNT support, use in cntlzw

[FioraAeterna]

Includes a unit test, but since my CPU doesn't support this instruction I haven't actually tested this yet.

[skidau]

Code looks good. Nice one.

[Sonicadvance1]

This gets hit fairly often in games.
I couldn't find a regression and it's a fairly simple optimization.

[shuffle2]

btw, common usage I've seen is documented in the CWG at 5.4 "The x = y predicate has the value 1 if x = y, and 0 if x ≠ y". It might be a good candidate for small local optimization... Basically, sequence is:

sub rz, ry, rx
cntlzw rz, rz
swri rz, rz, 5
rz = (rx == ry)

[FioraAeterna]

Wow, you're right; almost every single usage of cntlzw seems to be for that purpose.