[llvm] [AArch64] Enable preferZeroCompareBranch for AArch64 when we don't have fused cmp+br (PR #150045)

[via llvm-commits]

AZero13 wrote:

> From #145803, do you have any benchmark results? Codesize was better when I tried it, but AFAIU it does mean the branch starts a cycles later. And as I mentioned in the other ticket, Arm has a lsrs that sets flags as it shifts, which AArch64 doesn't have.

Unless cbz takes 2 cycles to do instead of 1, I don't see how 

https://github.com/llvm/llvm-project/pull/150045