[llvm] [PowerPC] Implement a more efficient memcmp in cases where the length is known. (PR #158657)

[via llvm-commits]

RolandF77 wrote:

The PR description should probably say that the case optimized is not just known length but EQ/NE only. Also that it extends existing handling to 64 bit up to 128 bits.

https://github.com/llvm/llvm-project/pull/158657