[llvm] [PreISelIntrinsicLowering] Lower bounded memcpy/memmove to masked load/store (PR #212710)

[Simon Pilgrim via llvm-commits]

================

----------------
RKSimon wrote:

Add a common MASK check prefix to the AVX512 cases:
```
; RUN: opt -mcpu=x86-64-v4 -passes=pre-isel-intrinsic-lowering -S < %s | FileCheck %s --check-prefixes=MASK,MASK256
; RUN: opt -mcpu=znver4    -passes=pre-isel-intrinsic-lowering -S < %s | FileCheck %s --check-prefixes=MASK,MASK512
```

https://github.com/llvm/llvm-project/pull/212710