Use MSVC vld1q_u8_ex and vst1q_u8_ex intrinsics for NEON code

[ccawley2011]

MSVC provides variants of vld1q_u8 and vst1q_u8 with an additional parameter specifying the alignment in bits. It might be beneficial to make use of this where available.

[mtl1979]

As far as I know, those _ex variants only make difference for AArch32 and are equivalent to non-_ex versions on AArch64.

Unless we get alignment exceptions in the code, I don't see any benefit in using the _ex variants. We only support ARMv7 when using MSVC for 32-bit ARM.