Enable AVX2 optimistically with AVX on NAOT

[saucecontrol]

This resolves the last of the issues that prevented optimistic enablement of AVX2 (and dependent ISAs) when AVX is included in the NAOT --instruction-set list.

[saucecontrol]

cc @tannergooding, thanks for talking this through with me. Changing getPreferredVectorByteLength() worked, but it does have the side effect of reducing the SIMD size for stack zeroing and cpyblk for AVX-only setups (diffs).

I don't know if it's worth introducing yet another vector size calculation method, but that's an option if we wanted to limit the change to only the Vector###.IsHardwareAccelerated JIT implementation.

[saucecontrol]

Never mind... What I really wanted to do was make sure JIT never tried to return NI_IsSupported_Dynamic for IsHardwareAccelerated (which ILC doesn't even look for anyway). It's easier to just say "don't do that".

(no diffs now)

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

[tannergooding]

CC. @dotnet/jit-contrib, @EgorBo