CN: Consistency cleanup

[Spudz76]

• Remove props.isHeavy() logic within the _gr_ functions as they are no longer called unless already CN_GR_[0-5]
• Add IS_CN_HEAVY_XHV to match existing IS_CN_HEAVY_TUBE and use it everywhere the long-form test was used
• Rename IS_HEAVY to IS_CN_HEAVY to better match the IS_CN_HEAVY_(TUBE|XHV)
• Use IS_CN_HEAVY the same everywhere props.isHeavy() was called separately
• Ensure IS_CN_HEAVY is also defined everywhere CN_STEP4 macro was used, as its props.isHeavy() was changed
• Flip ordering of cn_vaes_enabled tests for slightly improved readability, and allows to not validate the other bools unless VAES is available (now in roughly likelihood-to-be-false and importance order)

Tested compile, and valid accepts for cn-heavy/xhv and ghostrider but I do not have any VAES capable rigs, so it still needs a diligence proof test on something that does have it. Also don't have any pools to test the cn-heavy/0 or cn-heavy/tube but I think the cn-heavy/xhv test proves those by association.

Should be a tiny bit nicer since all these are validated once into constexpr-bool's the same way everywhere.

[SChernykh]

Flip ordering of cn_vaes_enabled tests for slightly improved readability, and allows to not validate the other bools unless VAES is available (now in roughly likelihood-to-be-false and importance order)

Do you know how compile-time checks are optimized by the compiler? They are essentially zero-cost.

[SChernykh]

Same goes for all pros.isHeavy() replacements - these are already compile time and your change doesn't change anything in the generated x64 code.

[Spudz76]

OK, following those rules, I've gone the other direction then.

Consistency is when everything is done the same way, whichever way that is. The latest commit pushes everything back the other direction instead (use props.isWhatever() for everything, drop all unneeded local constexprs shadowing the props).