Cubemap Coordinate Intrinsics

[magcius]

Is your feature request related to a problem? Please describe.
AMD has hardware instructions for calculating cubemap coordinates: v_cubeid_f32, v_cubema_f32, v_cubesc_f32, v_cubetc_f32 to calculate the cubemap face ID, major axis, and S/T texture coordinates of the minor axis. These are accessible through the SPIR-V and GLSL extensions AMD_gcn_shader.

Describe the solution you'd like
It would be nice if HLSL had intrinsics for these, both for performance reasons on AMD devices, but also so that people don't need to keep writing or copy/pasting the same fallback code.

Describe alternatives you've considered
A standard header or library for this would be OK too.