I've been reading throughout gd.net and Google about hardware depth cubes (rendered from an FBO) and how they are both hardware supported (or not) depending on which threads I read, and which GL/Shader/Extensions versions I am targeting. I'm targeting GL2.0 (GLSL 120) at the moment, and wondering if that's too low hardware requirement for them to be typically supported?
I'm trying to gather metrics at the moment to figure out if it's worth switching to GL3 (most likely 3.2), or not. I'm also wondering if it's just the depth texture lookup that's not supported or the sheer rendering to each of the six cube faces that's not supported for depth cubes? This is all leading up to getting omni-directional shadow mapping to work (ideally on an NVidia 8800 min-spec system), so any thoughts or links to papers/examples are appreciated.