NVIDIA's chip designation is based on the same cores (G80, G84, etc), but
they seem to be adding "GL onto them to differentiate Quadro's and I doubt
this is merely marketing. Older cards (GF6 series and before) were easily
modded to Quadro's with RivaTuner, but newer ones are a total pain in the
ass and professional applications like Solidworks may not detect them as a
Quadro.

It's more the case that the GL extensions and functions are enabled by the
Quadro BIOS and Quadro drivers, rather than the G80 architecture is
physically different. The GPU, shaders and memory have different different
clock settings. Memory size is larger, as is the PCB. The Nvidia CUDA
Programming Guide 1.1 has most of this stuff in it.