I don't think Bryce's (or any other program's) rendering will ever take advantage of graphics card features. I'm not aware of any flow of processed visual data back to the CPU from a graphics card. A graphics card processes data on the fly for the purpose of throwing it at a monitor (where it is shown for a fraction of a second and then dumped). Bryce already makes some use of hardware acceleration when running its work window in OpenGL or Direct3D. While I'm sure this can be improved in terms of work window textures, etc, it will never increase render speed. There are some special PCI cards about with inbuilt processors with good floating point performance. 3D programs can be written to take advantage of these to speed up rendering. But in our case that would require a Bryce6, and I don't think we're ever going to see one of those.