Last call for questions... Get them in today if you want them answered.
19. What is your experience with OpenGL drivers? Do you routinely encounter bugs on either vendor that force you to implement workarounds? Do you use multiple code paths, not only per hardware level (SM2, SM3, SM4), but also per-vendor?
20. This may a little too technical, but do you use a forward or a deferred renderer? If it is the latter, how did you implement FSAA? Is it done as a custom post-process edge-detect filter, or have you implemented some novel algorithm for this?
21. What approach do you take for GI (Global Illumination)? SSAO or something newer/better?
22. What are your thoughts on OpenCL? Are there any plans to use it for e.g. physics in the future?
Edit: 23. What is the triangle budget for a typical scene?