Luminescent said:
Simon F said:
The volumes themselves take the same sort of fillrate as stencils but this gets done in the ISP (i.e. very high fill rate) part and can carry on in parallel with the texturing.
I assume the ISP is some sort of ROP processor that could output more than 1 modifier volume op per clock in parallel with the rest of the rasterizer pipeline, thus the term "very high fillrate." Is this correct?
I wouldn't say ROP because that, to me at least, implies a 2D processor.
No, the ISP does the (for want of a better term) the rasterisation of triangles and the hidden surface removal (i.e. Z buffering). It can do this at up to a
peak fill rate of,
IIRC, 3.2 GPixels/sec. That's one reason why I said it was quite fast.
Is there stencil buffer support in CLX aside from the modifier volume functionality? If so, is the stencil buffer shared with the z-buffer?
No, the modifier volumes are all there are. You can consider them to be a more efficient but less flexible alternative to stencils.
The other reason why they are efficient is that the original geometry is only submitted once, unlike stencils where it is sent, say, 2/3/4 times. Being a deferred renderer, the visible geometry can be tagged with multiple shading parameters and the modifier volumes indicate to the shading engine which parameter to choose.
No.
Finally, how many internal raster passes (with accumulation buffer) could CLX complete before it sending final output to the framebuffer? Is it theoretically a limitless number of passes?
It'd be limited by memory since, with CLX, the entire scene had to be in memory before rendering commenced, though I suppose you could
- Render one tile at a time devoting all the memory to it and thus giving you a very complex scene
- Go insane, and manually tie display list in a loop and thus render for ever
