Difference between TBR and deferred rendering?

[Livecoma]

I was just wondering what the relationship between the two was. From reading other poster's comments it seems the two are not directly related.

So could you have a deferred renderer that does not use tiling?

Or could you have a tiler that does not use deferred rendering?


If any/both of those are true what would the pros/cons be???

Thanks

 

[multigl2]

TBR just breaks the scene up into small tiles to be rendered. Deferred rendering removes pixels before the rendering process. The two are fairly exclusive; it is just easy for the kyro to be TBR and DR because its onboard cache holds one tile.

 

[Althornin]

Is it correct that a TBR with a tile size equal to resolution (entire screen is one tile) would perfom exactly like a standard IMR? Assuming no DR, of course.

 

[Bjorn]

Is it correct that a TBR with a tile size equal to resolution (entire screen is one tile) would perfom exactly like a standard IMR? Assuming no DR, of course.

Isn't that basically a IMR then ?

 

[Dave B(TotalVR)]

Deferred rendering for HSR and TBR are different things completely, but to have the latter without the former is a no-brainer.

 

[MPI]

Doesn't deferred rendering just mean that you collect the entire display list before rendering?

I.e. you 'defer' the rendering until the scene is complete, for some type of analyzing(HSR algorithms, region division et.c.).


Immediate mode, means that you basically render each triangle as they come into the pipe, the chip needs not to know about any display list, no?

 

[MfA]

Deferred rendering isnt really standard terminology (outside this board it seems to usually mean deferred rasterization here). What PowerVR does is called deferred shading.