nVidia to have TBDR?

[digitalwanderer]

Please don't kill me if the title is wrong Baron, but that's what Warp2search is reporting that DataFuse is saying:

In regards to the tile based rendering, Nvidia now ownes this technology (happened during the Gigapixel -- 3dfx IP purchase). So just think of what might be possible, a GeForce 6800 Ultra with tile-based rendering? That would blow any type of competition out of the way by far. But of course all of this might not be as easy to integrate as you might think/hope.

 

[K.I.L.E.R]

Is there something in that article that we already don't know?
I haven't read it yet, just the quote which states what's been obvious for the last decade.

I bet if the R300 was a TBDR it would blow the NV40 out of the water.

 

[PaulS]

Please don't kill me if the title is wrong Baron, but that's what Warp2search is reporting that DataFuse is saying:

In regards to the tile based rendering, Nvidia now ownes this technology (happened during the Gigapixel -- 3dfx IP purchase). So just think of what might be possible, a GeForce 6800 Ultra with tile-based rendering? That would blow any type of competition out of the way by far. But of course all of this might not be as easy to integrate as you might think/hope.

Not for a couple of generations yet, AFAIK. May not even happen at all.

 

[Geo]

I hear Jimmy Hoffa is still missing too.

 

[martrox]

I hear Jimmy Hoffa is still missing too.

No....he and Elvis own a gas station in southern Montana.........

 

[Geo]

Well, someday the true and rightful King will come to pull the sword from the stone, and there will be joy throughout the land. Good to check in once in awhile on that I suppose.

 

[digitalwanderer]

Dumb question: if TBDR is such a bloody pantheon of video performance than why has no one used it in so long?

I'm serious. Everytime I hear people talking about it they're just raving about how great it would be, why hasn't anyone done it?

 

[Ylandro]

Totally agree Dig. If it's really so good, everyone would be using it allready. So there must be some really big disadvantages too!

Didn't the Kyro use tile based rendering? I remember it was blazingly fast on some benchmarks, (considering it was a budget product), but it was equally slow on most other benchmarks.

I can think of two problems:
1) It not as powerfull as some claim...
2) It requires such complete different architecture, that it's impossible to introduce it gradually into new generation videocards.

 

[Heathen]

2) It requires such complete different architecture, that it's impossible to introduce it gradually into new generation videocards.

That probably fits the bill me thinks

Real shame Imagination ever really pursued their TBDR product lineup.

 

[Geo]

I think that memory technology has just managed, barely, to stay "good enough" to not provide a strong enough incentive for that kind of major break in architecture for the big boys. Maybe if DDR had not been developed when it was a change would have been forced earlier.

 

[gkar1]

Dumb question: if TBDR is such a bloody pantheon of video performance than why has no one used it in so long?

I'm serious. Everytime I hear people talking about it they're just raving about how great it would be, why hasn't anyone done it?

That horse has been beaten to bones a million times in here already.
Theres a search link way up top of the page

 

[UKt@xman]

I would suggest that one of the biggest hurdles is that IMGtech own a lot of the patents for TBDR and that it'd be difficult to implement anything commercial without infringing those patents.

 

[Heathen]

That horse has been beaten to bones a million times in here already.

Yep, but we're all running out of X800 speculation.

 

[gkar1]

I would suggest that one of the biggest hurdles is that IMGtech own a lot of the patents for TBDR and that it'd be difficult to implement anything commercial without infringing those patents.

Gigapixel found a way to do it without infringin on those patents didn't they? If so, I don't see why they cant use the tech they acquired as collateral in the 3dfx buyout. It may just be that there's not much substance to the actual IP, after all this is what 4 year old tech now? Assuming it has been sitting on the shelf so to speak.

 

[akira888]

Intel of course found a way to avoid infringing on ImgTec IP with the Extreme Slideshow line of graphics decelerators. In general you can patent an implementation but not a theory or concept, although in practice you can patent just about anything so long as you bamboozle the Patents Office - which is of course not very hard at all. Time machines, faster than light transportation, perpetual motion machines - they're all successfully patented. Not to mention the XOR operator - also patented.

 

[Simon F]

I would suggest that one of the biggest hurdles is that IMGtech own a lot of the patents for TBDR and that it'd be difficult to implement anything commercial without infringing those patents.

Gigapixel found a way to do it without infringin on those patents didn't they? If so, I don't see why they cant use the tech they acquired as collateral in the 3dfx buyout. It may just be that there's not much substance to the actual IP, after all this is what 4 year old tech now? Assuming it has been sitting on the shelf so to speak.

FWIW, in my experience, the USPTO never bother to check for prior art outside of US patents, so if something has been patented first in Europe...

Mind you, I think I've seen 3 US patents for the same thing (from 3 different companies) so they don't even check their own database terribly well!

 

[StealthHawk]

Dumb question: if TBDR is such a bloody pantheon of video performance than why has no one used it in so long?

I'm serious. Everytime I hear people talking about it they're just raving about how great it would be, why hasn't anyone done it?

Because all the big boys have already invested, probably, hundreds of millions in R&D towards improving the efficiency of IMRs. It's hard to justify radical design changes after that, even if the benefits are real.

As for TBDRs being the holy grail of graphics rendering, I'll believe it when I see it. We haven't seen a high end TBDR in like, ever. If PowerVR shows up with Series 5 and its a contender maybe other IHVs will be forced to make the switch. Somehow, I doubt it though.