Lovely Toshiba..

[nAo]

Toshiba has so many patents on different GPUs and graphics algorithms that they HAVE to do some work for the PS3 in this regard..IMHO.
Here we go:

(A hw architecture devoted to subdivision surfaces)
Polygon generating apparatus and drawing system

(A patent filed by Toshiba guys assigned to IBM!)
Data storage apparatus, computer apparatus, data processing apparatus, and data processing method

(A GPU..)
ONE-CHIP IMAGE PROCESSING APPARATUS


(Another GPU!!)
Image drawing system and image drawing method

(What is this? no time to read it but it SEEMS interesting
System and method for processing image, and compiler for use in this system

I just found those patents 10 minutes ago..but no time to read them..or my boss will fire me

ciao,
Marco

 

[one]

My old post has a couple of more

 

[Panajev2001a]

FIG.11, pag 6 of the drawings: that's an F-Buffer.

 

[London Geezer]

FIG.11, pag 6 of the drawings: that's an F-Buffer.

How exciting is that?!?!?!!!!

 

[jvd]

FIG.11, pag 6 of the drawings: that's an F-Buffer.

How exciting is that?!?!?!!!!

An F buffer ? as in the patented tech in the voodoo 5 ? If true then tosibia ias going to have a lot of fun with nvidia in court

 

[Farid]

FIG.11, pag 6 of the drawings: that's an F-Buffer.

How exciting is that?!?!?!!!!

An F buffer ? as in the patented tech in the voodoo 5 ? If true then tosibia ias going to have a lot of fun with nvidia in court

http://www.beyond3d.com/reviews/ati/r350/index.php?p=3

SMARTSHADER 2.1 does not introduce in Radeon 9800 any new features, effects or increases in the number of shader instructions the hardware can handle in a single pass, and all this is the same as R300, but it does support, according to ATI, 'Unlimited Shader Instructions. This is facilitated by the use of the F-Buffer, or 'Fragment-Stream FIFO buffer'.

OMG, Nvidia will sue Ati, and then they'll buy them 1!1!1!

jvd, there's a lot of differents ways to implement a tech.

 

[jvd]

OMG, Nvidia will sue Ati, and then they'll buy them 1!1!1!

jvd, there's a lot of differents ways to implement a tech.

OMG your an idiot .

Yes i had the wrong buffer . Ati does have an F buffer , I'm pretty sure 3dfx had a buffer that allowed it to do some pretty cool things i think now it was the T buffer.

The point of the matter is if it infringes on any of these companys patents these companys can sue .

 

[Farid]

OMG your an idiot .

I was kidding... Insults were not "expected"... At least not coming from the forum's actual moderator. Anyway, it's not like i'm surprised or anything.

The point of the matter is if it infringes on any of these companys patents these companys can sue .

The point is simpler than that, you only heard the word F-Buffer (Confused with T-Buffer, doesn't change the(your) point) and said:

An F buffer ? as in the patented tech in the voodoo 5 ? If true then tosibia ias going to have a lot of fun with nvidia in court

A word doesn't give ANY hint about its actual hardware implementation...
So how would you know about any possible patents infringement?
You just shooted from the hip, i corrected your claims, that's all.

No need to try to twist the stuff... Or worse start name calling...

 

[jvd]

A word doesn't give ANY hint about its actual hardware implementation...
So how would you know about any possible patents infringement?
You just shooted from the hip, i corrected your claims, that's all.

No need to try to twist the stuff... Or worse start name calling...

SIgh panajev2001 said its an F-buffer. IF its an F buffer which ati has patented or is in the patent stages and its used in a device , any device ati can sue toshibia .

And the your an idiot reply was for the childish OMG that u used .

 

[nAo]

geeez..don't fight guys please...
Saying it's an F-Buiffer it really doesn't mean anything..
Often the same idea can be implemented in so many ways..

 

[Farid]

IF its an F buffer which ati has patented or is in the patent stages and its used in a device , any device ati can sue toshibia .

So what's the point now? IF the Toshiba's F-Buffer implementation is identical to Ati's one, THEN Ati would have the right to sue Toshiba for patent infringement.

...Yeah... But that's stating the obvious, and that wasn't the point at all, BTW.

The point was, simply, Toshiba doesn't risk any lawsuits just because they have a F-Buffer (or T-Buffer, or Z-Buffer, ...) as long as their hardware implementation of the F-Buffer differs from the patented ones.

And the your an idiot reply was for the childish OMG that u used .

I sent you a PM about that.

 

[one]

SIgh panajev2001 said its an F-buffer. IF its an F buffer which ati has patented or is in the patent stages and its used in a device , any device ati can sue toshibia .

What I found interesting is, has ATI patented it already? If not, look at the date of Toshiba patent (published and filed).

 

[London Geezer]

Errr... About my "How exciting is THAT?!?!". I was obviously joking, seeing how Pana got all excited after a F-buffer! I mean...

 

[passerby]

It seems that we have figured out quite a few mysteries already.

How can they appeal to the cg/visual content industry? It seems that the GPU mirrors a reyes-like renderer. If done properly, it will be very attractive to the studios that do this kind of work.

We were also wondering what wizardry must be involved to be able to squeeze so many 32GF APUs onto a piece of silicon. Since these APU parts are probably GPU components, they probably have specialized customizations dedicated to the work they are supposed to do. In which case we don't need crazy clockspeeds to satisfy the GF claims. Now it seems more credible.

Funny thing is we were wondering what the GPU is going to be like, but we were actually looking at it all along.

Now, me wonder what the 'CPU part' will look like. Maybe follow a 'Cell design' as well, though I doubt they will use the specialized 'GPU part APUs' for it.

OK I'll shut up now. It's still a looonnng way to PS3 release....

 

[nAo]

It seems that we have figured out quite a few mysteries already.

what?! how have we figured out mysteries?? I still don't have any idea on how PS3 GPU is.
To be fair I don't think it will be reyes-based or any other exotic thing.
I believe it will be a standard part..with less features than Xenon GPU.
I hope to be wrong..but according Sony experience in developing GPUs I don't see them to produce something that has:
full floating point pipelines, textures and frame buffers.
multiple render targets
a lot of textures per pass (max 4..)
decent AA..
anisotropic filtering.
[insert other item here..]

How can they appeal to the cg/visual content industry? It seems that the GPU mirrors a reyes-like renderer. If done properly, it will be very attractive to the studios that do this kind of work.

Maybe..but I'm not so sure..
HW itself doesn't appeal industry. They need some sw to run on the hw...you know

We were also wondering what wizardry must be involved to be able to squeeze so many 32GF APUs onto a piece of silicon.

In fact a lot of people think this is possible..but maybe not feasible in this time frame.

 

[j^aws]

It's quite conceivable that the GPUs APUs will have more SIMD units than the CPUs APUs but clocked lower and with more specialization? Personally I think the CPU + GPU will be flexible enough to do pretty much what you like. Moving on from the flexiblity provided by the EE+GS.

 

[nAo]

It's quite conceivable that the GPUs APUs will have more SIMD units than the CPUs APUs but clocked lower and with more specialization? Personally I think the CPU + GPU will be flexible enough to do pretty much what you like. Moving on from the flexiblity provided by the EE+GS.

I don't know..but I do know APUs as we know them from patents don't seem to be good candidates for pixel processing, imho.

ciao,
Marco

 

[Inane_Dork]

Can someone tell me a short history of Toshiba's GPU experience? After DeltaChrome and Volari, I'm not very optimistic about second-tier companies making 1st class products.




Reading this thread reminded me of all the self-imposed depression I got from the GeForce FX. For whatever reason, I had assured myself it stood a good chance of being a deferred rendering device. Not only was I dead wrong on that, but the card didn't really make up for it in other ways.

Now, I'm not saying the PS3 will be like that, but don't get your hopes up so high no one can reach them.

 

[Paul]

 

[j^aws]

It's quite conceivable that the GPUs APUs will have more SIMD units than the CPUs APUs but clocked lower and with more specialization? Personally I think the CPU + GPU will be flexible enough to do pretty much what you like. Moving on from the flexiblity provided by the EE+GS.

I don't know..but I do know APUs as we know them from patents don't seem to be good candidates for pixel processing, imho.

ciao,
Marco

I thought the APUs would deal with the vertex processing and the pixel processing would be dealt with these Salc/Salp babies...?

http://www.beyond3d.com/forum/viewtopic.php?t=12899&highlight=salc

This fits nicely with the REYES pipeline with massive amounts of vertex processors and relatively simple per pixel processors with massive amounts of fillrate (IIRC, that patent describes fillrate of the order of 10s of Gigapixel/sec).

Can someone tell me a short history of Toshiba's GPU experience? After DeltaChrome and Volari, I'm not very optimistic about second-tier companies making 1st class products.

Here's a recent link...

http://www.beyond3d.com/forum/viewtopic.php?p=366915#366915

Now, I'm not saying the PS3 will be like that, but don't get your hopes up so high no one can reach them.

Can't speak for anyone else but my hopes are pretty much irrelevant 'cause I'm ultimately interested in the games on all the next gen systems! ...But Cell/PS3 interests me 'cause it's not on any other technology roadmap like Xenon, Revolution and the PC! Variety is the spice of life!

Btw, it's not just Toshiba involved in this project but they have a pool of talent from both Sony and IBM!