Hardknock said:
Xbdestroya, lets put all semantics, unproven facts(either way) and patents aside for a second. I'm really curious of your honest opinion on this. PS2's CPU and GPU architecture were built from the ground up from a joint effort by Sony and Toshiba. Reason given by Sony is they wanted to control costs by holding all patents and manufacturing in-house. Now lets fast forward to PS3. Cell is being built by the same people again(+ IBM this time), and has similar design philosophies as PS2. They have shown off many graphic demos of things just running on Cell(why would they do that?). Cell was obviously made with graphic capabilities in mind(aswell as many other things.) Discussions with Nvidia weren't started till around 2003. And even then( in the link you posted) nothing was started with Nvidia yet. Sony stated they are investigating internal and external options and hadn't made a decision. What other internal graphics processing had they been developing or had a patent on? Going by their past and what they implemented with PS2 isn't it a no brainer? To me it only seems like common sense that Cell was going to not only be a CPU but a GPU aswell, and I know we don't have proof one way or the other. But what do you think?
I will happily address your points Hardknock.
1) My article simply indicates that word was NVidia was in talks with Sony. It neither confirms nor denies a deal signed by the parties, and Sony makes no statement either way - the "internal/external" comment was made by a third party, so don't read too much into it.
2) Yes, I believe that an internal solution (with Toshiba in the vein of the GS) was the original plan when the project started out. I've said so many times, and in fact if you go back in my history, I've started threads on the subject before. A Cell-GPU is different by the way than a Cell
acting as GPU, which is what you seem to indicate in your post.
BUT, the point of this debate right now is nothing other than to:
a) Indicate that Sony's decision was not necessarilly rushed, by pointing out that they have been in talks with NVidia for years.
and..
b) To show that for all of our speculation, there has never been anything to indicate that the GPU project with Toshiba went
bad as many have hypothesized, making the NVidia switch last minute. Indeed, we have zero official information with regard to such a project whatsoever - in spite of the general consensus (of which I am a part of as well).
It would make perfect sense to me for STI's efforts to have been primarily Cell-oriented for the last five years, with Sony and Toshiba getting to the GPU later on. Now, assuming in the formative stages of that that Sony was approached by NVidia, it stands to reason that NVidia could well have made a compelling case for itself. Lower R&D costs (by piggybacking on G70), use of OpenGL for PS3 development, and expertise from a company that designs these things regularly rather than every five years.
Again, I
do think that a Cell-based GPU was in the cards at one point or another - BUT, this debate is simply about the fact that the Cell-GPU exists in our minds only as a plausible scenario - not something that was ever announced. The Toshiba SCC is *not* an example of such a project.
And for the last two pages Powderkeg has been trying to prove that such a project *was* announced, to very poor effect - just like for the two pages before that, he was trying to prove that an NVidia and Sony relationship would have *had* to be mentioned in their quarterly proxy statements. (Incorrectly)
As for Cell's awesome imaging abilities, it's demonstrations have been ray-casting based, an area that right now rests with strong CPU power rather than GPU power. Since the Cell is being targeted towards the imaging markets, it only makes sense to show those demos. Plus of course, it's very cool.