Nv40 16 full pipelines- The Inq.

No, not really. Not directly.

Regardless, one is telling you all transistors, the other is telling you only logic transistors. One isn't any more truthful than the other, depending on the context.

 

Re: Watch your language

I think you're the biggest waster of space in this thread, I think you have no place to talk. Remove your constant ani-nVidia or highly tangential comments and you'll compact this thread treemndously.

This isn't yet another thread to bash nVidia in, or talk about the politic behind their parts. I don't see people raising ATI's past (eg. Pre-R300 when nVidia was kicking their ass around) in the thread on R420's pipelines... wonder why?

How damn hard is it to stay the hell out of every thread with nVidia implimented in the title to post some usless BS about how they didn't "earned the contempt/distrust/scorn of the graphics community." Who cares, start your own thread on that topic - this was about the NV40's pipeline configuration in a technical forum.

Great, nVidia's satan - won't get an argument out of me, just what the hell does it have to do in this thread? Start a concurrent thread. I see no reason why you need to read 15 posts which are nothing but the same tired, old, iterative anti-nVidia comment about how they pist off the community.

They're a company, their job is to sell product by several means. The objection (which I'd agree to) isn't a strictly technical problem, go to the 3D Industry Forum.

 

Leather Godesses Of Phobos - the Sultan's Riddle, and your companion's answer.

"Some say I'm pointless,
yet many are obsessed by me.
I have caused heroic gambles
and sown endless frustration.
Uncounted deaths have I caused.
What am I?"

"That's easy! A grapefruit!"

 

"An object at rest tends to stay at rest, and an object in uniform motion tends to remain in that state of motion, unless an external force is applied to it." --Isaac Newton

Still true after three hundred years. Also works for reputations. R420 and NV40 will be powerful external forces. Pre-release rumors are not. ATI earned the reputation they had before R300 and earned the reputation they enjoy today. Ditto NV re NV3x.

 

Re: Watch your language

Thanks for the input, always good to hear from another fan...but you didn't answer the question. I believe this forum is about 3D technology and hardware, how does trying to decide how credible a hardware rumor is NOT fit into that?

Dio-Thanks!

 

Re: Watch your language

A technical debate on why it can't be true (as several people conducted) is pefectly fine. But your and other's constant reference to nVidia's PR, for example Joe Defuria just provided a great example:

Don't belong here, this isn't technical. Take your extreme bias rooted in the great satan that is nVidia to the 3D Industry Forum where it belongs.

PS. Did you even make a single contibutive post in this thread which is technically rooted? I'm still looking but out of your 11 of so posts I still can't find one... Thanks for the help.

 

I'll second that...

 

The same goes for you Vince.

And please - take you loud mouth, boarish approach elsewhere. We've had discussions like this before and I've warned you before. If you can't converse with some civility then don't at all.

 

Re: Watch your language

You tell me great master of the B3D technical scorecard.

What you want a generic, "It could be 16 pixels per clock at just Z and stencil, and 8 pixels per clock in PS3.0 and texturing modes. This would make it absolutely fly in Doom3, twice the stencil fill-rate of R420 (clock for clock)." type answer? Sorry, I'm a color commentator and leave the technical stuff to those who know. 8)

 

Re: Watch your language

Um, Vince, the objection is criticial to evaluating / sepculating the technical probability of the transistor count and number of pipelines being accurate.

I just explained why it is inevitable.

At the same time, I see no reason why we have to listen to another one of DC's and yours old, iterative rants about rants either. So I guess everything's even then.

Wouldn't it be better to get your technical response to this rumor? We've asked you a few times now....

 

Re: Watch your language

 

I'll add my voice to the chorus of Nvidia pessimism. This is starting to remind me of the NV30 launch, with repeated delays and extraordinary performance claims. I hope they do actually have 16 pipes though, it would make for a great card and almost justify a $500 price tag :]

 

I'll second that, and that would put pressure on ATI for the R500 !

More power, more features, Isn't it all we claim ?

 

I do happen to think people are missing the point over my question if this is 16 real pipelines...

 

Sorry, in the confusion I got lost a bit...are you asking if we think it's a real 16x1 compared to an 8x2 or 4x4 that just 'fakes' it? Or are you telling us that it really is a real 16x1 card?

 

Not sure I understand...havn't many of us questioned that very thing? In other words, it wouldn't surprise me at all if the specs do say 16 pipelines, but it turns out to be more like 8x1 (or 8x2) / 16x0. (8 colored pixels, maybe 8 colored dual textured pixels, or 16 z-only pixels)

Is that totally missing your point?

 

Re: Watch your language

No, it's not at this point as you don't know where this comment came from. Your objection is only applicable if you're debating nontechnical aspects, not the probability of thids being possible vis-a-vis some form of logic gate statistsic.

It's only inevitable because you allow it to be inevitable. The quick turn-to-shit effect isn't inevitable in the comparable R420 thread, is it? Again, there is no reason past paradigms based on PR comments should be repeated in every thread.

PS. The comment regarding 11 useless posts wasn't intended at you. Sorry for the confusion.

 

Lets assume that its 16 pipelines in the way that NV30 is 4 pipelines, what ramifications would you think that would have elsewhere....?

 

Could be. Could not be. It all depends of a lot of factors.

But, I just see something. They seem to want to give a boost on Vertex shader's side : more transistors for it.

VS/PS 3.0 : that musn't be free. And now, 16 real pipes whilst nV3x is only 4 ? And this with less than nV36's transistor budget ?

So I summurize : it would be two nV38 + VS/PS 3.0 support with less than the double budget in transistor count than nV38.....

I'm doubtfull, unless they created redundant paratial pipes like in S3's Deltachrome.