• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

NV40 Only 8 Pipes?

quote:

"We understand that the NV40 will actually only have eight physical pipelines, but these will appear to act like 16 in certain games and, indeed, in 3DMark 2001. "

Doom3 and 3dmark 2001
 
If this ends up being true, it looks like nvidia is about to Pee in their own face again. Talk about Voodoo Graphic Techniques. Alas, all this sounds rather dubious and The Inquirer changes their story about every day, so I'll just hold off on further speculation.

Let both Cards drop to the public and then...let the pissing contest begin!
 
Alas, all this sounds rather dubious and The Inquirer changes their story about every day, so I'll just hold off on further speculation.
Click to expand...

Whatchoo talking about? The inquirer is always right, even when it changes it's story repeatedly and then back again ;) :D
 
Isn't that like the nv3x? It only has 4 physical pipes, but in certain cases (z-buffer operations) it can act like 8?

It's not a BAD design, and it would make sense nVidia would improve on it. Increase it from 8-pseudo pipes to 16-pseudo pipes, and increase what operations those pseudo-pipes can do from basic z-operations to some...I dunno, primitive texturing?

The thing to remember is that nothing, inherently, about the nv3x design was *bad*. It was just too slow, and tried to do too much too early.

I mean, having support for FP32 sometimes is great - it means greater precision than is needed is available for truly remarkable effects. And when lower precision can be used - terrific, peformance boost!

Ditto with the pseudo-pipes. Your rendering pipes aren't always doing the same thing - being able to get 'more' pipes for certain operations is a neat trick.

The problem was that the technology for the core just wasn't mature enough, and it did not have enough pixel shader power to compete.

Those are both things that can be fixed by another generation of technology.

IE., everything points to the nv40 being an evolutionary step up from the nv3x rather than the revolutionary change from the nv2x to the nv3x. But, that's okay! The nv3x wasn't a bad design, it just needed to be faster and more efficient. The *concept* was sound.
 
Isn't that like the nv3x? It only has 4 physical pipes, but in certain cases (z-buffer operations) it can act like 8?
Click to expand...

Yeah, but in this case it's more versatile apparently. The NV30 doesn't have it's fillrate effectively doubled by utilizing the vertex shader as a pseudo pixel pipeline.

The problem as theinq states it is that this is going to be of limited usefullness since the Vertex shader can't apply any reasonable filtering to the output. Still probably more useful than Z-stencil ops for most common games I'd think - D3 should be the exception.

I personally think they're probably wrong again.
 
You must log in or register to reply here.
Back
Top