The G92 Architecture Rumours & Speculation Thread

[fellix]

How you dare to be non-believer!

 

[CJ]

Sorry but I was so happy that RV670 can make almost same performance as ultra. I mean I never expected anything like this to come out.

And ofcourse you always play the popular game 3DM06. The HD2900XT was also a good performer in 3DM06, but I don't need to remind you how it did in most games compared to the GF8800GTX (especially with AA).

 

[seahawk]

But according to ATI/AMD AA perfromance is much better. It does look good for ATI this time.

 

[vertex_shader]

But according to ATI/AMD AA perfromance is much better. It does look good for ATI this time.

Wait for the reviews, without good driver and DEV support things can go backwards in realworld.

 

[trinibwoy]

But according to ATI/AMD AA perfromance is much better. It does look good for ATI this time.

I haven't seen any statements from ATI/AMD stating as much. Where did you see that?

 

[AnarchX]

AIB?-slides @ CHW:
http://www.chilehardware.com/foro/nvidia-vs-amd-t117534.html

 

[Arun]

So wait... Everything points at RV670 being as fast or faster than the HD 2900 XTX, and yet this implies NVIDIA sees the 8800 GT as being positioned much lower than R600?

No offense, but that's misinformation if I've ever seen any... Or just plain stupidity, perhaps. My guess is NVIDIA's definition of 'positioning' is exclusively price-related, not performance-related. Otherwise all this makes zero sense.

EDIT: I can't help but suspect a <=$299 pricepoint for the 'new' GTS...

 

[fellix]

Yup -- 8800GT seems to be mistargeted.
Aim for lower, Nvidia!

 

[Arun]

Mistargeted, as in NVIDIA is being stupid, or as in this slide is stupid, or as in NVIDIA is confusing everyone on purpose? You gotta admit you're being rather vague!

Anyway, I stick by my guess that the 8800 GT will, performance-wise, be positioned between the old 8800 GTS and the new one.

 

[fellix]

Anyway, I stick by my guess that the 8800 GT will, performance-wise, be positioned between the old 8800 GTS and the new one.

That's pretty narrow margin -- make sure you don't shake the mouse too much when benching, to avoid measurement error confusion.

 

[Topman]

Fruitzilla strikes again

http://www.fudzilla.com/index.php?option=com_content&task=view&id=3479&Itemid=1

No DX 10.1 for GF 8800GT?

:S

bye

 

[nicolasb]

Fruitzilla strikes again

No DX 10.1 for GF 8800GT?

To be fair to Fudo, even he clearly isn't entirely convinced by this and reckons further investigation is required.

 

[INKster]

Exposed: http://forums.vr-zone.com/showthread.php?t=192488

 

[AnarchX]

Exposed: http://forums.vr-zone.com/showthread.php?t=192488

The funny thing is, that cooler or just cooler-hole is bigger than on the 8800GT @ Mobile1.
But I bet it is an older version, like Sample-8800GTX with watercooling-plugs.

 

[fellix]

The fan is definitely larger now.
But it's a kind of weird, they still use heat spreader & this pointless metal frame.

 

[INKster]

The funny thing is, that cooler or just cooler-hole is bigger than on the 8800GT @ Mobile1...

The vents near the TV-Out jack also seem bigger.

 

[AnarchX]

The fan is definitely larger now.

Now or it was larger (see my edit)?

Maybe the overheating-rumors are outdated.

 

[Sound_Card]

That's pretty narrow margin -- make sure you don't shake the mouse too much when benching, to avoid measurement error confusion.

No, that should be a extreamly narrow margin.

In which case I say it will be slower than the old GTS.

 

[turtle]

I would think logic (where it was intended to be positioned) would say it's slower than the 'old' 8800gts by default, and with new clocks (because of fear of RV670 at same price range) or overclocking (a fortunate or unfortunate side-effect depending on how you look at it) can close to meet or slightly exceed the older GTS, hence the need for a faster GTS to keep from obsoleting the more expensive product.

 

[trinibwoy]

Well then I'd ask - why not just release a higher clocked 8850GTS-512MB based on G92 to replace the 640MB GTS?