NVIDIA GF100 & Friends speculation

Whoa, you're reading as a pre-launch event? Man your getting gullible! Its actually an announcement of an announcement of an revealing preview of a pre-launch event. They can't just go straight from endless reveal to pre-launch to launch. That won't take nearly as much time or focus from the journo community. Let them ease into it.

 

No, I'm talking about bechmarking NV's in-house (and unobtainable) demos on AMD hardwware as part of the review.

 

Heh that is pretty laughable. I wonder if anybody will...

 

They can't because they made sure to build PhysX into it (referring to the sled demo). Thus is could never be compared with a hacked wrapper or anything.

 

if this is true and the 470 is placed between the 5850 and 5870 in performance and price is at the5850 price. Then we will see price drops form ati.

And I will finally be happy !

 

AMD has its own demos - there's a decent chance they'll run on GF100, as I think they're regular D3D11:

​

​

Frame rates included. I've got no idea whether they're CrossFire friendly, i.e. whether his HD5970 is faster or slower than HD5870.

Hmm, one of those people out there with a GTX470 or GTX480 should try these, hint hint...​

Jawed​

 

Yep, I'm getting ~715fps on a GTX 285 so I'm not sure how the GTX 480 manages to do worse.

 

Myth Busted! Well done Jamie.

 

The plaster cast version of both GTX480 and GTX 470 will be available online,soooooooon.

 

Didn't stop Nvidia from making a slide with GF100 results vs 5870 in the hair demo that is said to be built around PhysX.

 

I've found those high-fps test to vary wildly with plattform. Seems like a lynnfield with P55 suffers here compared to AMD, Bloomfield and even Core 2 systems.

--
edit:
WRT to Metro 2033: There seem to be wild differences depending on how much advacend physics is going on.
http://translate.google.de/translat...GPU-PhysX/Action-Spiel/Test/Test/&sl=de&tl=en
(sorry, only google translation at the moment)
HD5k really… well, see for yourself.

 

They probably made a CPU code-path for the PhysX part, that is single-thread x87, most likely, just like in games.

 

Wow thats some of the biggest leads over the 285 I've seen from Cypress. And the 5770 overtaking the 4890 too, I knew my card would come good.

nVidia fanboys will be committing suicide over these results.

 

I haven't seen any Radeon demos willing to run on anything else really. Some of them do not even run on the next generation Radeon! These DX11 demos could be a different case though!

 

So they get criticized for using that "new media", but when they stop, you criticize them for that ?

Damned if you do and damned if you don't indeed...

 

It will be interesting to see the GF100 cards on this game as well.

As I read in the article, there is no heavy tesselation used, so maybe any tesselation advantage the GF100 may have over the Cypress, should not manifest here.

 

Do or do not, there is no try...

 

Perhaps it's not the fact they used the said "new media" but HOW they used it?

 

But the performance lost is very high: The min are 75% and the avg fps are 56% higher without Tessellation.

 

That's a quite low cost considering Heaven's tessellation can sink framerate by around 90% at low res and still 70% at 1920 No AA/No AF (53/39% avg)