next gen kepler to support dx 11.1, also take a year to rollout all cards - Page 23

Ninja Prime · 02-03-12, 11:26 AM

Quote:

Originally Posted by Robs07M6S

Yeah none of it sounded right to me either. Hopefully we get something concrete soon enough though.

There is a little bit of concrete info:

Quote:

Originally Posted by RussianJ
In about 4 hours or so there is a local invite only event where the Green Team CEO will be giving a Q+A session as well as a demo of upcoming Tegra, CUDA and Tesla performance. The preview email gave mention to a smaller process being talked of as well.
Anyone have any ideas for questions they would like to get answered?

And what came of that was this:

Quote:

Some answers for people.
CUDA demo on unspecified live 28nm card vs GTX 580 was roughly 28% faster. No clue again what card. Also showed 580 vs Tesla cards. Depending on how he rates paycheck cycles, either mid march or end of april for the launch.
When asked on the new cards launch date I was told "your next 6 paychecks so save up now." Take it as you will, I'm thinking under 45 days out.

28% faster than GTX 580 in CUDA, probably translates to 20-25% in real world gaming. About where the 7970 launched, though since driver revisions its a bit higher than that now. Depending on how he rates paycheck cycles, either mid march or end of april for the launch.

shadow001 · 02-03-12, 01:55 PM

Quote:

Originally Posted by Vardant

It's just a rumor. Even if it's true, it's nothing going to be dedicated to just PhysX, that would be stupid, they ported PhysX to Cuda for that purpose. If anything, they will tweak the shaders.

It's fluid based physics according to that article, wich is a step beyond what regular PhysX does unless you enjoy seeing fluid based physics in games running in realtime displayed at 2 Fps...

Quote:

It's 6 games vs none from the other companies. That's still a win in my books.

In how many years?....I rest my case.

Quote:

I don't think that's true at all. The extensions were needed, because OpenGL was too slow to implement new features, but developers needed them right then.

Also, PhysX doesn't need to become a standard, it's just a middleware. That would be like asking developers to only use Unreal engine.

You'll find this funny but OpenGL actually had more feature support at the begining since it actually supported features that DirectX didn't at the time, such as multitexturing wich Direct X5 didn't and OpenGL did and it was only added to direct X6, and it's those extensions that brought it down mainly because GPU vendors couldn't agree on features they wanted to add on their next generation cards, with each one going their own direction and game developers holding the ball by having to choose what features to use in their new game releases....It was a messy situation to say the least.

Making all features mandatory in DX10 and allowing no extensions at all was the best thing that could have happened to gaming, so now that companies can't mess with graphics features anymore to be DX10/11 compliant or risk the ire of microsoft, you have Nvidia seemingly trying to do it with physics instead to make their hardware seem special over the competititon, like they didn't learn a damn thing with GPU PhysX and are trying it a second time.

Features supported by a single card brand will never fly as developers are interested in selling their games to as many people as possible, many of whom might not be running Nvidia hardware at all, so unifying standards is the only way to go to see them used in more games ansd sooner, but from a marketing standpoint it doesn't make your hardware seem special, and Nvidia hates that with a passion it seems.

Muppet · 02-03-12, 05:00 PM

Quote:

Originally Posted by Xion X2

Yeah, same here. Just 2GB of VRAM would be disappointing. Even if it's 40-50% faster than 7970, as they claim, I won't buy it because I like to turn up AA all the way in my games, and 2GB will be a bottleneck soon.

And I think the 40-50% faster claim is just Nvidia PR. Just like the ATI PR claiming 50-60% faster than GTX580. Probably some canned bench they like to run that favors the hardware for some reason.

I expect it to be faster but not that much faster. Probably around 15-20%.

Yes, I like to turn up the AA too. I play everything in Surround, except BF3. That i play at 120hz. For Surround 2Gb, just won't cut it i'm afraid.

**K007** · 02-03-12, 05:21 PM

ultra bf3 at high res already goes over 2gb

Muppet · 02-03-12, 05:25 PM

Quote:

Originally Posted by K007

ultra bf3 at high res already goes over 2gb

That's why I play it on the 120Hz screen and not Surround.

Johnny C · 02-04-12, 04:22 PM

Quote:

Originally Posted by K007

Entire Nvidia Kepler Series Specifications, Price & Release Date
http://lenzfire.com/2012/02/entire-n...se-date-43823/

http://semiaccurate.com/2012/02/01/p...lergk104-fast/

I'm thinking that the numbers are correct from what you've posted but read the semi accurate link for a possible explanation of where that 45% may come from.

**K007** · 02-04-12, 04:32 PM

numbers could be correct but i think that is more like extreme cases where nvidia has a advantage on a certain game. they might be enabling certain features like phyx to max or something then comparing it to ati version while still having a high phyx setting.

Johnny C · 02-04-12, 08:50 PM

Quote:

Originally Posted by K007

numbers could be correct but i think that is more like extreme cases where nvidia has a advantage on a certain game. they might be enabling certain features like phyx to max or something then comparing it to ati version while still having a high phyx setting.

Like in Crysis 2 where they have an excessive level of tessellation for no apparent reason.

**Roadhog** · 02-04-12, 09:03 PM

Quote:

Originally Posted by Johnny C

Like in Crysis 2 where they have an excessive level of tessellation for no apparent reason.

I think you know the reason why.

Xion X2 · 02-04-12, 09:42 PM

Quote:

Originally Posted by Johnny C

http://semiaccurate.com/2012/02/01/p...lergk104-fast/

I'm thinking that the numbers are correct from what you've posted but read the semi accurate link for a possible explanation of where that 45% may come from.

So this "50% faster than 7970" claim is related to the supposed PhysX hardware that's on the card?

How many games even came out with PhysX last year? 6-7? Meh.

Muppet · 02-04-12, 09:48 PM

Quote:

Originally Posted by Xion X2

So this "50% faster than 7970" claim is related to the supposed PhysX hardware that's on the card?

How many games even came out with PhysX last year? 6-7? Meh.

I would take anything that Semiaccurate says to seriously. He has a beef with anything Nvidia. The same goes for Fud and anything they say about ATI. Both sites are really a joke.

Xion X2 · 02-04-12, 09:57 PM

Quote:

Originally Posted by Muppet

I would take anything that Semiaccurate says to seriously. He has a beef with anything Nvidia. The same goes for Fud and anything they say about ATI. Both sites are really a joke.

Yeah, I know. It didn't occur to me until after posting that it was semi-accurate where I had read that. lol.

02-03-12, 05:21 PM	#268
K007 Join Date: Sep 2004 Location: Australia, Sydney Posts: 9,405	Re: next gen kepler to support dx 11.1, also take a year to rollout all cards ultra bf3 at high res already goes over 2gb

02-04-12, 04:32 PM	#271
K007 Join Date: Sep 2004 Location: Australia, Sydney Posts: 9,405	Re: next gen kepler to support dx 11.1, also take a year to rollout all cards numbers could be correct but i think that is more like extreme cases where nvidia has a advantage on a certain game. they might be enabling certain features like phyx to max or something then comparing it to ati version while still having a high phyx setting.