Friday, March 19th 2010

NVIDIA Claims Upper Hand in Tessellation Performance

A set of company slides leaked to the press reveals that NVIDIA is claiming the upper hand in tessellation performance. With this achievement, NVIDIA is looking to encourage leaps in geometric detail, probably in future games that make use of tessellation. NVIDIA's confidence comes from the way its GF100 GPU is designed (further explained here). Each GF100 GPU physically has 16 Polymorph Engines, one per streaming multiprocessor (SM) which helps in distributed, parallel geometry processing. Each Polymorph Engine has its own tessellation unit. With 15 SMs enabled on the GeForce GTX 480 and 14 on the GeForce GTX 470, there are that many independent tessellation units.

NVIDIA demonstrated its claims in the presentation using the Unigine Heaven, where the GeForce GTX 480 was pitted against a Radeon HD 5870. In many scenes where tessellation is lower, the GPUs performed neck-and-neck, with the GTX 480 performing better more often. But in scenes with heavy tessellation (particularly the "dragon" scene, where a highly detailed model of a dragon needs to be rendered with densely tessellated meshes, the GTX 480 clocks nearly a 100% performance increment over the HD 5870. NVIDIA has been confident about the tessellation performance back since January, when it detailed the GF100 architecture. The GeForce GTX 400 series graphics cards will be unveiled on the 26th of March.


Images Courtesy: Techno-Labs
Add your own comment

145 Comments on NVIDIA Claims Upper Hand in Tessellation Performance

#1
Mussels
Moderprator
TAViX said:
Where did you find this garbage??? No cards released, yet you have the benchies???? WTF?! Another BS!:banghead:
its quite possible for it to not be fake. We know for a fact reviewers have the cards under NDA - someone could have leaked the images.
Posted on Reply
#2
Valdez
TAViX said:
Where did you find this garbage??? No cards released, yet you have the benchies???? WTF?! Another BS!:banghead:
I don't know its fake or not, i just found it on another forum.
Posted on Reply
#3
TheMailMan78
Big Member
I love the title of this thread. "NVIDIA Claims Upper Hand in Tessellation Performance!" Like Nvidia would ever say "NVIDIA Claims LOWER Hand in Tessellation Performance!" :laugh:

I mean F#$K it! Heres the new title.

TheMailMan78 claims new Claims Upper Hand in Tessellation Performance! Heres charts to prove it!
Posted on Reply
#4
Mussels
Moderprator
TheMailMan78 said:
I love the title of this thread. "NVIDIA Claims Upper Hand in Tessellation Performance!" Like Nvidia would ever say "NVIDIA Claims LOWER Hand in Tessellation Performance!" :laugh:
"nvidia claims upper hand in something no one cares about, cause they suck in everything else"

is how i read it.

You know they're reaching when marketing cant even make it sound good.
Posted on Reply
#5
TheMailMan78
Big Member
Mussels said:
"nvidia claims upper hand in something no one cares about, cause they suck in everything else"

is how i read it.

You know they're reaching when marketing cant even make it sound good.
Yeah but they do not have a cool chart like me. TheMailMan78 tessellation support beats AMD and Nivdia.
Posted on Reply
#7
TheMailMan78
Big Member
Damn with a graph like mine I figured everyone would want a MailMan!

DeathByTray said:
Let's say cat is a synonym for real and dog is a synonym for fake.

So, W1zzard, is this a cat or a dog?
http://i39.tinypic.com/14xp3qh.png



:wtf::p
Its a Hermaphrodite!
Posted on Reply
#8
dir_d
I wonder if they were using 10.3a in those Dirt 2 benches...I gain 15FPS over the 10.2 by switching to 10.3a. I was seriously amazed
Posted on Reply
#9
the54thvoid
http://www.techpowerup.com/reviews/MSI/R5870_HD_5870_Lightning/15.html

W1zz's own 5870 review puts the basic 5870 at 76 fps at that resolution on Far Cry 2 and thats with 9.12 drivers. So it's fair to say theres a lot of 'uncertainty' about the validity of the set up if not the authenticity of the graphs.

Also, the Dirt 2 graph states 4AA, the Far Cry 2 graphs states nothing about settings. Therefore, it's a crap post regardless. In science we need to make comparisons based on facts and numbers, not unsourced, unspecified benchs.

And for the hell of it, if you added 'dir d's' extra 15fps from the 10.3a drivers that makes 91 fps. Yes, i know thats not a valid statement but given how unspecific those graphs are, why not make it. And that makes it on a par with the GTX 480.

Game set and match to both my logic and my use of illogic.
Posted on Reply
#10
erocker
TAViX said:
Where did you find this garbage??? No cards released, yet you have the benchies???? WTF?! Another BS!:banghead:
I wouldn't say BS, but anyone can make a graph at anytime using believable or unbelievable facts and figures. Five Four and a half more days and the truth will be revealed. I don't think anyone will be surprised when it comes to the results. Anything until the 26th is speculation and or fake.
Posted on Reply
#11
the54thvoid
http://www.fudzilla.com/content/view/17754/40/

And this, with catalyst 10.2 gives Dirt 2 @ 1920x1200 with 4xMSAA at full settings on DX11 gives 59.3 fps which is 13 fps HIGHER than the GTX 480 gets from the unspecified charts. This is most clearly not a genuine graph and if it is, it severely compromises the performance of the 5870.
(It also gives 47 fps at 2560x1600 = beats GTX 480 still.)
Wins for me again.

(and I used Fuad's site - most def not an ATI site.)
Posted on Reply
#12
Valdez
erocker said:
Anything until the 26th is speculation and or fake.
Exactly.

Don't get angry at a graph which nobody claims as authentic.
Posted on Reply
#13
Benetanegia
the54thvoid said:
http://www.fudzilla.com/content/view/17754/40/

And this, with catalyst 10.2 gives Dirt 2 @ 1920x1200 with 4xMSAA at full settings on DX11 gives 59.3 fps which is 13 fps HIGHER than the GTX 480 gets from the unspecified charts. This is most clearly not a genuine graph and if it is, it severely compromises the performance of the 5870.
(It also gives 47 fps at 2560x1600 = beats GTX 480 still.)
Wins for me again.

(and I used Fuad's site - most def not an ATI site.)
Apart from what Erocker and Valdez said, you are aware that benching in a different spot of a map/level/track will always net very different results right? You can never compare benchmarks from different sites.
Posted on Reply
#14
the54thvoid
Uh huh. I know.

My point kinda points to the fact that a snippet of a bench or unsupported graphs are meaningless.

I quote myself

"Game set and match to both my logic and my use of illogic."

use if illogic - see, I know my comparisons aren't valid, i'm just saying look what unsupported info gives you. And yes, I'll wait for W1zz's review. Until then we only know it's late and hot.
Posted on Reply
#15
Benetanegia
I understand your point, but this is no science. This is a thread for especulation and everybody here undestands that.

There's no reason to believe those graphs are legit, but there's no reason to believe those graphs are fake either. Everybody here understands that too. There's no reason to put all the effort you put into proving those graphs could be fake, because no one has claimed them to be true, as Valdez said.

I'm not trying to be critical or offend you, but IMO the only problem with those graphs is your innability to deal with especulation. I'm not trying to be harsh, I'm just saying there's no need to prove or disprove speculation.
Posted on Reply
#16
DeathByTray
TheMailMan78 said:

Its a Hermaphrodite!
:nutkick:
Posted on Reply
#17
TheMailMan78
Big Member
Benetanegia said:
There's no reason to believe those graphs are legit, but there's no reason to believe those graphs are fake either.
Of course there is reason to think those graphs are fake. They are not from a TPU review. Anything less would be uncivilized...

Posted on Reply
#18
SetsunaFZero
Valdez said:
http://i39.tinypic.com/14xp3qh.png
even if this bench's are real, i would go ATI this time. Electricity isn't cheap in austria
nvidia fail this time in performance and power efficiency :laugh: ATI's HD5k is win IMHO
Posted on Reply
#19
the54thvoid
Benetanegia said:
I understand your point, but this is no science. This is a thread for especulation and everybody here undestands that.

There's no reason to believe those graphs are legit, but there's no reason to believe those graphs are fake either. Everybody here understands that too. There's no reason to put all the effort you put into proving those graphs could be fake, because no one has claimed them to be true, as Valdez said.

I'm not trying to be critical or offend you, but IMO the only problem with those graphs is your innability to deal with especulation. I'm not trying to be harsh, I'm just saying there's no need to prove or disprove speculation.
Dude, chill.

I am neither offended or feel chastised. All the GTX 4xx info thus far is necessarily speculation ( i refuse to use the term 'e'speculation). However, i thought unsourced info showing one side to be superior was akin to trolling.
I've seen folk do it here before.
And for the cheap seats - it took no effort to google two reviews. If i wanted to put effort into something, it wouldn't be on this. I'd rather put more effort into deciding which album i prefer - Ride the Lightning or Master of Puppets.

So lets hug and wait for the 26th? :toast:
Posted on Reply
#20
jmcslob
I hope those charts are completely true...

THINK ABOUT PRICES PEOPLE..

Plus that just means quicker ATI 6000 series which means even LOWER PRICES

and when I say low I mean i want my 5830 for $159 or LESS where it should be!!
Posted on Reply
#21
DeathByTray
I've heard Nvidia's driver aren't even beta worthy, can you comment on that W1zzard? CAT OR DOG? :laugh:
Posted on Reply
#22
Benetanegia
the54thvoid said:
Dude, chill.

I am neither offended or feel chastised. All the GTX 4xx info thus far is necessarily speculation ( i refuse to use the term 'e'speculation). However, i thought unsourced info showing one side to be superior was akin to trolling.
I've seen folk do it here before.
And for the cheap seats - it took no effort to google two reviews. If i wanted to put effort into something, it wouldn't be on this. I'd rather put more effort into deciding which album i prefer - Ride the Lightning or Master of Puppets.

So lets hug and wait for the 26th? :toast:
No offense, but you are again wrongly judging those charts (and the action of posting them) and in the process, giving them the importance you were trying to avoid with your posts. Trolling? He just posted some graphs, which may or may not be legit, but don't kill the messenger. Besides, the charts can easily be legit, or at least the performance lead they are showingcon be true. It's not far fetched, not at all. First we have FarCry 2, a game that runs better on Nvidia hardware and second we have Dirt2 on DX11, which means Tesselation, where we know Nvidia is much much faster. Dirt2 has little tesselation, maybe not even worth mentioning performance wise, but everything adds up.

The only thing I agree with is the last sentence, let's wait for the 26th, because only then we will know if the rumors and speculation are true or not. We know nothing or very little about these cards, and everything we know falls on the same bag: speculation. Showing some charts where one side is faster is hardly trolling, especially when there's no additional comment and... it might even be true. The posts akin to yours on the other hand are far closer to being trolling. They don't really help at all and the forums are full of them (which makes them redundant). The posts like the one that Valdez posted are the helpful ones, their content might be legit or not, but like I said, it's just naked info and people in these forums are intelligent enough to distinguish between (probably/maybe) fake and (probably/maybe) legit information.

No one needs to be reminded that rumors/speculation and pre-release graphs can be fake after every turn. We give no more importance to those charts than the one deserved. Like I said, the only one who gave them real importance in this thread was you so far and I only pointed out that you were using bad measures to treat/refute the info, because comparing benchmarks from different sources proves nothing.

Now it's clear you did got offended, so sorry for that, not my intention and neither is with this post. Let's wait a week and we will know which info was legit and which one was not.

PS. I love Metallica, you'd be better searching anything about them than trying to figure out just another way to demostrate how Nvidia has failed, while we don't even know if they trully failed yet.

PS2. Thanks very much for correcting my spelling. I really appreciate when people correct my english, because let's be honest, what kind of person is unable to write english properly? It must be some kind of illiterate, or worst :eek:, some kind of heretic for whom english is just his 4th language in order of importance or something wild like that.:rolleyes: Not mad at you BTW, only at people whom mother language is english, in general. I can patter more or less 7 languages and only with english you find people who correct you after every turn. Sadly you fell under that demographic with that comment and we do feel offended by that kind of comments. :toast:
Posted on Reply
#23
the54thvoid
PS2. Thanks very much for correcting my spelling. I really appreciate when people correct my english, because let's be honest, what kind of person is unable to write english properly? It must be some kind of illiterate, or worst :eek:, some kind of heretic for whom english is just his 4th language in order of importance or something wild like that.:rolleyes: Not mad at you BTW, only at people whom mother language is english, in general. I can patter more or less 7 languages and only with english you find people who correct you after every turn. Sadly you fell under that demographic with that comment and we do feel offended by that kind of comments. :toast:
Ben. Ben. Ben. Lets stop this now. I wasn't correcting your english! You used the term especulation. I thought the 'e' meant 'electronic' as in e-mail or e-tailer. I wasn't correcting you, lol. I thought it was one of those modern terms used for online stuff.

I'm not a spelling nazi - I only use one langauge (harsh language) so all credit to the multi linguists.

I think somewhere we've got wires crossed and perhaps the tone of my post has been lost long ago. I'm not going to even bother explaining anymore, i figure no matter what i say you'll have an equally long, 'yes but...' post. (When fact is we're probably singing the same words with a different tune).

So, to summarise - No offence to you - I honestly was not correcting your spelling - i thought you were using a tech term by prefixing the letter 'e' in front of another word. Like this e-discussion.

As for Metallica - I know it's Master of Puppets. :rockout:

And remember one thing - my use of english is minimal - I'm Scottish - Scandinavians sound better at englsh than us!
Posted on Reply
#24
nt300
Cant wait for CCC 10.4, something new is going to be added that will make HD 5000's special but cant tell right now :D
Posted on Reply
#25
the54thvoid
Something that will speed up loading times on BC2? DX11 and ATI+ BC2 = Slow load. I always miss the vehicles at map starts...
Posted on Reply
Add your own comment