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
ouch, so they cheat by running DX9? nasteh
Posted on Reply
#2
phanbuey
god will they just release this thing or what... i wanna see it beat a 285 by 35% :rockout:

I hate NDA's.
Posted on Reply
#3
TheMailMan78
Big Member
Benetanegia said:
Quad pumped yes, but 1 op/cycle anyway, because it needs 4 clocks to finish the work. It's the same for Nvidia (1 op/cycle) ones AFAIK and even if it wasn't the same it still has 16. If not quad pumped it would still have 4 times the tesselation capabilities.

Tesselators themselves are not the problem anyway. Triangle setup is far more important. You could have a tesselator hexa or hecta pumped (or 100 tesselators), I don't care, but it wouldn't make a difference if you can only do one tri/cycle as is the case with Evergreen. Nvidia didn't put 16 tesselators because they are needed anyway, they are there to increase availability and reduce latency by giving each SM one tesselator to operate with. Tesselation performance will still be limited primarily by the tri setup and secondly by hull/domain shaders anyway (both of which Fermi seems to deal better with thanks to its L1/L2 caches), but HD5870/HD5770 demostrate that Evergreen is heavily limited by tri setup/tesselator even with the low tesselation levels found in Stalker, Dirt2, etc and not by shader performance: when tesselation is enabled the HD5770 loses much less than it's biggest brethren.


DX9 FTW!
Posted on Reply
#5
Benetanegia
TheMailMan78 said:
http://www.semiaccurate.com/static/uploads/2010/03_march/Dirt2_scores.jpg

DX9 FTW!
So you pretend to say the charts posted are in DX9? Seriously?:laugh:
To pretend that the card is slow is one thing, to pretend that it performs somewhere in between a GTS250 and a GTX260... :laugh:

Anyway. Wow! We've really fallen under, Nvidia bashers, didn't we? Less than a quarter of a screenshot, with no single proof of which card is being used, coming from SA and Charlie Demerjian, both of which have not access to Nvidia presentations because they are never invited, but we are happy now with just that. we don't any more proof. A semi-properly presented charts are not believable, but this? Oh sure. Come on guys... it's PATHETIC.

The only truth is that CD had 2 days (1 now) to make noise and get attention for his new home SA, because once Fermi is launched he will not have anything relevant to talk about and he is deemed to oblivion, and you all fell in the trap... Poor boys...
Posted on Reply
#6
HalfAHertz
Benetanegia said:
So you pretend to say the charts posted are in DX9? Seriously?:laugh:
To pretend that the card is slow is one thing, to pretend that it performs somewhere in between a GTS250 and a GTX260... :laugh:

Anyway. Wow! We've really fallen under, Nvidia bashers, didn't we? Less than a quarter of a screenshot, with no single proof of which card is being used, coming from SA and Charlie Demerjian, both of which have not access to Nvidia presentations because they are never invited, but we are happy now with just that. we don't any more proof. A semi-properly presented charts are not believable, but this? Oh sure. Come on guys... it's PATHETIC.

The only truth is that CD had 2 days (1 now) to make noise and get attention for his new home SA, because once Fermi is launched he will not have anything relevant to talk about and he is deemed to oblivion, and you all fell in the trap... Poor boys...
God you sure are pregidous sometimes...
Posted on Reply
#7
erocker
Senior Moderator
I wold love a rule stating no links to Fudzilla or SemiAccurate. Seriously, why do people post them here? We all have the internet and have the capability to input their web addresses into our browsers and read their content on their site. It's so close to spam it feels like I'm eating a pork sandwich. Oh I miss the times when smart conversation was based on fact and not rumor and heresay.. if that time ever existed, I'm probably just crazy or something. Release is tomorrow, er.. well late tonight. You know you're excited.
Posted on Reply
#8
DaedalusHelios
HalfAHertz said:
God you sure are pregidous sometimes...
Says the guy with the signature built on propaganda. :roll:

Just look in the mirror man. :rolleyes:

I buy high end regardless of brand. I don't trash either company. I just share experiences.
Posted on Reply
#9
TheMailMan78
Big Member
Benetanegia said:
So you pretend to say the charts posted are in DX9? Seriously?:laugh:
To pretend that the card is slow is one thing, to pretend that it performs somewhere in between a GTS250 and a GTX260... :laugh:

Anyway. Wow! We've really fallen under, Nvidia bashers, didn't we? Less than a quarter of a screenshot, with no single proof of which card is being used, coming from SA and Charlie Demerjian, both of which have not access to Nvidia presentations because they are never invited, but we are happy now with just that. we don't any more proof. A semi-properly presented charts are not believable, but this? Oh sure. Come on guys... it's PATHETIC.

The only truth is that CD had 2 days (1 now) to make noise and get attention for his new home SA, because once Fermi is launched he will not have anything relevant to talk about and he is deemed to oblivion, and you all fell in the trap... Poor boys...
No I don't pretend. I just knew that would piss you off.

Anyway ATI not lowering its prices is all the proof I need that Fermi is a failure.
Posted on Reply
#10
HalfAHertz
DaedalusHelios said:
Says the guy with the signature built on propaganda. :roll:

Just look in the mirror man. :rolleyes:

I buy high end regardless of brand. I don't trash either company. I just share experiences.
Why are you right:

Yes I am buuuut I have the complete right to be so. I am simply tired of huge corporations(please note the plural form here) doing whatever they want completely ignoring the interest of their customers in the process.

Please do ignore all the bullshit and rumors on sites like Semiaccurate and Fudzilla and just look at the bare facts. We're discussing a product that should have been here about six months ago. It was delayed again and again and again.

Why are you wrong:

Why should you care and why is this hurtful to you? Well you were promised something and that promise was broken over and over again. The graphics card ecosystem has been broken, there is no healthy competition and the prices instead of falling, have continued to rise.

Everybody has the right of freedom of speech and having an opinion. You cannot simply label or insult someone only because his view on the matter differs from your own, unless of course you can deliver solid facts.

Do you have any proof right now that my belief is skewed? Can you honestly tell me that I am wrong, when all the facts are here and are staring you in the face: there is no competition on the market, the prices of video cards has sky-rocketed, the consumer simply does not get the same performance/$ as he did at the same time last year.

Could this have been avoided? Hell yes! In the face of obvious manufacturing difficulties Nvidia, and to some extent Ati as well, have failed to maintain sufficient stock of their older generation and instead of lowering prices, to at least try to create some competition in the mid-range, prices have continued to grow uncontrollably.If Nvidia was more open and honest they could have prepared us for the current situation.

Are Ati any better? No, because they have failed to keep their products at the originally announced price.

I don't care anymore if they will finally deliver on their promise tomorrow or two weeks after that. The damage to the market has already been done and they deserve everything that's coming to them.

Here's what I think will happen tomorrow: There won't be reviews from everyone, just from people that have been cherry picked, after signing some dubious agreement that they will show only controlled results that have been verified by Nvidia.

These are my own beliefs, you can continue to argue with me, prove me that I am wrong with solid facts or you can simply ignore me, but you should not label, demean or undermine me, or the fore-mentioned sites simply because you disagree.

/rage off

P.S. My signature was intended as a joke, if anyone asks me nicely, I'll gladly remove it.
Posted on Reply
#11
Benetanegia
TheMailMan78 said:
No I don't pretend. I just knew that would piss you off.
So you are a troll. I've been telling you for a long time, you were becoming one.
Anyway ATI not lowering its prices is all the proof I need that Fermi is a failure.
Considering your previous post, taking that as a proof is definately a step in the right direction :laugh:, but still nothing sort of a proof for failure. Nvidia said and did the same regarding HD5xxx cards (regarding price cuts).

HalfAHertz said:
These are my own beliefs, you can continue to argue with me, prove me that I am wrong with solid facts or you can simply ignore me, but you should not label, demean or undermine me, or the fore-mentioned sites simply because you disagree.
So let's see if I understand. You can say whatever you want without any proofs, but if someone argues with you he needs to provide solid facts.

IMO there are places for doing that and they are not called forums, they're called blogs. You can create one for free and say whatever you want there. In forums, you can't say whatever you want without proofs or someone will argue with you by either presenting you with the facts or exposing your lack of evidence. You are entitled to your beliefs, but others are too, even when those beliefs are about you.
Posted on Reply
#12
HalfAHertz
Benetanegia said:

So let's see if I understand. You can say whatever you want without any proofs, but if someone argues with you he needs to provide solid facts.

IMO there are places for doing that and they are not called forums, they're called blogs. You can create one for free and say whatever you want there. In forums, you can't say whatever you want without proofs or someone will argue with you by either presenting you with the facts or exposing your lack of evidence. You are entitled to your beliefs, but others are too, even when those beliefs are about you.
First of all the definition of forum: http://www.thefreedictionary.com/forum please note points 1b and 1c

Secondly you completely misunderstood me. I stated my proof - the obvious fact that the card has been delayed beyond any even remotely acceptable time frame. The only logical reason for that is that there were some problems with it - be it software, hardware,manufacturing or all of the above, it doesn't really matter, all that matters is that it is still 6 months late. And that is the undeniable truth, Charlie may be right or wrong and we may choose to believe him or not but that doesn't make us "Nvidia bashers" or "PATHETIC".

What really matters here is that the graphics card market has been foiled with and the consumer has been hurt. If you ask me, we should all stop buying gfx cards until the prices drop to pre-Dx11 levels and we have the same performance/$ we had back then...
Stick it ot the Man, man :P
Posted on Reply
#13
[I.R.A]_FBi
why cant we live together and show some man love?
Posted on Reply
#14
Benetanegia
HalfAHertz said:
First of all the definition of forum: http://www.thefreedictionary.com/forum please note points 1b and 1c
Yeah, discussion, so if you post your opinion and someone doesn't like it because it has no real proofs behind them, he will disscuss your opinion, which is mostly baseless. Nowhere you were talking about the delay.
Secondly you completely misunderstood me. I stated my proof - the obvious fact that the card has been delayed beyond any even remotely acceptable time frame. The only logical reason for that is that there were some problems with it - be it software, hardware,manufacturing or all of the above, it doesn't really matter, all that matters is that it is still 6 months late. And that is the undeniable truth, Charlie may be right or wrong and we may choose to believe him or not but that doesn't make us "Nvidia bashers" or "PATHETIC".
First of all, you had not posted about that ss so I was definately not talking about you. Besides I am not arguing against the critics, but on which ""proofs"" those critics are based off. Where's the GPU-z shot to begin with? Heck, where's the complete screenshot? Where's the proof that it comes from Nvidia and not from anyone else (or CD himself)? Like I said I know it didn't come from a Nvidia presentation, because CD has not been and will never be in a Nvidia presentation.

None of you have even considered those things, while those things have been asked for, in the case of the benchmarks. And not to say they could be (or could be not) false, which was my point, no, because of the lack of that info the graphs were directly labeled as false. That is what it is pathetic and yeah it makes those who act in that way bashers, because they are more that willing to bash without a single proof. The same people can't be claiming some graphs to be false because of x reasons and then claim Nvidia is cheating based on something that breaks those x reasons and then some.
Posted on Reply
#15
HalfAHertz
Ok then, those are valid points. The SS are most probably fake, but consider that we live in a digital age and the internet connects people... CD may not have been on any of Nvidia's presentation, but that does not mean someone else could not have sent him valid shots. Anyway they are irrelevant now because hopefully later today we'll see some real results.

Ignoring the screenshot, my other points still count valid, Nvidia is late, that hurts the consumer ergo Nvidia brought all that negatity on themselves because of the way they conduct their business...
Posted on Reply
#16
TheMailMan78
Big Member
Benetanegia said:
Yeah, discussion, so if you post your opinion and someone doesn't like it because it has no real proofs behind them, he will disscuss your opinion, which is mostly baseless. Nowhere you were talking about the delay.



First of all, you had not posted about that ss so I was definately not talking about you. Besides I am not arguing against the critics, but on which ""proofs"" those critics are based off. Where's the GPU-z shot to begin with? Heck, where's the complete screenshot? Where's the proof that it comes from Nvidia and not from anyone else (or CD himself)? Like I said I know it didn't come from a Nvidia presentation, because CD has not been and will never be in a Nvidia presentation.

None of you have even considered those things, while those things have been asked for, in the case of the benchmarks. And not to say they could be (or could be not) false, which was my point, no, because of the lack of that info the graphs were directly labeled as false. That is what it is pathetic and yeah it makes those who act in that way bashers, because they are more that willing to bash without a single proof. The same people can't be claiming some graphs to be false because of x reasons and then claim Nvidia is cheating based on something that breaks those x reasons and then some.
The reason people do that is because you were basing your shallow theories off of "positive" info about fermi but couldn't take it when someone shook those theories to the ground with fact. The sad part is your a fanboy in denial which is why you get trolled so much. People play you like a puppet on this forum but your ego is so large you don't even realize it.

I have no idea if Fermi will be good or not but rest assured Nvidia has fallen from grace lately from a PR point of view. That alone is bad for everyone. I honestly believe they rode that last architecture for FAR to long. I guess time will tell.
Posted on Reply
#17
Benetanegia
TheMailMan78 said:
The reason people do that is because you were basing your shallow theories off of "positive" info about fermi but couldn't take it when someone shook those theories to the ground with fact. The sad part is your a fanboy in denial which is why you get trolled so much. People play you like a puppet on this forum but your ego is so large you don't even realize it.

I have no idea if Fermi will be good or not but rest assured Nvidia has fallen from grace lately from a PR point of view. That alone is bad for everyone. I honestly believe they rode that last architecture for FAR to long. I guess time will tell.
:laugh: Nice theory. If only a single proof (fact) had been posted...
Posted on Reply
#18
phanbuey
this shall be settled in 8 hours and one shall emerge the winner :rockout:

im betting it will be shlow... and that they will argue immature drivers. Lets see how it OC's... Despite being a huge hog it may OC well - especially the 470.
Posted on Reply
#19
afw
phanbuey said:
this shall be settled in 8 hours and one shall emerge the winner :rockout:

im betting it will be shlow... and that they will argue immature drivers. Lets see how it OC's... Despite being a huge hog it may OC well - especially the 470.
And we'll also have to see what sort of drivers they'll (reviewers) be using on ATi cards ... the 10.3 have shown some good gains over the 10.2 so ... we must watch out for that ... some of the pro nVidia sites might use old drivers to widen the margin .... :laugh: ...

I'm eagerly waiting for W1zz's review .... :rockout: ... since its gives more accurate performance summary charts (due to testing on a wide array of games - 18 infact ) :respect: ... I hope he uses the 10.3 driver set ;)
Posted on Reply
#20
jmcslob
Oh i can't wait for the Actual Fanboi Reviews Followed by Real world reviews So I can finally see if this card is worth something
Posted on Reply
Add your own comment