AMD V Nvidia by Richard Huddy

[Lonbjerg]

So you know for sure that Huddy is responsible for none of the positive decisions made at AMD. Of course you do.

I have no idea why you hate Huddy and AMD so much, very strange. For every one thing you might post that is positive about AMD, you post 1000 negatives and bash them from every possible angle. :thumbsdown:

Because he lies.
He claimed NVIDIA paid devs to use PhysX...but offered no documentation.
He claimed NVIDIA hobbled PhysX with X87 code...but leaving out the history of the PhysX API.
Now he claims NVIDIA forces Ubisoft to not use AMD's "tessellation light code".

It's FUD...and bad FUD.

 

[badb0y]

Now he claims NVIDIA forces Ubisoft to not use AMD's "tessellation light code".

It's FUD...and bad FUD.

Yea, since he never said that. LOL

 

[Idontcare]

Earlier, I caught myself wondering if people on forums like this spend more time arguing about video cards or more time using their video cards for what they were (mostly) designed for.

Not to get too technical on you, it is happy-hour in my house right now and my margarita glass has been emptied three-times over by now, but this is technically the VC&G forum, not the PC Gaming forum.

Presumably we are all here to bitch about the things we don't have, the things we don't want, or the things we don't want to hear about not having as it pertains to video cards.

If we wanted to talk about the kickass stuff we are doing with the things we have then we'd be over in PC Gaming being all hugs and giggles with each other

[/needs 'nother 'garita]

 

[Lonbjerg]

Yea, since he never said that. LOL

Wanna bet he was invovled in this?:
http://hardocp.com/news/2010/10/20/benchmark_wars

And I take you agree he lied on the 2 first counts, since you didn't object to those?

 

[badb0y]

Wanna bet he was invovled in this?:
http://hardocp.com/news/2010/10/20/benchmark_wars

And I take you agree he lied on the 2 first counts, since you didn't object to those?

I don't see Huddy anywhere in there. But since you responded, for you it's not suspect at all that 1 day before the release of 6800 nVidia sent out the benchmark and encouraged reviewers to use it?

 

[Lonbjerg]

I don't see Huddy anywhere in there. But since you responded, for you it's not suspect at all that 1 day before the release of 6800 nVidia sent out the benchmark and encouraged reviewers to use it?

So you agreed he lied in the first 2 cases?
I can ask you again, but have to wonder why you won't answer that question...:hmm:

And since Fuddy's title is Dev. Rel. you think he had no contact to the devs?

 

[badb0y]

So you agreed he lied in the first 2 cases?
I can ask you again, but have to wonder why you won't answer that question...:hmm:

And since Fuddy's title is Dev. Rel. you think he had no contact to the devs?

Nah I never commented on those two because I know nothing about them, I can't comment on something I don't know anything about.

Probably does(it's his job) but why would you imply he is e-mailing reviewers? Do you have any proof of this?

 

[Lonbjerg]

Nah I never commented on those two because I know nothing about them, I can't comment on something I don't know anything about.

Probably does(it's his job) but why would you imply he is e-mailing reviewers? Do you have any proof of this?

I have more proof that Fuddy has to any of his claims...and if you know nothing of the previous statements (as in you don't reallu know Fuddy at all) why do you try and defend him?

Kinda begging for trouble defending an unkown entity...wouldn't you say?

And you DO comment on things you know nothin gabout.
You have access to HAWX2 source code?
If not, why are you posting...against you own claims of not talking about things you know nothing about?

 

[Scali]

Probably does(it's his job) but why would you imply he is e-mailing reviewers? Do you have any proof of this?

Huddy tends to be the guy talking to the press when it comes to more technical/development-related issues, being the head of devrel at AMD.

 

[Aristotelian]

Okay, so what was he trying to say?

Read it yourself. Ask your English tutor for advice if you're still confused about the difference between 'whining' and what was happening in that article.

Also, I find it interesting that there was no comeback to the remainder of my post. Nor can there be. See? The reviews aren't complaining about the lack of tessellation and image quality and so on, but some forums posters are on a crusade to drive the point home as if the potentiality of increased tessellation usage at some point in the future necessarily entails that one ought to buy the better tessellator now. I mean, get real.

 

[Scali]

but some forums posters are on a crusade to drive the point home as if the potentiality of increased tessellation usage at some point in the future necessarily entails that one ought to buy the better tessellator now. I mean, get real.

Hold on just a second.
Let's make it crystal-clear that I never gave ANY advice on what card to buy, or why, okay? Answer me, did I give anything remotely resembling purchasing advice to anyone anywhere?

 

[Gikaseixas]

forums posters are on a crusade

...to shift focus from more relevant issues such as actual gaming performance.

I mean didn't AMD just launch a GTX 460 / GTX 470 rival? And a great one...

 

[Lonyo]

Because he lies.
He claimed NVIDIA paid devs to use PhysX...but offered no documentation.
He claimed NVIDIA hobbled PhysX with X87 code...but leaving out the history of the PhysX API.
Now he claims NVIDIA forces Ubisoft to not use AMD's "tessellation light code".

It's FUD...and bad FUD.

the developer has chosen

Holy force Batman!

 

[Lonbjerg]

Holy force Batman!

So you too agree he lied on the first two points?

And if you would car to look in this thread and others, you can see people posting about how this is "NVIDIA, a TWIMTBP game and deliberate code to make AMD look bad"...you are to intelligent to pull the "dumb" card.

 

[tviceman]

Not to get too technical on you, it is happy-hour in my house right now and my margarita glass has been emptied three-times over by now, but this is technically the VC&G forum, not the PC Gaming forum.

Presumably we are all here to bitch about the things we don't have, the things we don't want, or the things we don't want to hear about not having as it pertains to video cards.

If we wanted to talk about the kickass stuff we are doing with the things we have then we'd be over in PC Gaming being all hugs and giggles with each other

[/needs 'nother 'garita]

Very true I was just rambling out loud. I feel like so many of these arguments are opinion based (RC Cola is way better than Pepsi and Coke, btw) and fueled mostly by loyalty. But Margaritas are nice and I'm glad you are enjoying your afternoon! I'll go back to staying on topic now.

 

[Aristotelian]

Hold on just a second.
Let's make it crystal-clear that I never gave ANY advice on what card to buy, or why, okay? Answer me, did I give anything remotely resembling purchasing advice to anyone anywhere?

As far as I can tell, you've never GIVEN any advice (I'm making the stronger claim, to show my sportsmanship) on which card to buy and who should buy it and so on. Note: I wasn't calling you out as if you're explicitly on some card pushing mission. I don't believe you work for Nvidia or have any agenda to push here regarding credit card sales. That's why I have an affinity for your posting style, because you're seeking a truth behind the various levels of games that are at play in these discussions (marketing, rhetoric and so on).

However, I have an issue with the notion that 'AMD has a problem' (you've used this or a similar phrase on a number of occasions recently). I agree(d) with you straight away that IF lots of games in the future start to become more and more dependent on powerful tessellators AND IF AMD does nothing to improve their tessellation performance (let's see how the 6900 series fare: there seem to have been decent improvements in this 6800 refresh/whatever), THEN THERE WILL BE a problem for AMD, because their cards will underperform relative to Nvidia's offerings which (at least for now) have superior tessellation performance.

But, and I really mean but, tessellation is not the totality of what gaming is (yet). Since it is not (currently), and this is illustrated perfectly in the benchmarks in these latest reviews, AMD cards still play games well NOW. So, if I'm going to agree with you that AMD might at some point in the future have a problem, that is dependent on the conditional above (if, regarding tessellation performance, if games need it, and so on). It was your phrasing of AMD has a problem. Has entails currently. What is AMD's current problem? I don't see it. If games were as dependent on tessellation as Nvidia would like them to be right now, then we would see huge performance differences across many games (not just isolated cases or benchmarking tools that are currently irrelevant). We don't see such performance differences. So AMD's cards are competitive even with weaker tessellators, and they are doing fine. I don't see their current problem. Understand?

 

[Scali]

As far as I can tell, you've never GIVEN any advice (I'm making the stronger claim, to show my sportsmanship) on which card to buy and who should buy it and so on.

Good, then that is perfectly clear.
I don't care what people buy... with the current 6800 cards and their prices, it's pretty hard to resist AMD anyway, at this point.

But this thread is not about which card to buy, it is about some statements that Huddy made in an interview, and a lot of it revolves around tessellation.
And that is what I was talking about, how AMD's tessellation differs from nVidia's tessellation in terms of implementation and performance characteristics.

If you go pros and cons, then tessellation may still be a con, but the pros might outweigh that, and you'd still arrive at choosing AMD for your next videocard, that's all fine by me. That's up to everyone to decide for themselves.

But, and I really mean but, tessellation is not the totality of what gaming is (yet).

Yes, but I'm afraid I'm a bit ahead of you there already.
What is nVidia good at? Developer relations!
They have this pretty successful TWIMTBP program, where they will try to push their latest features into any game they possibly can.
They've been doing it with PhysX, 3D Vision, and now tessellation.
So those games will come.

It was your phrasing of AMD has a problem. Has entails currently. What is AMD's current problem? I don't see it.

Well, you have to understand that I'm a developer, and that is my perspective on things.
For me, the future is now.
In fact, I was already working on a tessellation implementation using Cuda sometime before DX11 was out. I had already decided years ago that it would be the future of graphics.

So you may not see it yet (but that would mean you are blind to synthetic benchmarks as well), but I do. And I think you will be seeing it soon.

 

[Dark Shroud]

Well, you have to understand that I'm a developer, and that is my perspective on things.
For me, the future is now.
In fact, I was already working on a tessellation implementation using Cuda sometime before DX11 was out. I had already decided years ago that it would be the future of graphics.

The problem with your statement is that you only speak for yourself. I'm stopping here because I could but do not want to start a brand new argument on what you've said.

So you may not see it yet (but that would mean you are blind to synthetic benchmarks as well), but I do. And I think you will be seeing it soon.

Two problems for you here.

A) How long is "soon?" Because you can bet AMD already has the 7000 & 8000 series cards in development. In addition said AMD 7000 series will be released in roughly a years time on a new 28nm process. So unless AMD does absolutely nothing in tessellation performance their cards will do enough for the mid range market. You should at least wait until the 6900 series cards have been released and benchmarked before making comments like this. As the GTX 480 & GTX 460 do not have the same tessellation performance, which is the same for the HD 6800 & HD 6900 series.

B) Synthetic benchmarks are just that, synthetic. In the real world many devs cut corners and are developing for Consoles first. So while your crystal ball depicts problems for AMD "soon" the masses will not see said problems on their consoles; most of which use AMD GPUs.

 

[Scali]

The problem with your statement is that you only speak for yourself.

Firstly, is there a problem with that? I think every forum user normally speaks for themselves.
Secondly, am I really just speaking for myself? I mean, if I'm a developer, and I think this is the future, wouldn't it be perfectly possible that I, being a professional, working with a team of other professionals, have given this issue a lot of thought, and arrived at this conclusion... that all over the world, many other developers have reached very similar conclusions? I mean look at the marketing material from both AMD and nVidia. Both are promoting tessellation, and exactly in the way I described... Huddy's early DX11 blog says *exactly* what I'm saying. Huddy just changed his mind after Radeons started getting owned at this (even in AMD's own code, in AMD-sponsored DX11 games) by Fermi.
It's clear what's happening here.

A) How long is "soon?" Because you can bet AMD already has the 7000 & 8000 series cards in development.

That will probably be a year off or so. I think the games will arrive sooner than that.
I don't see what the big deal is anyway. In the end it will only mean that AMD users will have to play with lower, or even no tessellation in these games.
The real problem is that the AMD camp is not willing to accept the fact that nVidia does this one thing better (which is pretty hypocritical after the many months of gloating over being the only one with DX11).
If you can just simply accept the fact that your card may not be able to run all games at maximum settings, then really, what is the big deal?
Heck, I bought a GTX460 recently. It may be relatively good at tessellation, but being a mainstream part it's not going to be able to run most games at full detail anyway. I don't care.

B) Synthetic benchmarks are just that, synthetic.

Nope, they aren't. But I already grew tired of arguing that point with laymen back in the GeForce FX days. I have to admit, I got a nice sick sort of satisfaction out of the fact that they all bought GeForce FXes against my advice, and then came running back to the forums crying when all SM2.0 games completely tanked on their useless cards.
Synthetic benchmarks eh? If you're dumb enough to think that, you deserve what you get: a card with useless performance.

 

[Aristotelian]

The real problem is that the AMD camp is not willing to accept the fact that nVidia does this one thing better (which is pretty hypocritical after the many months of gloating over being the only one with DX11).

Right, and given our recent discussion here I hope that you do not consider me 'in the AMD camp'.

Nope, they aren't. But I already grew tired of arguing that point with laymen back in the GeForce FX days. I have to admit, I got a nice sick sort of satisfaction out of the fact that they all bought GeForce FXes against my advice, and then came running back to the forums crying when all SM2.0 games completely tanked on their useless cards.
Synthetic benchmarks eh? If you're dumb enough to think that, you deserve what you get: a card with useless performance.

Isn't that backwards? If someone buys a 6850 to play games now having looked at how the card does in gaming benchmarks now and you say (perhaps rightly) that in the future lots of games will use more and more tessellation how is the card 'a useless performer? I'm not accusing you of pushing cards, but the extension of what you are saying seems to be that you think it's better to buy Nvidia's parts now, even though AMD's parts compete well in almost every game, because they are better tessellators. Yet it seems the tessellation itself is the 'useless performer' if it is simply a potential waiting to be actualized in the future.

Do you only develop for PCs? How is this tessellation push impacting console game development, or did the 360/PS3 etc ship with great tessellation power prior to this...media push?

 

[SirPauly]

For me, this is as subjective as subjective can be and simply allow the market to answer which is the correct or wrong way of implementing tessellation for its specific window of time.

Views may differ and debate is always healthy and constructive as long as it doesn't get personal.

 

[Scali]

and you say (perhaps rightly) that in the future lots of games will use more and more tessellation how is the card 'a useless performer?

Well, that was with regards to the GeForce FX. You effectively couldn't run SM2.0 games on it, you had to run them at DX8.1 quality, else the performance was very bad, pretty much unplayable.

With tessellation, you may get something similar, where you need to run on lower tessellation settings than similarly priced nVidia cards, to get the same framerates. Although the visual impact will probably not be as dramatic as SM1.x -> SM2.0 was.

Technically the cards aren't 'useless', they're just not very good at what they were meant to be (DX9 and DX11 cards respectively).

I'm not accusing you of pushing cards, but the extension of what you are saying seems to be that you think it's better to buy Nvidia's parts now, even though AMD's parts compete well in almost every game, because they are better tessellators.

No, what I'm saying is: AMD's tessellator is just considerably less powerful than nVidia's. So if you buy AMD, be aware that you will most probably not be running the latest games with full tessellation on.
Obviously as we all know, nVidia's cards have their own share of issues, such as the power consumption and heat being considerably higher than similarly performing Radeons...
Whichever you choose is up to you, but somehow the consequences regarding nVidia seem obvious to everyone, while most people seem to be in denial about any of AMD's shortcomings.

Do you only develop for PCs?

Yes

How is this tessellation push impacting console game development, or did the 360/PS3 etc ship with great tessellation power prior to this...media push?

As far as I know, the PS3 has no tessellation abilities. The 360 has a tessellator (it uses an AMD GPU, and a very similar tessellator could also be found on AMD's DX10 GPUs), although it's not as powerful as the DX11 one.
To be honest I have no idea whether 360 games actually make use of the tessellator at all. But I think that by itself is telling... because I see two options:
1) Games use the tessellator, but apparently it doesn't make enough of an impact that we would actually notice, so it's not such a spectacular feature.
2) Games don't use the tessellator, so apparently developers think it's not even worth using.

 

[Scali]

For me, this is as subjective as subjective can be and simply allow the market to answer which is the correct or wrong way of implementing tessellation for its specific window of time.

Despite what AMD wants people to believe, the issue isn't even about ways of implementing tessellation.
It's actually about scalability and maximum levels of detail.

And for that, I think there's only one answer: more is better.

 

[dust]

Isn't that backwards? If someone buys a 6850 to play games now having looked at how the card does in gaming benchmarks now and you say (perhaps rightly) that in the future lots of games will use more and more tessellation how is the card 'a useless performer? I'm not accusing you of pushing cards, but the extension of what you are saying seems to be that you think it's better to buy Nvidia's parts now, even though AMD's parts compete well in almost every game, because they are better tessellators. Yet it seems the tessellation itself is the 'useless performer' if it is simply a potential waiting to be actualized in the future.

I think you cleared it for me as well. So all we have to do is to buy the Fermi Herrenvolk now, put the hands together, pray and wait for tessellation to take over the world. It would be foolish to buy the AMD Untermensch now.

Ty, gb, gl and most of all gg!

 

[Scali]

I think you cleared it for me as well. So all we have to do is to buy the Fermi Herrenvolk now, put the hands together, pray and wait for tessellation to take over the world. It would be foolish to buy the AMD Untermensch now.

Ty, gb, gl and most of all gg!

This was NOT any kind of buying advice, as WAS ALREADY MADE PERFECTLY CLEAR IN THE PREVIOUS POSTS

The point is that AMD, by means of Richard Huddy, is trying to stop the progress of tessellation in games. Not because it wouldn't deliver better image quality AND better performance, but simply because their hardware can't handle it.
They don't care about gamers, and delivering the best possible gaming experience. They're just trying to sabotage their competitors. Instead of offering GPU physics, they've only been on a PR campaign to sabotage GPU physics, and now they're trying to sabotage tessellation, rather than trying to develop good hardware for it (which is a STANDARD DX11 feature).

Again, it's not about what YOU buy, I don't care about that.
It's about how I disapprove what AMD and Huddy are trying to do.