Richard Huddy - Gaming Scientist Interviewed

[GoogalyMoogaly]

nvidia never said that amd had access to the code, they said devs who bought the licence have access to it, and they also said AMD doesnt need the source code to optimise, because there are other ways, all of this implies that devs are under NDA and cannot share the code with a 3rd party who doesnt have the licence, things that huddy explained and pointed out that devs told his engineer team that the nda forbids them from giving it to AMD.

And Nvidia have said that developers aren't forbidden from working with AMD to optimise.

 

[Marine-RX179]

So cards would need to be mantle certified to run it? Kind of like how we have dx9/10/11 certified cards at the moment?

Say nvidia and Intel adopt it tomorrow and their new graphics archs support mantle, does it get less efficient each year? What about in 10 years? If dx11 is already bloated from 4 generations of cards (from an architecture count that's not many per vendor) how many cycles is it going to take to bloat it to the point it becomes inefficient again?

No I that's not what I meant at all. The mentioned of dx being "bloated" is because of all the codes it accumulated over the years (regardless of being efficient or not). Mantle has advantage in terms of position because it had a "fresh start", and is designed and developed from ground up in the more recent years.

In a way it is similar to why Milton Keynes seem more modern than most town/city in the UK- it was developed almost like developed from ground up and doesn't have a long history dating back centuries ago with historic towns, buildings and landmarks that cannot be rid of with any new developments MUST be built going around them....the same apply for Shanghai in China as well.

 

[opethdisciple]

My summary:

Makes me wonder if Richard has actually played BF4...

 

[arc@css]

Interesting to watch. An articulate presentation by Dicky Huddy, though I do wish someone would give him a fisherman's friend.

 

[AlamoX]

And Nvidia have said that developers aren't forbidden from working with AMD to optimise.

to optimise what ? a DLL ? thats the whole point behind this, devs dont have the engineering expertise for a vendor's hardware, why do you think devs often call for gpu hardware vendor support... devs having the code is fine and all, but that doesnt mean they can optimise it correctly especialy if nvidia forbids them from telling AMD about the code.
yes developers arn't forbiden to work with AMD to optimise they game in general, but when it comes to gameworks they are forbiden to share with AMD, now just enlighten me how does this colaboration works with the NDA in place? im curious!

 

[humbug]

Thanks Matt, will listen to this

 

[Gregster]

Cheers for the breakdown Googly. I don't have time to watch it.

 

[_Alatar_]

So much blatant PR with tons of stupid talking points.

It's almost like listening to american politicians.

Couple of points:

1)
He basically said that if NV and intel were to adopt the API AMD would allow them to suggest some small fixes and stuff like that but other than that the "future" of the API would still be completely controlled by AMD.

And that of course basically means that if NV and intel were to be tied down to mantle AMD would have the capability of changing and updating the API whenever they want in a way that would cripple the competition. NV's next architecture is going to focus on X? Let's make doing X with mantle difficult.

And that's why it's never going to be a universal API between the manufacturers. If that is what AMD wanted they would have to hand it completely over to a 3rd party. Because right now if they're complaining about the possibility of foul play with Gameworks I have no clue how they can justify making the mantle comments considering mantle adoption by intel or nvidia would allow AMD to do things that are 100 times worse than Gameworks is even theoretically capable of.

And well, not to mention the fact that despite AMD telling the press that they want intel and Nvidia on board at least intel reps have come forward saying that they've been denied access to info and specs they've asked many times.

2)
Complaining about overtesselation is extremely petty. It's not like AMD backed teams don't go out of their way to play to the strengths of AMD software/hardware at the moment.

Take star swarm for example, it has an extremely inflated amount of draw calls purely to make mantle look better. And absolutely nothing is stopping devs from doing this exact same thing in games just to trump up mantle.

But you know what Nvidia did about it? They didn't complain but instead came out with DX11 improvements that have allowed NV to tie/surpass AMD in thief, surpass then in star swarm and come close in BF games. Not to mention getting improvements in tons of other titles that don't even have mantle. Maybe AMD should just improve their tesselation instead of complaining about batman's cape being too hard to render.

3)
Huddy's talking points about GameWorks were the exact same as hallock's. No proof, just some blatantly false assumptions and accusations designed to do nothing but change public perception.

No regard for actual facts.

btw if someone wants to talk to someone who has actually licensed gameworks we have one guy on OCN: http://www.overclock.net/t/1496631/...ls-mantle-an-open-source-api/30#post_22437361

All in all I find it sad that Ryan didn't call out Huddy when he was so extremely obviously spewing just utter PR/talking point garbage.

Money talks and bullcrap walks I guess.

 

[Martini1991]

But the difference is mantle don't effect nvidia users at all, unlike Gameworks is supposed to.

Hasn't it been proven this simply isn't the case?

 

[Rusty0611]

It has but they hope if they keep repeating it, it will hopefully become the accepted norm.

 

[_Alatar_]

It has but they hope if they keep repeating it, it will hopefully become the accepted norm.

Exactly. It's the oldest trick in the book if you're into politics.

Here it's just applied to hardware/software discussions.

 

[Martini1991]

It has but they hope if they keep repeating it, it will hopefully become the accepted norm.

Amongst a few, it is the accepted norm.

Sometimes being part of team red hurts.

 

[Orangey]

So much blatant PR with tons of stupid talking points.

It's almost like listening to american politicians.

Couple of points:

1)
He basically said that if NV and intel were to adopt the API AMD would allow them to suggest some small fixes and stuff like that but other than that the "future" of the API would still be completely controlled by AMD.

And that of course basically means that if NV and intel were to be tied down to mantle AMD would have the capability of changing and updating the API whenever they want in a way that would cripple the competition. NV's next architecture is going to focus on X? Let's make doing X with mantle difficult.

And that's why it's never going to be a universal API between the manufacturers. If that is what AMD wanted they would have to hand it completely over to a 3rd party. Because right now if they're complaining about the possibility of foul play with Gameworks I have no clue how they can justify making the mantle comments considering mantle adoption by intel or nvidia would allow AMD to do things that are 100 times worse than Gameworks is even theoretically capable of.

And well, not to mention the fact that despite AMD telling the press that they want intel and Nvidia on board at least intel reps have come forward saying that they've been denied access to info and specs they've asked many times.

I think we can safely assume now that Mantle will never be open and neutral. With all these different codebases and forks for Intel, NV etc you would think control by a working group like Khronos is the only sane thing to do. Do AMD seriously expect their direct competitors to just trust them??

 

[triss]

I think we can safely assume now that Mantle will never be open and neutral. With all these different codebases and forks for Intel, NV etc you would think control by a working group like Khronos is the only sane thing to do. Do AMD seriously expect their direct competitors to just trust them??

I thought when it was launched it was going to be open and handed over?

 

[Orangey]

I thought when it was launched it was going to be open and handed over?

IIRC they merely mused about possibly maybe thinking about it?

Someone is gonna have to find the Q&A or whatever where they talk about it.

 

[andybird123]

IIRC they merely mused about possibly maybe thinking about it?

Someone is gonna have to find the Q&A or whatever where they talk about it.

http://www.techradar.com/news/compu...g-api-to-become-the-industry-standard-1218560

"After that phase is done, we do hope that Mantle becomes an industry standard. We'll be releasing a public SDK later this year, and hope that others adopt it. If they don't adopt it itself, then we hope they adopt APIs similar to it that become an industry standard for PC gaming."

It's a nice sentiment, but does that mean AMD is willing to allow an entity traditionally viewed from behind enemy lines make use of Mantle?
"It's hard to say," Hallock said. "If we want it to be an industry-wide API or inspire an industry-wide API, that would mean adoption from Nvidia in some way, shape or form. I can't speak from an architectural level what that would require of them to change."
"But for the good of gamers, [we] would want one ultimate specification that is either Mantle itself or one similar to it. As Highlander said, 'There can only be one.'"

there have since been reports that AMD are quoted as saying that with DX12 looming (e.g. due for release to developers around the same time as the mantle SDK) that it is unlikley that mantle will become open and have cross vendor support - which totally makes sense, because there would be little to no benefit for Nvidia to support mantle if DX12 already does, and AMD already saying that they don't know what architecture changes nvidia would need to make to support mantle (e.g. suggesting that they would)

listening to this interview, he says that AMD will always "own" mantle, meaning they are not going to pass it off to a 3rd party, which means it will never be considered an open standard

 

[tommybhoy]

Hasn't it been proven this simply isn't the case?

Depends what side of the fence you sit on, both have fingers in ears in discussion of the matter, there is zero conclusive proof either way.

 

[llamadal]

Hasn't it been proven this simply isn't the case?

Who needs facts when you have assumptions!

 

[_Alatar_]

Depends what side of the fence you sit on, both have fingers in ears in discussion of the matter, there is zero conclusive proof either way.

lol...

The only question that we should be asking is "Is there proof of Gameworks crippling AMD performance?"

The answer to that is no.

You can't possibly be suggesting that people should also ask if there's proof of it not doing that? That's the guilty until proven innocent argument that never really works. You can apply it to almost anything and make anything look bad for any reason.

 

[Rusty0611]

The thing is Tommy (and I do agree with your sentiment) is that if you're accusing somebody of doing something 'bad' you must provide proof of the wrong-doing. The proof that was put forward before has been shown to be nonsense. So therefore just because you can't prove that nVidia aren't doing anything wrong doesn't mean there's any particular reason to be suspicious.

That would be like saying just because you can't prove that on the 10th May 1985 Bob Marley didn't fly without wings after jumping off the Eiffel Tower therefore we must talk about it as if it is likely. OK the example is extremely exaggerated to make the point but the point is important. The argument put forward by the people who so think that GW is harming AMD is flaky at best.

Could it be used in future to potentially harm AMD? Definitely yes but until that happens there's not much point discussing it.