How will NVidia counter the release of HD5xxx?

[Dreamweaver]

So on a 2-year refresh cycle... well if in those first 2 years, the DX11-content doesn't exceed the DX10-content much (and if the move from DX9 to DX10 is anything to go by...), then yes, PhysX could actually be the better value (better performance, better graphics, all the main selling points for DX11?).
At least in the first year of this DX11 hardware, I don't expect to see much in terms of physics support, and probably not too much DX11 usage either.
So yes, what I'm saying is: is this DX11 future proofing going to be relevant for the lifetime of this particular card?

For whatever reason though (pricing mainly, bang for buck etc) Joe Public has already voted against PhysX and NVIDIA in the last generation with the 48xx series, (I believe) outselling the more expensive GT series, or are you saying that PhysX is reaching a critical point where it is going to become more important to Joe Public to the point where they will choose previous generation products over DX11 future products?

I just don't think it is prime time for PhysX, Havok or whatever...

take care,
J.

 

[Scali]

For whatever reason though (pricing mainly, bang for buck etc) Joe Public has already voted against PhysX and NVIDIA in the last generation with the 48xx series, (I believe) outselling the more expensive GT series

I hope you realize that sales figures don't directly correspond to the customer's perception of PhysX.
48xx just had very attractive pricing, and for most of the 48xx' lifetime, the customer had little or no idea about this thing called PhysX in the first place.
DX11 cards will likely be more expensive than competing DX10 products, so the situation will be completely different.

or are you saying that PhysX is reaching a critical point where it is going to become more important to Joe Public to the point where they will choose previous generation products over DX11 future products?

It might, yes. With titles like Mirror's Edge, Cryostasis and Batman AA, the customers are slowly becoming aware of the existence of PhysX, and realizing that there's some extra visuals to be had in various games.
Customers have no idea about DX11 however, and we don't know how long it will take until DX11 offers anything worthwhile over DX10.
They'll just buy whichever card gives them the best visuals and fastest performance... In PhysX titles it's no contest. They probably won't actually know that they're buying a DX10 card instead of DX11, but they do know they are getting better visuals/performance. Don't forget that the end-user experience is what matters, not who has the highest DX number on the tin.

 

[Mariner]

(As I already said, who cares about 'faster'? It should be about creating more detailed, more dynamic environments).

This would obviously be of benefit but current physics middleware doesn't seem to provide more dynamic environments, does it? Just graphical 'bling' which is a pity really.

 

[Scali]

This would obviously be of benefit but current physics middleware doesn't seem to provide more dynamic environments, does it? Just graphical 'bling' which is a pity really.

Cloth, smoke, flying paper/leaves, destructable floors/walls, it's all dynamic (as in: it actually responds to character movements and actions).
Call it 'graphical bling', but it is dynamic graphical bling, and that's the point.
Without accelerated physics, DX11 titles will have a far more static and 'empty' appearance to them.

 

[John Reynolds]

Arguing that PhysX support outweighs DX11 support at this point in time, for this upcoming GPU generation, is not a position I can get behind.

 

[Scali]

Arguing that PhysX support outweighs DX11 support at this point in time, for this upcoming GPU generation, is not a position I can get behind.

That's odd, since there's already various titles with PhysX out, and no DX11 yet. This is a simple and obvious fact.

 

[digitalwanderer]

That's odd, since there's already various titles with PhysX out, and no DX11 yet. This is a simple and obvious fact.

Considering Windows 7 and DX11 isn't out yet is it all that surprising?

 

[Scali]

Considering Windows 7 and DX11 isn't out yet is it all that surprising?

No, that's not the part that surprised me...

 

[digitalwanderer]

No, that's not the part that surprised me...

I give up, which part surprised you? :|

 

[digitalwanderer]

Without accelerated physics, DX11 titles will have a far more static and 'empty' appearance to them.

What makes you think DX11 titles won't have accelerated physics?

Now I know they won't have physX, but you really think they won't have physics when it's built into the new API?

 

[Scali]

What makes you think DX11 titles won't have accelerated physics?

Now I know they won't have physX, but you really think they won't have physics when it's built into the new API?

There is no physics built into the DX11 API. There's DirectCompute (which works on DX10 hardware aswell), but so far I haven't heard of any physics library that supports it.
DX11 *could* have accelerated physics... but not at introduction, most probably not in the first year, and possibly not in the second year either.
Havok with OpenCL is a more likely way to accelerated physics for AMD hardware, at least on the short term. And even that doesn't look like it's going to happen anytime soon.

So, that means that I don't see AMD's DX11 hardware doing ANY physics in any major titles for the first 6-12 months, and if they do get physics afterwards, it will likely be Havok with OpenCL (which also works on DX10 hardware, obviously), not DX11.

 

[digitalwanderer]

Doh, sorry. I got confused and I unfortunately can't explain it to you yet, ask me in a couple of weeks and I'll explain my misunderstanding.

 

[SiliconAbyss]

They probably won't actually know that they're buying a DX10 card instead of DX11, but they do know they are getting better visuals/performance. Don't forget that the end-user experience is what matters, not who has the highest DX number on the tin.

So, Joebob Dumbpants is too stupid to know the difference between DX10 and DX11, but knows that PhysX is going to give a better "end-user experience"? I think you've crossed over from questionable posting to downright nonsense. Essentially what you are saying is, DX11 will never offer anything better or compelling vs. DX10, and will most certainly never be as important as PhysX.

Well one thing is certain, people that predicted Nvidia would push PhysX ad nauseum to try and wet blanket ATI's impending launch were bang on.

 

[madyasiwi]

Just took a quick look, but strangely they don't market it as a complete physics engine ... and Force Unleashed also uses Havok. I guess using the DMM method for everything is too expensive. Still, it's "just" flexible/breakable joints ... I don't see why they can't do the simulation with havok, maybe the artist tools for DMM are a lot better?

Although it's a cool touch that they support non elastic deformation too ... don't see that too often.

Havok didn't have cloth simulation and destruction features back then. These features only available since mid 2008. And AFAIK they still don't have soft body physics either, neither force fields, let alone fluids, and don't event think about turbulence.

On the other hand Pixelux DMM is a very specialized physics engine, just like Havok Cloth and Havok Destruction (which are distinct products that need to be $purchased$ separately from the core Havok Physics), so it still needs another physics engine to do everything else.

PhysX has this kind of cloth based "non elastic deformation", but is not as realistic. Although it could be more suitable for certain other cases.

 

[Scali]

So, Joebob Dumbpants is too stupid to know the difference between DX10 and DX11, but knows that PhysX is going to give a better "end-user experience"?

They 'know' what they see.

I think you've crossed over from questionable posting to downright nonsense.

Not at all, you however have crossed into personal insults. Go away.

Essentially what you are saying is, DX11 will never offer anything better or compelling vs. DX10, and will most certainly never be as important as PhysX.

I never said never. Try to actually read my posts. I know it's difficult for you, but please, just try. I've always mentioned a 1-2 year window. Crucial to the rest of the story.
What I said was that initially, DX11 hardware (regardless of who makes it), will either be running straight DX9/DX10 code, or DX10 code that is ported over to DX11 with little or no usage of the extra features. Current games will make no use of DX11 at all, obviously. Future games may make use of DX11, but it remains to be seen when such games arrive, and what they will actually be offering.
I think it's pretty safe to say that at launch next week, there's not going to be any DX11 game. There aren't any now, and there haven't been any announcements that any will be released soon.

Still with me? Good. Now, I've been saying that it's probably going to take about 6-12 months before any DX11 games will come out, so during that time, DX11 will not be offering anything better or compelling, as you'll just be running DX9/DX10 games.
In the second year you *might* get some mileage out of the DX11, but that remains to be seen. If the initial DX11 offerings will be as weak as the DX10 ones were (the GeForce 8800's success had nothing to do with the DX10 features), then for the first 2 years, DX11 isn't going to be much of a deal. By then a new generation of cards will have come out anyway.

Where does PhysX come in? Well, that's simple... because there ARE games with PhysX, which you can already use, and given the way nVidia has been promoting PhysX so far, it looks like there will be plenty more games with PhysX in the coming 2 years.

So what does it boil down to?
Your DX11 card will be running DX9/DX10 games with CPU physics only.
Your DX10 card with PhysX will be running the same games, where some of them will have extra effects through GPU physics.
Result: some games will have more dynamic effects and more eyecandy on these DX10 cards, while none of the games will look any better or more dynamic on the DX11 cards.
So what does the end-user see? The DX10-card with PhysX has the better graphics.
Can't argue with that, really. All you can do is give some "might have, could have" on the future of DX11, Havok, Bullet, and all that... but I will keep pointing out that since there have been no announcements made, it is unlikely that much is going to happen in the next 6-12 months on that front. By that time, Intel and nVidia will likely also have DX11 hardware out, and any physics solution that will be running on AMD's hardware, be it Havok, Bullet or something else, will be running on OpenCL or DirectCompute, and as such also work on the competitors (even the DX10 hardware).

 

[FrameBuffer]

PhyX is the future of gaming. We have DirectX 11 right where we want it.. in it's last throughs.


MODEDIT FOR EVERYONE: Unless you have proof, please refrain from acusing other members of doing an IHV's crowd control and questioning his or her motives

 

[green.pixel]

Is there a possibility that nVidia could port PhysX to DX Compute or OpenCL?

 

[FrameBuffer]

Is there a possibility that nVidia could port PhysX to DX Compute or OpenCL?

there is always the possibility however unless there is a distinct performance advantage, there is no real incentive for them to do so.

 

[Silent_Buddha]

Can pretty much guarantee that if there's even the slightest chance that PhysX ported to Dx11 or OpenCL will perform better on ATI or Intel hardware that Nvidia will not allow it to be done.

Regards,
SB

 

[ECH]

Havok didn't have cloth simulation and destruction features back then. These features only available since mid 2008. And AFAIK they still don't have soft body physics either, neither force fields, let alone fluids, and don't event think about turbulence.

On the other hand Pixelux DMM is a very specialized physics engine, just like Havok Cloth and Havok Destruction (which are distinct products that need to be $purchased$ separately from the core Havok Physics), so it still needs another physics engine to do everything else.

PhysX has this kind of cloth based "non elastic deformation", but is not as realistic. Although it could be more suitable for certain other cases.

The most important aspect of DMM (but certainly not the only aspect) is that anyone can play the game with it. As long as you meet the hardware requirements. Again, I am not sure why this isn't being marketed for PC games as of yet though. Perhaps this is AMD's opportunity to either buy them or promote them along with their developer's relationship program.

In either case if Nvidia is trying to counter with just Physx AMD can counter with DMM IMO. Inwhich we will now have equality as far as having 2 physic technologies that were already developed.