• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Bye bye.. Ageia :)

Why would they make the ability available on the video card when the P5N-SLI Premium will have an extra pci-e that was intended for an Nvidia based physics card? HMMMMMM

Also it states that Quantum is supposed to compete with Ageia. Since Ageia really has survivability on its own in the first place, why didn't Nvidia just do the nice thing and acquire them instead of inevitably forcing them to their grave?
 
TekSomniaK said:
Why would they make the ability available on the video card when the P5N-SLI Premium will have an extra pci-e that was intended for an Nvidia based physics card? HMMMMMM
Click to expand...

Well nVidia didn't design that board, ASUS did. ASUS may have also have thought that nVidia would do something like ATi said they would and simply use a third GPU for it. That slot near as I can tell is no different than any other x8 PCIe slot in an x16 form factor.
 
Dan_D said:
Well nVidia didn't design that board, ASUS did. ASUS may have also have thought that nVidia would do something like ATi said they would and simply use a third GPU for it. That slot near as I can tell is no different than any other x8 PCIe slot in an x16 form factor.
Click to expand...

Yes I realize that, but this whole thing is stupid. This whole fiasco, while it might save some money for consumers, is just like if Intel were to develop their CPUs so that video cards are no longer needed.
 
The article doesn’t really say much but hey, making sensationalist posts that exaggerate the facts is the cool kid thing to do right?

You do know that physics involves API’s right? There isn’t just this magical thing out there called “physics” that any hardware can run. If a game uses the PhysX engine it will either be accelerated by the PPU or it will use the CPU in software mode.

What do you think is going to happen when you pop in a PhysX API game with your 1337 new all powerful physics powerhouse G80? Likely about the same as running a Voodoo3 as far as physics is concerned. Of course that could go the other way with games supporting Havok or a proprietary NV/ATI API and not PhysX but my point is that this isn’t like Intel vs. AMD where you can run the same software on both and make an even comparison.
 
GotNoRice said:
The article doesn’t really say much but hey, making sensationalist posts that exaggerate the facts is the cool kid thing to do right?

You do know that physics involves API’s right? There isn’t just this magical thing out there called “physics” that any hardware can run. If a game uses the PhysX engine it will either be accelerated by the PPU or it will use the CPU in software mode.

What do you think is going to happen when you pop in a PhysX API game with your 1337 new all powerful physics powerhouse G80? Likely about the same as running a Voodoo3 as far as physics is concerned. Of course that could go the other way with games supporting Havok or a proprietary NV/ATI API and not PhysX but my point is that this isn’t like Intel vs. Nvidia where you can run the same software on both and make an even comparison.
Click to expand...

Yeah, but then Nvidia is with Havok, which a ridiculous amount of games are using now. Ageia has what? UT2007 and what else?

How long will it be until we have a standardized physics handling unit? I mean Alan Wake will utilize the extra cores on Kentsfield processors to handle its physics, there will be Ageia's Physx, and then now Nvidia and ATI with their own things. WTF?
 
TekSomniaK said:
Why would they make the ability available on the video card when the P5N-SLI Premium will have an extra pci-e that was intended for an Nvidia based physics card? HMMMMMM

Also it states that Quantum is supposed to compete with Ageia. Since Ageia really has survivability on its own in the first place, why didn't Nvidia just do the nice thing and acquire them instead of inevitably forcing them to their grave?
Click to expand...

Why wouldn't they? Why worry about not being able to fit a PhysX card into your case because of your oversized GPU HSF(s), when the whole thing could be built right into the video card. I for one think this is awesome, of course not for Ageia, it kind of makes me feel bad for them, but shit happens I suppose.
 
“Yeah, but then Nvidia is with Havok, which a ridiculous amount of games are using now.”
None of those Havok game are going to benefit from Nvidia doing physics. The game needs a patch or update to use the GPU for physics and no major games are announced to support Nvidia and havok physics on the GPU. What use is Nvidia physics is no games support it? Its got less support then Ageia so far and I dont see it getting lots of support due to the high cost.




“Ageia has what? UT2007 and what else?”
There are lots of Ageia games out but assuming you mean PPU only UT and all the games based of its engine. Warhammer Online could turn out to be big as Warhammer is a popular franchise. Also Scared II could be a good game as Scared I did well. Not a triple AAA game but a good game.
 
Arcenite said:
Why wouldn't they? Why worry about not being able to fit a PhysX card into your case because of your oversized GPU HSF(s), when the whole thing could be built right into the video card. I for one think this is awesome, of course not for Ageia, it kind of makes me feel bad for them, but shit happens I suppose.
Click to expand...

Its a new feature they can add to GPU's and a new feature they can charge us money for.

If they can do it without spending money on buying out Ageia, then even better.
 
If people think that buying 3 video cards or even buying 2 and having one dedicated to physics is a better solution than aegias then I am sorry, but you lot a bunch of nubs if you think that.

do you really think having a video card working as a dedicated ppu is a good solution ? i would rather have them both working in sli and have the ageia for physics than them working as gpu and ppu.

thats my personal opinion, but no way in hell will i buy 2 or 3 video cards to turn one of them into a ppu.

just not on, and a video card cannot do physics and act as a gpu to any great extent, not like ageias card can do physics, they can do one or the other to a good degree, but both together they will not be able to pull that off and have leet physics and leet fps at the same time.

What people sem to forget is, when and if there is a standard physics api, Aegia will just start to use that instead, as it can basically use anything that it is programmed to use, so the ageia physics card is far from being retired.
 
If people think that buying 3 video cards or even buying 2 and having one dedicated to physics is a better solution than aegias then I am sorry, but you lot a bunch of nubs if you think that.
Click to expand...

It's the fact that Ageia doesn't have many games for it to shine, while Havok has been a player for years.

It's about how many games utilize a certain API rather then the actual ability for the hardware to calculate the actual physics.

If Ageia can do 10 trillion-billion physics calulations in a 1/10th of a second but only has two games with utilize this, it's kind of lame to buy a $300 peice of hardware of two games.

If nVidia/ATI can make realistic physics (while not nearly the ability of Ageia) but has hundreds of games and can do it for $150, then, it kind of makes more sense to use a GPU.

It's not about the ability of the hardware to do physics, it's about the ability of games to use the hardware.

Once there's a standard API, I'll jump on. But, there needs to be an API (most likely Microsoft will develop and intergrate it into Vista) before the PPU becomes a 'must have' item for games.
 
I think that the smartest solution would be to wait and see. After all, we are comparing a not-yet released card that may support hardware accelerated physics on not-yet released titles with a released card, which will support hardware accelerated physics on not-yet released titles.

I think I will make up my mind when I get to compare the two solutions.
 
EQTakeOffense said:
It's the fact that Ageia doesn't have many games for it to shine, .
Click to expand...
Do you know what your saying ?

Ageia is the only Physics card to have any games available for it, nvidia and ati dont have any games, in the future yes, now, NO

I agree with you on the standard api, which we will just have to wait and see what happens there, but as for games, you cannot even think about saying ati or nvidia have more, as they have 0 at the moment, even tho that will change in the future, but the future isnt now.

Havok = we have that now without any need for anything else than a video card, i,e, halflife2 physics is havok, and we dont need a ppu for that.

Havokfx = needs hardware to be able to run with any degree of leetness.

Thing is, Ageia will take off, as the ps3 has ageia physic in it, just like the xbox360, so it isnt going anywhere, as the 360 and the ps3 are both gonne sell millions upon millions of consoles, so ageia isnt going anywhere anytime soon.

the big coup for ageia was having it intergrated into consoles, 360/ps3, as it will then extend over to pc's to a large degree sooner or later, whether there is games available for it now is really irrelative, as once the ps3 comes out then you can bet that the physics game will really start..
 
it still amazes me how pp wish to see ageia fail when they were the ones to really push the idea of accelerating physics.

i hope everything goes well for them, and i hope us, as end users see the concept of accelerated phyics come to fruition.

saying otherwise is nonsensical.
 
Some of you actually seem hurt and taken aback by this astonishing :rolleyes: concept. I don't see why. Ageia stirs up a ruckus by trying to create a new market; other companies take notice. Assuming accelerated physics does turn out to be noteworthy, Ageia would stand to make a grab at some pretty serious coin in an existing market as well. e.g. when I build my next puter, i might alot 500 bucks for gaming acceleration - right now, most of that would go to a graphics card. but a year from now, 250 of that might have to go towards a physx card and thus only half of that gets into ati's (or nvidia's to be fair) pocket. It's coming down to a choice - dive into a potential market and share the benefits, or ignore it and lose all around. For the big two, the right choice is obvious, and for that matter, healthy.

Besides, who else is gonna drive down that ridiculous price tag? ;)
 
I don't really care whether it's AGEIA, NVIDIA, or ATI that promotes the PPU (or GPU for physics. I think the solution presented in the article is a great idea, if and only if it gives the customer options. The board from ASUS looks promising as it would allow for two GPUs along with a PPU (or extra GPU for physics), therefore allowing someone who wants the best performance possible to have SLI graphics and physics. The next step down would be one GPU for physics and one for graphics. Finally, gamers on a budget could buy only one GPU and use it partly for physics and partly for grahics.

A standard API is a must before the PPU catches on. I hope that MS releases one in a patch soon after Vista. DXPhysics would ensure future games support and prevent a PPU format war, which is something that no one wants.

AGEIA's solution still looks more economical to me, however. Buying a second or third video card would be more expensive. The only way AGEIA would be more expensive that NVIDIA or ATI is if you can use an old GPU and have on to use. Therefore, I hope that AGEIA at least stays in the market.

What we really want is options people. What we really want is competition between AGEIA, ATI, and NVIDIA, with MS DXPhysics to make sure game are compatible. You shouldn't want AGEIA to die. It is good if they stay in the market and provide more options and competition. I really just want to see options for:

1. SLI GPUs and a PPU - high-end
2. SLI GPUs with one for physics, or one GPU and a PPU - mid-end
3. One GPU, for physics and graphics - low-end.

The high-end can be provided by AGEIA's PPU, the ASUS board, or ATI's Triple Play. The mid-end solution can be provided by AGEIA or NVIDIA. The low-end can be provided by NVIDIA only.

As long as there are options for every performance teir, I'm okay with it. But I don't see why people should want AGEIA to die, therefore limiting competition and reducing the number of options for gamers.
 
drizzt81 said:
I think that the smartest solution would be to wait and see. After all, we are comparing a not-yet released card that may support hardware accelerated physics on not-yet released titles with a released card, which will support hardware accelerated physics on not-yet released titles.

I think I will make up my mind when I get to compare the two solutions.
Click to expand...

I agree with this. Right now, we have:

1. A solution by a huge company with no working product and no game support.
2. A solution by a small company with a working product and limited game support.

Which will win? It's too early to tell. The ideal situation, however, would be that both companies survive, and agree on a common standard set by Microsoft.
 
Very very nice, god i love this forum, keeps me updated all the time lol..
 
For the record the Ageia PPU is flexible as well, allowing the capability to accelerate other APIs if software developers utilize it that way. Both technologies are more open than you think. It all comes down to the software engineers and of course games.
 
mashie said:
The physics on the G80 reminds me of when nVidia introduced the Texture&Lightning engine to the GeForce 256. The first year it was only something that you could use in thech demos, however a few years later when the install base was big enough all games started to require it.

It is the same thing with physics, you can't sell software until you have hardware to run it on and by adding it "for free" the install base will explode in numbers so in due time you will have to have a physics GPU / PPU to run games. It is just a long way to go still.
Click to expand...

The big difference being that T&L was always a standard part of 3d rendering. Physics has practically nothing to do with 3d rendering. Would it be interesting to see an actual dedicated physics core on a video card capable of doing real gameplay physics? Probably. However, that would mean you have to pay more for the video card every time you upgrade.

Personally, i'm not a big fan of the idea of using the video card for physics. For one thing, it's a waste unless you have an extra card that is capable of doing it (otherwise you're paying a lot for DACs, DVI connectors, video decompression capability, etc. that have nothing to do with physics). And there's no chance in hell that I would buy an extra video card just for effects physics. If I'm dropping $hundreds for physics, I want to be able to interact with it (like Cell Factor) not just see fancy explosions. Also with a dedicated and fully capable accelerator, you can probalby keep it for years since the performance should be good for a while. If you buy ATI / nVidia current physics you will have to buy another card later on to get the real gameplay physics.

I'd bet Aegia could make a lot more money if they sold their cards for $150 or so. It can't be that expensive to make, with only 128MB of relatively slow RAM, slow PCI interface, and a chip taht certainly doesn't need a 2-slot cooler. If you can get 100 customers at $300 each, or 1000 customers at $150 each, which is better? And there's no way i'm paying more for their card than I paid for my 7900GT!

I think all of this nonsense will settle down once a standard (DirectX / open physics API). And once that happens the current nVidia / ATI graphics-only physics will probably go the way of the dodo </rant>
 
Just wait for quakewars....It's supposed to have agiea support.


I don't see ageia going away any time soon ether.
 
wera said:
Just wait for quakewars....It's supposed to have agiea support.
Click to expand...
Never heard of this one...

I am strongly of the opinion that it's just another case of "rumours on teh internets" or do you have a link to a reliable source?
 
People seem to be forgetting that IF the G80 has a physics engine on the die it will mean that the gpu will be able to accelerate gfx AND physics at the same time. The great thing about this is you WONT need an extra card for physics acceleration. Plus if you have the cards in SLI you will get both extra gfx horsepower AND increased physics acceleration.

To me that sounds a lot better than a seperate PPU. Admittedly the actual physics performance COULD well be slower than a dedicated PPU, but that will change in time with hardware refreshes and driver updates.

As for the whole one gfx card for gfx and one for physics argument. Its fine if you dont use SLI ( i certainly cant afford it) and besides the SLI seems to be leaning toward single card SLI (2+ cores on one board). Wont be long till we see dual core gpu's out there. A year tops i reckon.

I myself have a crossfire mobo and an X1900XTX, when i get a DX10 card i would love to use my 1900 for physics, assuming my DX10 card will be ATI and not Nvidia. But who knows. I like the G80's rumoured specs and the built in physics hardware is VERY interesing to me.
 
There's only one way this is ever going to come to a sensible end and thats if Microsoft make a decent scaleable API for physics and build it straight into DirectX or make it an extension of DirectX

Then all the different solutions out there, Agiea's physics card, ATI and Nvidia Graphics cards, and whatever else is floating around, can all adopt revisions of their hardware to use it.

There's 2 simple undeniable truths here:

1) There's no way all these companies are going to all agree on using just 1 physics API between them, they all have invested interest in the API they're currently using.

2) Theres no way the consumers are going to buy 3-4 different physics solutions to ensure they can play all their games with advanced physics options enabled.

If Microsoft don't introduce an API then I personally see the video card manufactures winning, theres one large reason behind this in my mind, and thats it's a multifunctioning card.

With Ageia if the game doesnt support their physics then the card sits there and doesn't do anything, with a GPU the card can be enabled to help deal with the extra frame rate and essentially give the user SLI, in this respect all games can make use of the GPU and the user can best decide how they want to enhance the gameplay experiance.

Another problem with effects done that utilise a lot of physics is that this causes extra rendering power, as seen with the effects in GRAW, the movement of the particles is delt with the PPU but the effects still need to be rendered and 1000's of little pieces of rubble need to be drawn which puts strain on the video card. I'm no engineer but I expect theres some fairly clever things that can be done with a GPU rendering particles that it's also calculating the motion of using a physics engine, obviously thats just pure speculation but the engineers for the GPU's are clever people and I'd be supprised if there wasn't some optimisations to be made there.
 
MIcrosoft introducing a physics API will not magically standardize physics in games. It will merely add another, probably free, option for developers to use when making games. Even with DirectX, there is still a not insignificant number of developers using non-DirectX components to do something that DirectX is capable of (opengl, openal, sdl). AFAIK, DirectX is used most prominently because it is one of the easiest packages to develop with - i.e. arguably the best for Windows.

Ignoring open APIs (which I think would be the optimal solution), every hardware developer should be able to develop for HavokFX, since it interacts with the hardware via glsl and hlsl (which is an open spec). So it's not exactly like we don't have an API which couldn't become the standard.

addendum: I should also mention that the reason I think a truly open physics engine would be ideal is to ensure reasonable portability between architectures (i.e. PS3, Wii, Linux, Apple, etc.). I'm sure most of you don't care about that though.
 
Skirrow said:
People seem to be forgetting that IF the G80 has a physics engine on the die it will mean that the gpu will be able to accelerate gfx AND physics at the same time. The great thing about this is you WONT need an extra card for physics acceleration. Plus if you have the cards in SLI you will get both extra gfx horsepower AND increased physics acceleration.
Click to expand...

How do you know that? Their big thing seems to be HavokFX which has nothing to do with gameplay physics. Seems like it would be a big change of direction before even releasing their first option for physics acceleration! They could easily make a small part of the core dedicated to effects physics without working around the problem of getting good 2-way communication between the GPU and the PCI-E bus. Also there's no guarantee that the physics engine on the die will be anywhere near as fast as the PPU even if it is fully-featured.

To me that sounds a lot better than a seperate PPU. Admittedly the actual physics performance COULD well be slower than a dedicated PPU, but that will change in time with hardware refreshes and driver updates.
Click to expand...

So now you can tack an extra $100 or $200 per card on to the cost of upgrading your video to keep up with modern games every 6 months - year, instead of buying a new physics card every few years.

Frosteh said:
Then all the different solutions out there, Agiea's physics card, ATI and Nvidia Graphics cards, and whatever else is floating around, can all adopt revisions of their hardware to use it.
Click to expand...

They wouldn't even need hardware revisions, only driver modifications (and probably small ones at that).

jimmyb said:
Ignoring open APIs (which I think would be the optimal solution), every hardware developer should be able to develop for HavokFX, since it interacts with the hardware via glsl and hlsl (which is an open spec). So it's not exactly like we don't have an API which couldn't become the standard.
Click to expand...

If they don't care about physics that the player can actually interact with, that's great.

Also Havok FX is a commercial product and while I don't know what it costs i'm sure it's expensive. I know Ageia's API is free and I believe they also support at least one open physics API.
 
Skirrow said:
I myself have a crossfire mobo and an X1900XTX, when i get a DX10 card i would love to use my 1900 for physics, assuming my DX10 card will be ATI and not Nvidia. But who knows. I like the G80's rumoured specs and the built in physics hardware is VERY interesing to me.
Click to expand...
and you will get a 1500W (exaggerated number to get my point across) psu to go with that?
 
jimmyb said:
AFAIK HavokFX can do "gameplay" physics.
Click to expand...

http://www.havok.com/content/view/187/77/

Note that they also have a product called "Havok Physics" but it is NOT Havok FX and it is NOT hardware accelerated. Havok FX is strictly effects physics (i.e. smoke, explosion, particle effects, etc.). You can not have a stack of boxes or a torn piece of cloth or a landslide / avalanache / etc. simulated in the game unless it is purely a graphical effect and does not interact with the actual game world.
 
Ah, you appear to be correct from their info page. HOCP has an article which seems to suggest that Havok does accelerate gameplay physics, but Havok's own info page tells otherwise. I'm inclined to agree with you and the link you posted. Very interesting.
 
Pyromaneyakk said:
How do you know that? Their big thing seems to be HavokFX which has nothing to do with gameplay physics. Seems like it would be a big change of direction before even releasing their first option for physics acceleration! They could easily make a small part of the core dedicated to effects physics without working around the problem of getting good 2-way communication between the GPU and the PCI-E bus. Also there's no guarantee that the physics engine on the die will be anywhere near as fast as the PPU even if it is fully-featured.
Click to expand...

HavokFX has everything to do with gameplay physics. ALL physics API's can do both 'effects physics' and 'gameplay physics'. The physics calculations for both are exactly the same.

I have to say though that if the physics engine on the G80 wasnt almost as fast as a PPU then there would be no point in it being there in the first place. Which is why i think this rumour may be wrong. I imagine Nvidia are only stating that the card is physics capable. In the same way that Ati said that their X1000 line of cards can accelerate physics and that it may have been taken out of context.

Pyromaneyakk said:
So now you can tack an extra $100 or $200 per card on to the cost of upgrading your video to keep up with modern games every 6 months - year, instead of buying a new physics card every few years.
Click to expand...

If the rumour is true then i expect the price to be about $50 more than they are now. They just wont sell cards if they hike the price up too high. Prices for the high end cards are too high as it is.

And the only reason we havent seen a new physics card is because 1. there are only a handful of games that support them and 2. Ageia has no competition on the market yet so they dont need to. Once ATI and Nvidia's solutions are available you can bet your ass Ageai will announce a new PPU.

Pyromaneyakk said:
Also Havok FX is a commercial product and while I don't know what it costs i'm sure it's expensive. I know Ageia's API is free and I believe they also support at least one open physics API.
Click to expand...

Yeah HavokFX is commercial and developers have to buy a license. Ageias API only supports the Physx PPU and software. But MS is developing a universal physics API that will support both PPU and GPU as well as in software.

What i'd like to know is if HavokFX will be able to accelerate previous versions of the Havok
engine. That would be cool.

drizzt81 said:
and you will get a 1500W (exaggerated number to get my point across) psu to go with that?
Click to expand...

Yeah probably.
 
3D acceleration never really had much of a chicken/egg to slow the adoption of the hardware, as games took relatively small amounts of coding in order to add a hardware renderer, after which you 'immediately' saw huge benefit in quality and speed.

With physics-only cards, we have a chicken or the egg situation. Physics hardware needs content to be specifically written for it, and gameplay needs to be designed/adjusted in such a way to take account for it. So developers are going to be very slow about sinking a lot of time/money into part of a game to make use of hardware that virtually no one owns.

With GPU-based physics, people are going to already own hardware that can provide some acceleration to physics, so you have less issue with that chicken/egg problem. Within a couple years, a lot of gamers are going to have physics capable GPUs, with quad+ core CPUs, and PPU's adoption won't be a whole lot further along than it is right now. In that situation, what are the developers going to code physics for first? GPUs? CPUs? PPUs?
 
Pyromaneyakk said:
http://www.havok.com/content/view/187/77/

Note that they also have a product called "Havok Physics" but it is NOT Havok FX and it is NOT hardware accelerated. Havok FX is strictly effects physics (i.e. smoke, explosion, particle effects, etc.). You can not have a stack of boxes or a torn piece of cloth or a landslide / avalanache / etc. simulated in the game unless it is purely a graphical effect and does not interact with the actual game world.
Click to expand...

'Havok Physics' is an api used in CG Animation and is the same as the Havok api used in games. The Havok 4 api (which Havok FX is based on AFAIK) is intended for game integration and actually does support hardware acceleration as evidenced in the recent videos that have been linked in other threads. (http://www.hardforum.com/showthread.php?t=1034977 and http://www.hardforum.com/showthread.php?t=1105316)

But having looked at the link you posted to the HavokFX site i'm actually confused. Manufacturers are saying HavokFX is capable of calculating gameplay physics on the GPU, (http://enthusiast.hardocp.com/article.html?art=MTA5NywxLCxoZW50aHVzaWFzdA==) yet the HavokFX website is saying otherwise. As i mentioned in my last post, the physics calculations needed for effects physics and gameplay physics ARE the same, since they follow the exact same rules. So i cannot see why they cant be accelerated by the GPU.

:confused:
 
Skirrow said:
HavokFX has everything to do with gameplay physics. ALL physics API's can do both 'effects physics' and 'gameplay physics'. The physics calculations for both are exactly the same.
Click to expand...
I was under the same impression, but that appears not to be the case.

From the Havok FX product page:
Using Havok FX and GPU technology, game developers can implement a range of physical effects like debris, smoke, and fluids that add immense detail and believability to Havok&#8217;s already unparalleled game-play physics system.
Click to expand...
According to Havok's webpage, it is just effects. This is reiterated further down the page:
Do PC Gamers Need To Purchase Proprietary Physics Hardware To Accelerate Game-Play Physics?

Not for game-play physics. Havok believes that fast game-play physics starts with excellent software technology, and gets faster with increased general CPU computing power &#8211; specifically from multi-core and multi-cell computing architectures. The very clear trend from AMD, Intel, and even IBM going forward is to provide scalable core architectures that can be leveraged by truly multi-threaded software systems. Havok has introduced HydraCore&#8482; technology in its physics system to take advantage of these architectures now and in the very near future. So, for acceleration of game-play physics, along with the core game-logic, special purpose physics hardware is not the answer. Multi-core CPU architectures are the path to speed.

For effects physics, however, the GPU offers the most compelling and sustainable promise to added visual fidelity in games. Relative to proprietary hardware, GPUs also have a clear advantage as a pre-existing technology familiar and readily available to consumers and game developers, providing other benefits such as wide-spread availability, commodity pricing, and mature standards for hardware and software interfaces.
Click to expand...
 
Skirrow said:
'Havok Physics' is an api used in CG Animation and is the same as the Havok api used in games. The Havok 4 api (which Havok FX is based on AFAIK) is intended for game integration and actually does support hardware acceleration as evidenced in the recent videos that have been linked in other threads. (http://www.hardforum.com/showthread.php?t=1034977 and http://www.hardforum.com/showthread.php?t=1105316)

But having looked at the link you posted to the HavokFX site i'm actually confused. Manufacturers are saying HavokFX is capable of calculating gameplay physics on the GPU, (http://enthusiast.hardocp.com/article.html?art=MTA5NywxLCxoZW50aHVzaWFzdA==) yet the HavokFX website is saying otherwise. As i mentioned in my last post, the physics calculations needed for effects physics and gameplay physics ARE the same, since they follow the exact same rules. So i cannot see why they cant be accelerated by the GPU.

:confused:
Click to expand...

The calculations are the same, but ther eis one key difference: For effects physics, the video card can render it and send it directly to the screen. For gameplay physics, the accelerator needs to be able to send the information back to the CPU for processing. Current video cards are not able to do this efficiently, if at all.

So, while the videos may appear to display "gameplay physics", they are probalby not actually having an effect on the gameplay environment. Good demonstration of the performance capabilities of the system, but not evidence of interaction. I think the first nVidia "bricks" demo is evidence of what i'm saying -- listen to the guys talking they say "the ball is controlled by the CPU, the CPU informs the bricks, the bricks interact with each other". I.e., the bricks are simply an effect, and one effect is able to interact with another, however it is unable to interact with the gameplay world (i.e. if the ball continued to bounce around the screen it would not land on top of the bricks but rather go through them since the ball would have no way of determining the shape and size of the pile of bricks).

The effects are nice, but if I have to buy 2 or 3 video cards just to get it, I'd rather buy a PPU, or not have it for now.

On a side note, it seems there is a lot of ignorance / f*nb*yism on this board, especially bashing Ageia. I am by no means a fan of them (you couldn't convince me to buy one of their cards now) but that doesn't mean they don't have a nice piece of equipment. It seems like all of the zealots just bash each other without getting a good understanding of each solution, its advantages and limitations. Ultimately I think physics will become standardized and things will settle down, just like video cards (remember the big TNT2 vs Voodoo3 war with people raging like mad over things like 16 vs 32-bit color, RAM sizes, etc.). The title of this thread I think is a perfect example of this. I don't see this rumor as being any solid proof that nVidia has a solution that completely dominates Ageia's PPU. What it says is that a $600-700 video card might have effects physics built in. It seems that people want the company to fail for some reason.

Well remember this: The more competitors on the market, the lower the prices are for us, and the greater the technology improvements will be. Why don't we just kill off ATI and AMD and just let nVidia / Intel rule the scene? Well if we want to pay $1000 for a mid-range CPU and $2000 for a video card then maybe that's a good idea.
 
Well remember this: The more competitors on the market, the lower the prices are for us, and the greater the technology improvements will be. Why don't we just kill off ATI and AMD and just let nVidia / Intel rule the scene? Well if we want to pay $1000 for a mid-range CPU and $2000 for a video card then maybe that's a good idea.
Click to expand...

QFT

Guys, here's what happens if:

1. AGEIA fails. You only option for more physics is to either buy a new GPU or give up graphics performance. Only two competitors are left on the scene. If you don't keep your old GPU after you upgrade, too bad, you'll have to buy a new one or sacrifice graphics performance. The same applies if you only have one PCI-E16 slot.

2. NVIDIA and ATI fail. Your only option for more physics is to buy a PPU. If you don't have enough slots, you'll need a new mobo, too. Only one competitor is left in the market. If you happen to have last generation's GPU laying around somewhere, you still can't use it.

3. Everyone fails. You notice a slight performance boost as developers utillize your extra CPU core for physics. It isn't much, however, and you don't really notice the improvement. You have an old GPU sitting around and wonder what it would be like if you could use it for physics. If you don't own a dual-core CPU, you are forced to upgrade. When developers decide to start utillizing four cores, you have to upgrade again.

4. Everyone survives. Microsoft comes in and creates a standard in a future Vista patch. If you own an old GPU, you can use it for physics. If you don't want to spend any money, you can sacrifice some graphics performance and use your GPU partly for physics. If you want better performance, you can buy a second GPU and get a boost in graphics and physics performance. If you want the best performanc, you can get SLI along with a PPU, leaving your GPUs completely free to do graphics rendering. The MS standard could also make use of your extra CPU cores (did anyone consider this possibillity?). That way, if you spend $1000 on a quad-core CPU, you don't need to give up as much GPU performance as you would if you only had one core.
 
TekSomniaK said:
is just like if Intel were to develop their CPUs so that video cards are no longer needed.
Click to expand...

Intel already makes it so expansion video cards are not necessary, intel is the largest manufacturer of integrated graphics chips..
 
You must log in or register to reply here.
Back
Top