Ageia's PPU's

[akugami]

I think besides heavy FPS and action games, the biggest benefit would be to MMORPG's where interaction with the environment would bring a whole new level of submersion. Think about it, today's MMORPG's have very very limited interaction with one's environment.

 

[nitromullet]

Originally posted by: John Reynolds
Yea, but with that perspective how would you explain the success of the Voodoo 1 back in '96? It too was essentially a one-trick pony that was useful only for entertainment purposes (games).

I agree, though, that AGEIA has a tough hill to climb.

I think that about sums it up, and is clearly evident by all the people claiming they would be glad to see AGEIA fail. I'm certainly not one of those people. I'm quite interested in this tech actually.

 

[tuteja1986]

was asus suppose to launch one arround this time :?

 

[biostud]

Originally posted by: tuteja1986
was asus suppose to launch one arround this time :?

apperently nothing will be launched until the AAA titles taking advantage of it are launched, which makes pretty good sense IMHO. This way they avoid buyers complaining about not beeing able to use their new hardware and giving them a negative public start.

 

[NoStateofMind]

I see this PPU as a major step forward in games. I think it will change gaming for a much more "in the game " feel.

That being said, another person posted about it being integrated on the mobo, this seems very logical to me. I think motherboards with this PPU would be in high demand.

Previoiusly I mentioned the demo for the physics engine and it's phenominal at showing what it can do. It shows two dual core processors being used in the test. The first test showed that without the PPU, it used 100% of 3 cores for physics, and about 25% of the fourth core for rendering. In the second test they used the PPU with the four cores, the result was staggaring to me. Only used one out of four cores at 25% for rendering only. Now think about how much it will help future gaming realizm. I'm definatly excited. cheers!

 

[gac009]

hmmm, if they could somehow intergrate it onto the actual videocard itself.....

 

[Jeff7181]

Originally posted by: munky
Of course it needs developers to explicitly code for it. Otherwise how would you know what code to unload to the ppu and what not to unload. It will probably need it's own API, or possibly become incorporated into future versions of DX if it gains popularity. The problem is there's not much demand for the thing, and rightfully so. It's a game we're talking about, not some mathematical modeling of a chain reaction. You can just as easily unload a lot of the calculations on the gpu because in current games the vertex processors are just sitting twiddling their thumbs half the time, waiting for the pixel processors to finish rendering the frame. Add multi-core cpu's to the equation, and I hardly see a reason to spend $200+ on additional physics HW.

Not from what I've read. I've read to make use of Ageia's PPU the game only needs to use their PhysX physics engine. Of course it would need drivers... what piece of hardware DOESN'T need drivers?

*EDIT* Also, your opinion about the GPU and CPU being adequate for handling the physics is based on current physics engines and technology. If you throw 10 times the physics calculations at the GPU and CPU, things are going to slow down. I think the most attractive thing about the PPU wasn't it's ability to create a destructable environment, but it's potential for better fluid simulation. Rather than having puffs of transparent textures and shader effects for smoke and water, smoke particles and water drops could be actual objects, not just shader effects and transparent textures.

 

[apoppin]

from an old post:

Originally posted by: apoppin
r520 can already replace the PPU

AEGIA might have a run for it's money if ATI has anything to say about it. One AIB commented today that the idea of a dedicated scalar mathematics processor for game physics could already be replicated on ATI's R520 series silicon, although drivers for such a project only exist in R+D departments (the vendor wouldn't let us have them, we tried). The idea of offloading math to a GPU is not a new idea; many projects exist for Linux for this already. However, the indication we had was that ATI could actually do physics calculations on the card with the graphics processing simultaniously -- the bandwidth is already there. AEGIA's physics processor has already been delayed well into Q2 next year.

:Q

competition
:thumbsup:

UPDATE:

Havok to compete with AGEIA for physics

many have heard of AGEIA and its startling announcement: they will produce a processor used exclusively to process physics related computations. Called the PPU, or Physics Processing Unit, its role will be to offload highly intensive mathematics such as realistic water movement, realistic character physical reactions to objects and the world, from the CPU to a dedicated processor. This all seems like the natural progression of things, since dedicated sound, network and other processors are commonplace.

Today, however, most processors spend their time mostly idling - you're rarely ever pushing your hardware to its limits consistently. Thus Havok, a company that's well known to game developers, has announced that it has plans to do for you what AGEIA promises, but save you money and maximize your dollar spent at the same time. Indeed, Havok has confirmed with us that they are competing with AGEIA.

The Havok FX engine is what Havok claims will provide the functionality of a PPU, but its approach is entirely different than AGEIA's. What's special about Havok FX is that it's a software engine that is currently based on Havok's widely used physics engines. However, Havok FX is designed to offload many intensive physics functions from the CPU to the GPU. Using technology available in Shader Model 3.0 and beyond, the Havok FX engine will be able to take advantage of unused resources of today's powerful GPU's and put them to use.

[/quote]

 

[rise]

i don't get it (surprise, lol). what unused gpu resources are they talking about? if games are gpu limited how are there extra resources lying around?

 

[Munky]

Originally posted by: Jeff7181

Originally posted by: munky
Of course it needs developers to explicitly code for it. Otherwise how would you know what code to unload to the ppu and what not to unload. It will probably need it's own API, or possibly become incorporated into future versions of DX if it gains popularity. The problem is there's not much demand for the thing, and rightfully so. It's a game we're talking about, not some mathematical modeling of a chain reaction. You can just as easily unload a lot of the calculations on the gpu because in current games the vertex processors are just sitting twiddling their thumbs half the time, waiting for the pixel processors to finish rendering the frame. Add multi-core cpu's to the equation, and I hardly see a reason to spend $200+ on additional physics HW.

Not from what I've read. I've read to make use of Ageia's PPU the game only needs to use their PhysX physics engine. Of course it would need drivers... what piece of hardware DOESN'T need drivers?

*EDIT* Also, your opinion about the GPU and CPU being adequate for handling the physics is based on current physics engines and technology. If you throw 10 times the physics calculations at the GPU and CPU, things are going to slow down. I think the most attractive thing about the PPU wasn't it's ability to create a destructable environment, but it's potential for better fluid simulation. Rather than having puffs of transparent textures and shader effects for smoke and water, smoke particles and water drops could be actual objects, not just shader effects and transparent textures.

As far as having more accurate physics modeling, that might be the case in the future, but I dont see it happening in the near future (like this year). Current games always make trade offs using approximations so that the effect looks real enough but not actually being modeled in full fledged reality. For example, things like normal maps, shadow maps, and transparent textures are often used to approximate the appearance of much higher detail objects, surfaces, and lighting. I think in 2-3 years when the hardware becomes much faster the developers will incorporate increasingly more detail into the scenes, and possibly these approximation models will be replaced by full detail models and effects. Similarly, the physics will likely be much more accurately modeled, and then we might see dedicated physics HW gain popularity, or it might be integrated into other components like the mobo, the gfx card ot the cpu.

 

[Munky]

Originally posted by: rise4310
i don't get it (surprise, lol). what unused gpu resources are they talking about? if games are gpu limited how are there extra resources lying around?

Vertex shaders are rarely stressed in any modern games. But then again, the future cards will have unified shaders, so that should provide even more flexibility on the gpu platform but at the same time those unified shaders might all be too busy rendering the graphics and not have enough slack to do physics as well.

 

[the Chase]

It also depends on just how far a certain game wants to go with physics in the game. If they want super realistic water but stop short of wanting/needing every other item destructible by the player then a Havoc/software/superGPU/4 core CPU or combination of these would probably do it. If they want the whole ball of wax I think the GPU's/Havoc's claims/dual cores would slow to a crawl. It seems Ageia is so far the only player to put enough time and development work into the area to produce the results in their samples we have seen. I do like the idea of incorperating it onto the motherboard or graphics card. It may be the only way to make it fly is by licensing the tech to ATI/NVIDIA, or whomever.

 

[PingSpike]

I don't know...with dual cores out we have a lot of general processing power thats still largely untapped all of a sudden. I think they are introducing this idea at the wrong time. Are people going to buy a $200-300 graphics card (most people aren't buying SLI GTXs) that they use in all their games, a dual core processor and then a $200-300 physics card that a handful of games support?

 

[dug777]

TBH i think they are dead in the water, as has been said, a lot of their need has been taken by dual cores for the immediate moment, and anyone remember the havok engine and it's future use on both ATI and NV cards? IMO with that option people will be very unkeen on a very expensive discreet solution.

 

[rise]

i dunno, even with the games patched for dual cores you don't see that kind of improvement and thats not adding anything to the enviornment or gameplay.

the cpu obviously is practically unused in gaming as i can run 2x rosetta while running any game out. but it doesn't seem it can do the same work these ppus can/will.

otoh, it will be able to crunch for me when not gaming

if it sells like a sound card does, $75-150 then i'm in. if its up around $300 then it'd have to be unbelievable for me to consider it and i would hope it would allow me to run a low-mid range card with it.

 

[rise]

Originally posted by: John Reynolds
I've said this before, but what AGEIA needs at the time of hardware launch is a AAA title that really shows an appreciable benefit from supporting the PPU, something that goes way beyond increased frame rate from the physics code being hardware rather than software accelerated. But what that title could be is a mystery to me: UT2007? Warhammer Online? It's gotta be a high profile title and there needs to be a really well thought out marketing push behind it.

oblivion on my mind bump

i kinda thought oblivion would be one of their AAA titles. it made even more sense when oblivion and the ppu wee both delayed late last year.

now the developers are cutting back on oblivion shadows, specifically interactive objects shadows. needless to say, i'm no graphic guru so my question is, would the use of a ppu have allowed them to leave the hdr shadows that they were showing off in the demo?

 

[Mocib]

The Elder Scrolls 4: Oblivion uses a current-gen version of the Gamebryo 3D game engine that uses Havok for physics. The next-gen version of Gamebryo 3D will use AGEIA PhysX SDK, though: http://www.emergentgametech.com/index.php?source=20060117

Gothic 3, on the other hand, uses AGEIA PhysX SDK and will support the PPU with a patch after the game is released (end of 2006): http://forum.jowood.de/showthread.php?p=1694472#post1694472

 

[otispunkmeyer]

i wonder if they could use this in engineering?

fluid dynamics and models of fluid flow etc are very taxing, and it shure would be nice to have someway of speeding these models up. at our uni they use the network computters in a folding @ home style set up to crunch numbers for big theoretical model calculations etc....would be awesome to do it on one machine

for $300 though it would have to dance on my desk every now an then.

 

[otispunkmeyer]

also put it this way

with SLI, an Audigy 2 and a digital TV tuner there is simply no more room for another AIB. this is promising tech, but i think they need a better way of getting it integrated into peoples systems.

 

[akugami]

I'm of the opinion that Ageia may ultimately fail BUT nVidia and ATI will take up where they left off with more simpler physics routines initially that adds a little bit to physics processing. This will in turn become more complex as time goes by. After all, we're getting close to the limit on how far a tradtional GPU can get better without it becoming a bit ho-hum.

To me, current graphics cards are nearing the limit on how desirable the new features will be in another year or two. I mean, more polygons, whee, faster shader processing, whee, higher resolutions, whee, more graphical effects, whee, higher frame rates, whee. What's after that? What new features are going to make it compelling to purchase new graphics cards? I mean, once you get past a certain level of graphics effects and can get a steady 60fps+ on it, then what? I can easily see a 2560x1600 resolution at 60fps being done by graphics cards in 2 years with all the features you see on current cards and then some. I feel most graphical improvements after that will be incremental and not compelling enough for most gamers to upgrade until full photo realism in 3D is achieved.

I agree that more realistic physics will be the next big thing. I'm just not sure a $200-300 add-on card will be that. It seems much more realistic that ATI and nVidia (or maybe Matrox rising back from the grave) will be able to do that. Imagine a FPS where you could literally blow holes through walls to make a new entranceway if you don't feel like walking in the front door and you got a rocket launcher. Grenades and other explosions would make holes in the ground, allowing you to make your own trench. Blowing up small buildings and killing everyone inside as the roof collapses instead of having to shoot everyone individually. All in a realistic manner. The biggest boon could probably be to MMORPG's with superb world realism, though they'd have to limit how much you can impact the environment.

 

[PingSpike]

Originally posted by: rise4310

Originally posted by: John Reynolds
I've said this before, but what AGEIA needs at the time of hardware launch is a AAA title that really shows an appreciable benefit from supporting the PPU, something that goes way beyond increased frame rate from the physics code being hardware rather than software accelerated. But what that title could be is a mystery to me: UT2007? Warhammer Online? It's gotta be a high profile title and there needs to be a really well thought out marketing push behind it.

oblivion on my mind bump

i kinda thought oblivion would be one of their AAA titles. it made even more sense when oblivion and the ppu wee both delayed late last year.

now the developers are cutting back on oblivion shadows, specifically interactive objects shadows. needless to say, i'm no graphic guru so my question is, would the use of a ppu have allowed them to leave the hdr shadows that they were showing off in the demo?

Probably not. I would imagine the shadows were almost entirely a videocard limited affair. It seems like the items are still interactive, which should be handled by the cpu anyway...they just won't all be casting shadows all over the damn place anymore. But thats just a guess.

 

[VERTIGGO]

The thing is, when video cards were conceived, people ran around saying there was no reason we needed a replaceable card; IGPs were plenty capable of anything imagineable. It's the same with everything they come up with. "Who needs more than 2 speakers? Stereo sounds great!" It's going to happen again, but only when the software available will make them money by giving us a reason to buy a PPU. (Eventually games will have physics options as extensive as current graphics options) Of course no one wants to spend more money, but we still have a choice. Nothing is free. Personally, I would be glad to shell out a few hundred, to see thousands of bone fragments fly and hair that is teased by the wind like Aslan's mane in his recent flick. Anyone who won't spend money on it, can just pretend they don't care how realistic their games look.

 

[PingSpike]

Well, the comparison to 3d accelerators is valid and it isn't. Its the same kind of idea, a seperate focused device that will take the load off the processor and do it better at the same time. But this isn't a graphics card. And it isn't 1996. There's a very different climate out there now.

I remember when my friend came over and played dark force II: jedi knight on my PC with a 3dfx voodoo and my parents with integrated graphics. He said "The graphics are good on that other one...but they're AMAZING on yours." Now JK looks pretty horrible these days, but he was right. 256 colors at 320x240 and bad framerates versus 16bit at 640x480 and smooth framerates. It was night and day. I grumbled about having another expensive component to buy, but it certainly offered a very tangible benefit. And it wasn't long before 3d accelerators weren't even an option anymore. The 3d accelerator was probably the biggest thing to happen to PC gaming, I'd say in its entire history. Only the sound card would come close IMO.

Is the physics card going to provide this same night and day difference? I don't know. But before you answer consider that even with the 3d accelerator there WAS a transition period. Everyone here talks about exploding buildings that crumble, endless 3d objects bouncing all over the place, etc. But if this thing wants to be accepted games can't REQUIRE it, at least not right away. So initially we're going to have to have software physics tandem along with it, and its going to cater mostly to the lowest common denominator.

Whether people are going to see the difference between relatively simple physics being dumped onto the now readily available second CPU core and the presumably superficial enhancements offered with the PPU remains to be seen. I say superficial because they probably wouldn't want to break the game for non-PPU folks or lengthen already very long development times by changing the game signifigantly with enchanced physics, so we can probably assume that we'll see the extra PPU power used on boring extra things. You're not going to see 25 boulders instead of 3 rolling at you with the PPU. You'll probably just see some pebbles that wouldn't make any difference in the completion of the level. This is just a guess based on past precidents though.

Also of note...there's no directPhyiscs or OpenPL to program for to my knowledge. Part of the reason 3d accelerators took off is was it was easy enough for developers to put them to use...indeed it became easier TO use them to not use them. But then again, ageia could just be the "Glide" for PPUs.

I don't know. I could see the PPU being useful. But I'm not sure it was introduced at the right time to be successful. Developers have been happily shoveling the workload onto the GPU for a few years now since CPU speeds have lost their speed of growth. Now they have double the speed, with some real work being required to harness it. Is there really room for another processor at the developers table? It looks like we're seeing both the developers and Ageia waiting for the other to flinch right now. There's a few games said to use it on the horizon, but the part isn't even out there.

Time will tell though. It might only take a few killer apps to get the market momentum going. I think price will play a big part. If it costs between $50-100 and there's a handful of games that really benefit from it I could see it getting a solid install base and taking off from there. If it costs $200-300 or something I think its completely doomed unless they can trick microsoft into slipping it into the vista system requirements or something.

 

[Raduque]

All I have to say about this thing is the tech demos look neat, but at $300, I'm not buying a ppu that costs more then my video card (6800 agp, $150)

 

[SheepMan]

I agree that $300 is alot, but I think tht once games start to use the PPU having a completely interactive envrionment, and playing in enormous battles (maybe online?) in the tens of thousands :shocked: (from what I've read) might be just worth it.

http://hardware.gamespot.com/Story-ST-17585-1998-x-x-x