..don't buy em. They'll be phased out shortly I'm sure. The nice thing is their solution should allow them to enable the additional physics effects when you have an NVIDIA card or the standard effects without. I don't see the big deal for ATI users, as there had already been effects for certain cards that they could take advantage of and NVIDIA cards could not (combination AA/HDR). It's not like ATI/AMD, Intel, and NVIDIA were going to get together to standardize physics in games anyways.

This comment was edited on Feb 5, 10:35.

Now that name takes me back, back to when 3Dfx still ruled and Derek Parez was chief FUD-master at nvidia. The lies this man told, he shamed even politicians with his audacious and voluminous prevarications, lies upon lies upon lies. Believe anything he says at your peril.

The concept of standalone cards purely for physics was a wildly optimistic idea. I suspect they wont survive.

..don't buy em.

Too late! Well, my hope would be that the result of this acquisition is more widespread support for the Ageia Physx library, which would hopefully make that card get more of a workout more often...

Is nVidia released a 9xxx series card with PhysX built in I would pee my pants.

I bought one too! Even though everyone told me not to, I didn't listen. I seen little to no change in any game that uses Ageia Physx. Oh well, it's not the first time I bought useless gaming peripherals... kinda like those 3-d glasses I bought back in '99...

I think nVIDIA has decided that it wants to get physical, physical, nV wanna get physicaaaaal. Maybe they'll get Olivia to host the press conference (topless? nah).

I don't remember how long ago the physx PPU came out but it's been a hella long time in HW years(like dog years but more).

Ageia didn't want to / couldn't make any next gen cards because it would cannibalize their only customers. No doubt advances could have been made to the PPU's, just wasn't economical or possible in their failing business.

I only hope that nvidia will not be afraid to make a next gen PPU that will be a clean break instead of leaving devs trying to stay compatible with the aged HW. A PPU today should be able to blow the doors off the original with what has been learned as well as just normal engineering and tech advances.

Hopefully they can make a clean break with their first implementation, as sad as it may sound to exisitng owners. I wouldn't want to see devs stuck having to support the old card's requirements when something a magnitude more powerful will be available, yet won't be utilized by devs due to them trying to support legacy HW.

It's like devs still having to support Pentium 4's instead of todays C2d's or quads. NVIDIA's first product would be a great time for a clean break. Then devs won't feel like they have to support equivalent of 6800 GPU's, if they start with a PPU equivalent to an 8800 GPU. Then we won't have to wait 5+ years to see what physics can really do, but 2 or 3 years instead with a clean break and uber fast physics HW.

Actually physics is a natural for processing on the CPU, which is exactly what it's fucking there for. A) physics aren't remotely complex enough, like graphics, as to necessitate the need for external processing and B) physics complexity in games will never, ever outpace the CPU's ability to handle it.

If anything we should start buying AI cards, but then developers would have to, you know, start actually coding AIs.

Wow, physics aren't as complex as graphics? You get my vote for most ill-informed post this year.

It is rather funny to see NVIDIA do a 180 from last year when they were saying how AGEIA wasn't necessary to now all the sudden physics necessitates the parallelization techniques which are best suited to a PPU on their GPU's.

Does kind of make me snicker. Devs just have to become more proficient in multithreading and gamers have to get off their ass and buy a decent multicore CPU.

Havok has alot of potential still if Intel doesn't stifle them. We need good API's, SDK's and toolkits for implementing it then things will be golden.

This comment was edited on Feb 5, 19:23.

physics aren't remotely complex enough, like graphics, as to necessitate the need for external processing and B) physics complexity in games will never, ever outpace the CPU's ability to handle it.

Bullshit. Physics are probably more complex than graphics. Rendering transformed polygons, rasterizing to 2d scenes, blending layers and layers of polygons of varying materials is naturally suited to a dedicated GPUs.

Physics are the same way. Games have had to avoid many aspects of physics due to their complexity and inability to process in a real time game. True geometric shattering, fluid dynamics, accurate deformation of models, and countless other aspects of physics remain untouched due to the complexity of modelling them in real time. CPU's are not up to the task of all that. The only thing that keeps current CPUs ability to handle it is by developers careful use of only specific aspects of physics. We've only scratched the surface of physics in games. More realistic collision responses by material types, vehicles, various joint types are but a small portion of what can be done with physics. Even within those domains the implementations in games is simplified quite a bit.

Phyics on the GPU will probably remain as an eye candy focus for a while at least, but it's not impossible to have game effecting physics on a seperate processor(GPU or PPU). Reading back object transforms, or getting collision results back is well within the capability of GPU read operations, as it doesn't approach the complexity(in terms of bandwidth) of reading back textures or procedural meshes, etc. Until there is enough market penetration from this sort of thing developers probably won't use it for gameplay effecting features though. In the end whatever reduces CPU load ultimately gives the rest of the game more headroom. Physics and collision happens to be one of the more expensive aspects of games apart from rendering. The quicker we parallelize that, whether on the GPU or multicore CPUs, the more is opened up for the rest of the game, like AI.

Also regarding AI, as I am a professional AI developer(Pandemic Studios on Mercs2), as well as hobby developer(Omni-bot), I can mention that collision and physics are a large cost of AI, for vision, pathfinding, cover determination, etc... and the cheaper that ultimately gets the better it is for that aspect of AI. If NVidia helps us work towards putting game effecting physics onto the GPU it's a win for developers in general, just as offloading graphics was way back with the introduction of video cards.

Wow, physics aren't as complex as graphics? You get my vote for most ill-informed post this year.

They are not.
The mathematics of Newtonian physics is incredibly simple. A 20 year old solar powered calculator has enough power to calculate the change in velocity, in all three spacial directions, as a function of time, mass, gravity, and air resistance; more then 60 times a second.

Computers are super at basic math like that.
To represent 3D images on the screen is just as simple math on the hardware level, in reality; it just takes a vast large more of those simple calculations.



Alternating Logo (GreaseMonkey script):
http://www.ualr.edu/szsullivan/scripts_/BluesNewslogo.user.js

Physics are probably more complex than graphics.

Sure they are, if you want to create a real to life simulation for scientific modeling. The simple Newtonian mathematics I talked about in my previous post start getting really complicated.

Not for gaming. Things can be super simplified; as you eluded to. It is about representing something on a screen as what we expect to happen. In the calculation of the traction of a car on asphalt, we can represent the surface of the tire as flat. In the real world, depending on how far the wheels are turned, the rotational speed of the tires, the angle of the street surface, the temperature of the tire, etc, all effect the area of tire in contact with the street surface; and thus traction. None of this is required for an accurate representation in a game of the physics of a tire on asphalt.

Where as those simplifications can not be made in graphics. Example, in games where there is dynamic lighting and the light source hits a wall; but there is no reflection from that wall lighting up other parts in the room. That is obvious and looks horrible to us. That next step of multiple reflections from surfaces is needed to represent the game. And a surface that only reflects light at one angle is also really obvious and horrible looking.


Sure physics is complicated, but in a gaming environment, even to have amazing physics, at a mathematical level it is not as resource intensive and graphics.

It is just done on the CPU currently; which has to do tons of other things. Including telling the GPU what in the hell to do. And in many ways (but not in all) modern GPU's are more complex then modern CPU's.






Alternating Logo (GreaseMonkey script):
http://www.ualr.edu/szsullivan/scripts_/BluesNewslogo.user.js


This comment was edited on Feb 5, 21:20.

Oh please. A calculator doing basic physics calculations for an object is nowhere near the complexity of a real time physics engine. Physics engines in games often operate at fixed rates, say 60hz or 100hz where they must detect collisions for large numbers of dynamic objects, and iteratively resolve penetrations between large numbers of independently moving objects, on top of the more advanced functionality of maintaining constraints of joints or other physical modelling of motors, pulleys, etc. The amount of work involved in that is far above punching in some velocity math on a calculator.

I'm not saying there is desire to model every detail of real world physics complexity in a game. I'm saying there is a lot in the area of physics that has yet to be explored in games. Realistic shattering, destruction, fluid dynamics are all heavily physics based and aren't yet feasible in real time. Given how closely graphics and physics are ultimately related in many ways, it also makes it a good fit for the GPU. In any complex physical operation, like fluid modeling, deformation, shattering, etc that games will eventually begin using, there exists a huge amount of graphics information that need to be preserved as part of those processed. Any aspect of the game that results in deformed or broken meshes mean that texture coordinates, normals, etc need to change as well in order to render it properly. Having them both located on the GPU can accelerate that process.

Huge simplifications are made in every facet of graphics. They have been since the beginning and still are today. Any complex feature of graphics is greatly simplified. Shadows, reflections, and even the basic lighting model is still not comparable to realistic lighting. It looks damn good these days no doubt, but with every advancement in hardware there is no shortage of things to put the power to good use on. The same applies to physics.

The GPU being the only real surviving addon card these days, it's really the only feasible option for integration of hardware physics acceleration, on top of simply being a good candidate, given how parallel and math heavy they operate by design.

Point is, that whether it stays as an effect only acceleration, or eventually becomes more it's good news for gamers and developers.

This comment was edited on Feb 5, 23:22.

Physics are the same way. Games have had to avoid many aspects of physics due to their complexity and inability to process in a real time game. True geometric shattering, fluid dynamics, accurate deformation of models, and countless other aspects of physics remain untouched due to the complexity of modelling them in real time. CPU's are not up to the task of all that. The only thing that keeps current CPUs ability to handle it is by developers careful use of only specific aspects of physics. We've only scratched the surface of physics in games. More realistic collision responses by material types, vehicles, various joint types are but a small portion of what can be done with physics. Even within those domains the implementations in games is simplified quite a bit.

You just proved my point. Physics in games equates to basic acceleration do to gravity, minor implementations of friction, and "ragdoll" effects. At the rate at which physics in games is evolving (read: not really evolving) it will be a lllooonnngggg time before substantial, realistic physics implementations are in place.

Of course that could just be because physics in games is relatively new and has yet to really mature. Obviously graphics in games have been able to do that since the days of Zork.

But I still don't foresee a physics implementation with enough muscle to require anything more than maybe another expendable core on my CPU.

a smart move by nVidia but for us consumers it will be at _least_ a year away before we have any reason to buy a new video card with this physics chip on it.

probably will...

how is it anyway ??

did you download warmonger? if it behaves the same with and without the HW please let me know what CPU u have..

Oh please. A calculator doing basic physics calculations for an object is nowhere near the complexity of a real time physics engine. Physics engines in games often operate at fixed rates, say 60hz or 100hz where they must detect collisions for large numbers of dynamic objects, and iteratively resolve penetrations between large numbers of independently moving objects, on top of the more advanced functionality of maintaining constraints of joints or other physical modelling of motors, pulleys, etc. The amount of work involved in that is far above punching in some velocity math on a calculator.

I am sorry. But everything you just said, is mathematically easy and not very computer resource intensive.

What you just said, the raw numbers can be crunched at 100 times a second on a 10 year old computer with proper algorithms. That is one of the things engineers starting using PC's for years ago.

Trying to display that raw data as graphics is where the problem comes in. The difference is the engineers would get the raw data and they would the analyze the raw numbers. Now display it as a real world representation on the screen.




Alternating Logo (GreaseMonkey script):
http://www.ualr.edu/szsullivan/scripts_/BluesNewslogo.user.js


This comment was edited on Feb 6, 19:24.