NVIDIA to acquire AGEIA

[+BeLGaRaTh Subscriber¹]

US, February 4, 2008 - NVIDIA sent out word today that it has signed an agreement to acquire AGEIA Technologies. The move puts AGEIA's PhysX software, a widely used SDK for gaming phsyics, under NVIDIA's belt. According the press release, more than 140 games that use PhysX are either shipping or in development for PS3, Xbox 360, Wii and PC. The acquisition will allow NVIDIA to introduce an integrated GPU and physics card into the market, giving them a leg up on the competition.

"The AGEIA team is world class, and is passionate about the same thing we are?creating the most amazing and captivating game experiences," stated Jen-Hsun Huang, president and CEO of NVIDIA. "By combining the teams that created the world's most pervasive GPU and physics engine brands, we can now bring GeForce?-accelerated PhysX to hundreds of millions of gamers around the world."

[SOURCE]

[Yusuf M. Veteran]

GeForce accelerated PhysX? Interesting. I wonder how AMD-ATi will react.

[+BeLGaRaTh Subscriber¹]

From what I recall AMD-ATi tried to purchase them late last year for the exact same thing, an integrated card, they must be mighty PO'd

[Raa]

Which is interesting, because AGEIA denied this rumor last month!

Good to see though, should be fun :D

When can I have my Geforce 9800 PPU? :p

[rajputwarrior]

i still think PPU are useless, just get a core of a cpu to do it...

[jimbo11883]

Try getting a core of a CPU to do what a GPU does... Instead, a PPU works like a GPU works, they are finely tuned to their tasks. Throwing both GPU and a PPU on a card will re-define what the card is... instead of Graphics Processing Unit, it could be Game Processing Unit. :p

[Berserk87]

damn, guess ill put off upgrading till series 10 comes out now :p

awesome move for nvidia, i wanna see what they do with it.

[Andre]

i still think PPU are useless, just get a core of a cpu to do it...

Exactly. And the AMD Fusion project will most likely just do that.

[Ayepecks]

Hey, maybe now it's possible their hardware will finally be adopted. :laugh:

[giantpotato]

Isn't ATi already doing physics processing?, but only in Crossfire. http://ati.amd.com/technology/crossfire/physics/index.html

[Thrawn]

Cool, we'll have a GeForce 10800 GTX w/PPU.

[NienorGT]

Great, PhysX libairies where starting to well be in use, and now, Nvidia cards will have it.

I'm not sure if nvidia will keep producing PPU cards, I bet they will make SLI Physic finally a reality.

[null_]

Great, PhysX libairies where starting to well be in use, and now, Nvidia cards will have it.

I'm not sure if nvidia will keep producing PPU cards, I bet they will make SLI Physic finally a reality.

Hell, they could possibly add the PPU chip to the PCB board of the graphics card or build it straight into a single chip, who knows with NVIDIA.

[Smigit]

i still think PPU are useless, just get a core of a cpu to do it...

The same arguement could be made for not having Video cards at all and getting the CPU to do the processing. In practice processors that are dedicated to a sole task and optimised to perform it will perform alot better than a general processor designed for a multitude of different task.

[NienorGT]

Hell, they could possibly add the PPU chip to the PCB board of the graphics card or build it straight into a single chip, who knows with NVIDIA.

Probably useless, well I hope they wont, since GPUs can almost do CPU stuff now (IE: Folding@Home) with GPGPU stuff, I think it's already not so hard to tweak the Ageia Physic driver to use the GPU to do the thing other than the PPU.

We all know that the nice thing with Ageia PhysX API, is that it work also on X86 CPUs.

Anyway, with the GT200 (Next GPU after the G100/D9E) having a rumored 250W TDP, it's not the time to put another little chip with a TDP of 30W on it lol. I REALLY doubt it's possible to cool 280W (+ RAM) on a Dual Slot Air cooler :s)

Edit: Here the nVidia CUDA thing: http://en.wikipedia.org/wiki/CUDA

Basically, I see it as running C code on the GPU...

Whatever, I may be a big Hardware fan, but all that API stuff is out pacing me :s

[spaceelf]

I hope they do keep it a seperate chip and license it out to the other card companies.

[revvo]

GeForce accelerated PhysX? Interesting. I wonder how AMD-ATi will react.

They probably won't react.

Unless Nvidia combines both the physics chip and the geforce gpu's on a single card, and eventually they will, ATI/AMD will sit where they are. How many people care about the physics card, seriously out there how many games make use of this? And are the benefits worth the purchase? If Nvidia pushes this by putting its features into upcoming Geforces, that could change a thing or two but as long as its is own seperate card, it won't do squat.

[Smigit]

Unless Nvidia combines both the physics chip and the geforce gpu's on a single card, and eventually they will, ATI/AMD will sit where they are. How many people care about the physics card, seriously out there how many games make use of this?

Your right when you say noone really cares but NVidia aren't stupid either and probably realise theres not a huge market for these cards as individual items so I'd really expect them to be pushing for the integration into their graphics cards in the future. I doubt NVidia bought them with the intention of selling millions of stand alone cards.

[The_Decryptor Veteran]

Intel has Havok (and HavokFX, which runs on the GPU) and now Nvidia have AGEIA, I wonder if ATI will follow suit.

[Smigit]

Intel has Havok (and HavokFX, which runs on the GPU) and now Nvidia have AGEIA, I wonder if ATI will follow suit.

NVidia may possibly license out the technology which would benefit them and possibly ATI by having a uniform standard between the two manufacturers for programers to program to, making the cards more likely to be utilised.

[The_Decryptor Veteran]

Yeah, a unified way to access it is really going to benefit adoption.

ATI and Nvidia do work together on making things work, if they both care about Physics (they have both stated that) then they'll hopefully work on it (and hopefully Intel will join the party if they do)

If they do all chose a method, I'm hoping they document how to do it, not every physics library is closed source.