NVIDIA Acquires AGEIA

eric66

Gawd
Joined
Sep 16, 2007
Messages
920
'"NVIDIA is the perfect fit for us. They have the world's best parallel computing technology and are the thought leaders in GPUs and gaming. We are united by a common culture based on a passion for innovating and driving the consumer experience," said Manju Hegde, co-founder and CEO of AGEIA.'

http://www.nvidia.com/object/io_1202161567170.html
 
How much do you bet that all the people who bought PPUs find themselves looking at complete obsolescence within a few months? Yay $200 for a tech demo! :rolleyes:
 
How much do you bet that all the people who bought PPUs find themselves looking at complete obsolescence within a few months? Yay $200 for a tech demo! :rolleyes:

I could have told you that over a year ago.
 
Maybe the separate PPU add in card is dead, but maybe GPU physics gets moving. I don't care what device is rendering physics, as long as I can shoot a rocket into a building and it blows up and falls down accordingly.
 
Maybe the separate PPU add in card is dead, but maybe GPU physics gets moving. I don't care what device is rendering physics, as long as I can shoot a rocket into a building and it blows up and falls down accordingly.

Yeah, I'd like to see GPU physics, although I really think software physics will be there by the time we see real support for GPU physics. Intel will have 8-core processors out before we see a GPU physics engine.

You'd better hope that Nvidia makes it an open standard, or ATI will cry foul and find their own GPU physics engine to promote. Can you say standards war?

And since the market is basically owned by Havok, they would both be clawing their way from the ground-up. A standards war never makes this easy.
 
You'd better hope that Nvidia makes it an open standard, or ATI will cry foul and find their own GPU physics engine to promote. Can you say standards war?

For some reason I doubt Nvidia will make it a open standard. I also don't think they'll give a shit about ATI crying foul.
 
For some reason I doubt Nvidia will make it a open standard. I also don't think they'll give a shit about ATI crying foul.
Right, ATI's not gonna cry foul, but AMD probably will. They've got a long teary history of crying over Intel performing better than them for one reason or another for years.
 
How much do you bet that all the people who bought PPUs find themselves looking at complete obsolescence within a few months? Yay $200 for a tech demo! :rolleyes:

Wow! Someone actually bought one of these? :rolleyes:
 
I would like to see an open physics API stanard, possibly built into the DirectX standard, but what says, microsoft, intel, or nvidia are willing to play ball for even a second.
 
Is this the PPU / Physics Processors forum?
If not litterally and techniacally you are wrong :p

But it is now a subsidiarity of nvidia, which is in the realm of "video cards" which this forum pretains too. I think the separate PPU addon card is dead, but the PPU will most likely move either on the GPU as another instruction or as an add in chip to the board itself working in conjunction with the GPU.
 
But it is now a subsidiarity of nvidia, which is in the realm of "video cards" which this forum pretains too. I think the separate PPU addon card is dead, but the PPU will most likely move either on the GPU as another instruction or as an add in chip to the board itself working in conjunction with the GPU.

When such a product comes, perhaps you will be right,..but I relate to reality..and in the real world the PPU is not a GPU..nor have NVIDIA said ANYTHING about their palns..so this is the wrong forum...we are talking PPU/Physics...not GPU/graphics.
 
Question. The Physics Processor, While a good idea, was it only TWO games that really took advantage of it? and were both of these Ghost Recon? If this is the case why is NVidia Purchasing them? or is it mainly to incorporate the features into future Ge-Forces?


I think I may have answered my own question.
 
Question. The Physics Processor, While a good idea, was it only TWO games that really took advantage of it? and were both of these Ghost Recon? If this is the case why is NVidia Purchasing them? or is it mainly to incorporate the features into future Ge-Forces?


I think I may have answered my own question.

Don't forget UT3.
 
Yea, I was thinking about buying one, good thing I didn't. If they could incorporate such technology in a regular GPU, now that would be cool, sure would give Nvidia another edge in gaming market.
 
Wasn't there a setting back in the day... for I think it was ATI cards in crossfire.. or even NOT in crossfire where one could be set to a dedicated physics card? Didn't it process ageia code?

With all of the programmable processors on the nvidia GPU's (8000 series.) Wouldn't it be feasible to have a soft patch that simply lets you use a portion of your processors for physics and the rest for graphics?

Especially in a TRI SLI setup.
 
I never did follow up closely on PPU so one thing never came clear to me. Which one did suck for Ageia, their drivers or hardware? Or maybe both. Good luck to Nvidia, they have some work to do :p
 
But it is now a subsidiarity of nvidia, which is in the realm of "video cards" which this forum pretains too. I think the separate PPU addon card is dead, but the PPU will most likely move either on the GPU as another instruction or as an add in chip to the board itself working in conjunction with the GPU.

So we should post about 780i mobo here too? :rolleyes:
 
This move was coming for awhile, its a great move for Ageia since they has been earning for this kind of market potiental for awhile now that there onboard I believe we will start seeing second generation PhysX Chips right on-board the next "G"en Large Nvidia Flagship G100 Anyone???.

Also it will be interesting to see if this will now push Game makers to actually support PhysX Technology.
 
This move was coming for awhile, its a great move for Ageia since they has been earning for this kind of market potiental for awhile now that there onboard I believe we will start seeing second generation PhysX Chips right on-board the next "G"en Large Nvidia Flagship G100 Anyone???.

Also it will be interesting to see if this will now push Game makers to actually support PhysX Technology.

If it did, it would certainly give nVidia an edge over ATI, as their GPU physics tech is nowhere to be seen in the real world. At least nVidia would have a head start with support/games.
 

But that was made for Havok FX, which was terminated when Intel bought Havok. If anyone wants to know what GPU pysics dried-up, it's because ATI and Nvidia were waiting on the release of Havok FX.

If we're going to use Havok as THE physics standard, it's going to be on general-purpose CPUs only. And Intel is just fine with that.

Again, this is why I think Nvidia may move to make PhysX an open-standard:

1. PhysX has a software physics engine that competes with Havok
2. PhysX is VERY far behind Havok in terms of game uptake, even for games that just use the software PhysX engine.
3. A situation where only one of the two top GPU makers supporting PhysX will not help their goal of market domination.

If Nvidia wants to take Intel and Havok out of the picture and really make a step for GPU physics, they need to open the standard right now.
 
Reading their press release

"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."

I am thinking either a on card physx solution, or possibly and more likely a rebranded physx card.....
 
Also it will be interesting to see if this will now push Game makers to actually support PhysX Technology.

Certainly more so now, also AGEIA had this stupid habit of only having PCI classic available even after stubborn Creative broke down an made pcie cards widely available.. Anyone who would be building a high end computer would almost certainly already have plans for the one or two left on motherboards these days(even if that only meant having one blocked by dual slot GPU cooling :p ).
 
Reading their press release ...
I am thinking either a on card physx solution, or possibly and more likely a rebranded physx card.....

I read it quite differently.

The Nvidia press release states "we can now bring GeForce®-accelerated PhysX to hundreds of millions of gamers around the world", and "Applications such as physics, ... are enabled through CUDA ..."

To me this is a pretty clear statement of intent to draw on their CUDA expertise to port/merge the PhysX SDK into the GeForce drivers -- i.e. it becomes a driver extension like PureVideo is. I just don't see them selling a significant number of dedicated cards.

And just as PureVideo can be done partly in software and partly in hardware (think 8800GTX with VP1 vs 8600 with VP2) I wouldn't be at all surprised if in a year or so they implement the PhysX hardware within a new GPU -- after all, the PhysX PPU is only 125M transistors (and that includes the I/O logic).

Actually this could fit well with product development cycles -- release the software on the existing platform thus enabling game development now, by the time they are out in the market (running on users existing, non-hardware-accelerated video cards), the next generation of hardware will be ready to go onstream.

... AGEIA had this stupid habit of only having PCI classic available ...

I read that the initial PhysX design was PCIe-x1 but there were compatibility issues, so Ageia went back to PCI.
 
Question. The Physics Processor, While a good idea, was it only TWO games that really took advantage of it?

Nope, go read up.

...and were both of these Ghost Recon?
No, wrong again.

If this is the case why is NVidia Purchasing them?
Because AGEIA has a working, dedicated hardware physics solution plus software and a API.
And because Intel bought Havok, killing SLI/Crossfire physics....if it ever was more that a PR stunt.

...or is it mainly to incorporate the features into future Ge-Forces?
First sensible thing you have said.


I think I may have answered my own question.
But you post is still full of flaws.
 
But that was made for Havok FX, which was terminated when Intel bought Havok. If anyone wants to know what GPU pysics dried-up, it's because ATI and Nvidia were waiting on the release of Havok FX.

If we're going to use Havok as THE physics standard, it's going to be on general-purpose CPUs only. And Intel is just fine with that.

Again, this is why I think Nvidia may move to make PhysX an open-standard:

1. PhysX has a software physics engine that competes with Havok
2. PhysX is VERY far behind Havok in terms of game uptake, even for games that just use the software PhysX engine.
3. A situation where only one of the two top GPU makers supporting PhysX will not help their goal of market domination.

If Nvidia wants to take Intel and Havok out of the picture and really make a step for GPU physics, they need to open the standard right now.
so current in game physics are done by cpu and that quantum physics thing that came with 8000 series is unused
 
so current in game physics are done by cpu and that quantum physics thing that came with 8000 series is unused

Unless you own a PPu and play a game supporting the PhysX API..then you have next-gen gaming...today ;)
 
Unless you own a PPu and play a game supporting the PhysX API..then you have next-gen gaming...today ;)

But I'm sure he doesn't (nobody does) and I'm sure he's not (there's barely a single worthwhile game that supports PhysX). So, I'd assume next-gen gaming is simply "that" far outside his reach ;).
 
I would like to see an open physics API stanard, possibly built into the DirectX standard, but what says, microsoft, intel, or nvidia are willing to play ball for even a second.

There has to be really, and I thought there was some whispered talk of a DirectX physics api being in progress behind the scenes. We're not going to see any serious adoption of hardware-accelerated physics until there's a standard, hardware-independent API, it doesn't matter who owns what.
 
Back
Top