AGEIA Can USE MORE THAN ONE CARD?

[CaptNKILL]

here's what's next... me and the Infinium boys have been talking... and, well, they gave me what they call the "secret jpeg". I swore on my life that I wouldn't show the world... and so I thought about it, and ya know, this place isn't the whole world... so:















http://www.nvnews.net/vbulletin/attachment.php?attachmentid=21156&stc=1&d=1161047883


:p

gaaaahahahahaha... (pirate) (pirate) (pirate)

[superklye]

Man, if they keep up with SLI (Oct-SLI) and PhysX cards can be paired...we're going to need bigger mobos.

[Tikiman44]

Man, if they keep up with SLI (Oct-SLI) and PhysX cards can be paired...we're going to need bigger mobos.

And bigger cases.:captnkill:

[SH64]

PhysX SLI ?? NO :wtf:

[Tikiman44]

PhysX SLI ?? NO :wtf:

My thoughts exactly. "Coming to a store near you Physx SLI:retard:!"

(crazy)

[jAkUp]

Yea I think this is a bad decision. I could see maybe 2 PPU's on 1 PCB, but asking for another PCI slot in an enthusiasts rig just aint gonna happen.

[Tikiman44]

Yea I think this is a bad decision. I could see maybe 2 PPU's on 1 PCB, but asking for another PCI slot in an enthusiasts rig just aint gonna happen.

The man speaks the truth. :D

[Redeemed]

Yea I think this is a bad decision. I could see maybe 2 PPU's on 1 PCB, but asking for another PCI slot in an enthusiasts rig just aint gonna happen.

Yet again you guys aren't realising that this technology probably isn't targeted for the hardcore-uber-e-wanker gamer. They are probably thinking of commercial fields where this would be far more applicable than any other alternative right now.

With on-board sound and networking, you can easily have two cards for SLi (or Quad-SLi), and two more for PhysX SLi as well. Any more and you would run into problems obviously.

Okay, so they might need a SCSI card or such. So, then the two cores on one PCB would be the better solution.

I think you guys are reading far too much into this. And, I'm also sure tha AGEIA is merely trying to "think ahead". PhysX in its present state, can't remain as it is forever. Even if games do start really utilizing this card in the future, eventually PhysX will have to have a successor. And then another one after that. And eventually, when the physics in game are complex enough, two cores on one PCB might actually be beneficial. Who is to say it wont be?

Obviously, I can see the logic in this- and it isn't at all related to gaming.

[Gnu_Raiz]

I have to agree with redeemed, lets face it Nvidia, ATI, Havoc still have not matched what Ageia has done so far. Even if they do bring out some uber Physics card they will be in the same situation that Ageia was in a about a year ago, with no software.

Now this comes to light about having multiple cards, or cores in the drivers, this has got to be a wakeup call to Havoc, and the others. Who are still faced with software that does not exist. you also have to remember that the prices will get cheaper if more games use the PhysX card, and I bet some of you who laugh at it now might pick up a card if it hits say 60-80 dollars.

Have any of you looked at the cell factor demo's on the net, doesn't that look pretty cool? Then you have the science aspect of this card, come on do you really think that serious researchers will want a Havoc, or a Nvidia which depends on the gpu to crunch the numbers.

You have to consider that multi-core is not really going to solve the problem as some recent interviews state that software development for multi-core is going to be a big deal. Since developers have a hard enough time with 2 cores, they will want to use what is available. I think it might be just cheaper to license the Ageia software, and not worry about how to implement it for multithreaded app's and put the special sauce for more then 2 cores.

[Redeemed]

I have to agree with redeemed, lets face it Nvidia, ATI, Havoc still have not matched what Ageia has done so far. Even if they do bring out some uber Physics card they will be in the same situation that Ageia was in a about a year ago, with no software.

Now this comes to light about having multiple cards, or cores in the drivers, this has got to be a wakeup call to Havoc, and the others. Who are still faced with software that does not exist. you also have to remember that the prices will get cheaper if more games use the PhysX card, and I bet some of you who laugh at it now might pick up a card if it hits say 60-80 dollars.

Have any of you looked at the cell factor demo's on the net, doesn't that look pretty cool? Then you have the science aspect of this card, come on do you really think that serious researchers will want a Havoc, or a Nvidia which depends on the gpu to crunch the numbers.

You have to consider that multi-core is not really going to solve the problem as some recent interviews state that software development for multi-core is going to be a big deal. Since developers have a hard enough time with 2 cores, they will want to use what is available. I think it might be just cheaper to license the Ageia software, and not worry about how to implement it for multithreaded app's and put the special sauce for more then 2 cores.

Currently Ageia only has a handfull of applications (games) that can make use of the PPU, and all of them are rather disappointing in regard to the physics... except for maybe CellFactor demo. But that gets old rather quick.

HavokFX has quite the list for software support, not to mention it has already been used in multiple 3DMark versions. Then again, Ageia has an impressive list of titles that should support the PhysX card- one of which is expected to be blockbuster title, UT2007.

But, already Ageia is getting pretty awesome and fierce competition. Look at how the buildings crumble and fall apart in HL2: EP2. That is all done via the CPU, I believe (though I may be wrong on this). Alan Wake takes a second, with CrySis right behind those two first titles. And all three of these games are using the CPU for physics.

Physics far more complex can be had when using the PhysX chip than a mult-core CPU or a graphics card, but it is up to the devs to prove this to the gaming community. If UT2007 proves to be uber-spectacular with its usage of PhysX, then Ageia has a far greater chance of success.

One thing many gamers are not aware of is that by having all these awesome physics effects that PhysX can offer, it will create more work for your graphics card. Sure, G80 sounds like it will be a beast in regards to performance. One G80 might be enough to play Crysis, Alan Wake, and HL2: EP2 with most eyecandy. But when playing UT2007 and the CellFactor game with max physics effects enabled, that single 8800GTX verywell may come to its knees due to all the objects having to be rendered. This is where we will probably see the greatest gain in G80 SLi... top notch graphics along with top notch physics.

If UT2007 can prove PhysX's worth to the gaming community, then future titles are sure to implement support (and full use of) PhysX. Once this happens, physics in games are bound to become more and more complex. As such, eventually PhysX will become bogged down. This the addition of a second PPU, either on the same PCB or as a dual-core PPU, could become a more pheasable option.

Then that brings us back to the beginning where we have and are implementing such an amount of physical effects and properties that the GPUs become the bottleneck. Here is where G90 SLi or even Quad SLi will become popular, I'm betting. ;)

Just my take on all of this. :)

[Daneel Olivaw]

This is just crazy talk in my opinion. But there you can see it right in the drivers.

:barf: Multiple Physx Cards.Well, being a developper myself, I actually see it as a good thing. It shows their drivers and API are properly built and support hardware scaling. Which is a requirement of good middleware IMO.

Of course, using the multiple PPU argument for marketing purposes right now is just plain silly, and I agree with you there.