Let me explain the business model here...

B

BBA

Supreme [H]ardness
Joined
Apr 25, 2003
Messages
5,382
It seems no one gets it. People are harping on things like games will have different gameplay with Ageia physics and those without will be unable to play...or the physics will be eye candy only, not interactive...etc...

Well, thats all hysteria, let me tell you what is going on here: Ageia is marketing a technology. Plain and simple, nothing more.

Ageia is NOT a hardware company. The hardware PPU card is only for proof of concept, to stimulate the market.

I would bet money the 'REAL' real plan is Ageia wants to sell licensing and get royalties.

Once everyone supports the processing it's a done deal...and guess what, both nvidia and ATi are starting to do just that, incorporating Ageia physics programming in their future drivers.

What does this REALLY mean?

It means you won't need an Ageia PPU to play games that have the Ageia physics technology...all you will need is driver support for it in the video card (although it would take some graphics performance overhead and may effect playable graphics settings)


Commone sense would say that yes, you will get better performance by using a dedicated PPU, but it won't be a requirement. Eventually, a hardware PPU won't be needed at all, as graphics cards become more advanced, then Ageia will be a licensing mogul (not unlike the Dolby Labs business model)
 
How about we just slap a PPU and a GPU on the same PCB? $900+ video cards anyone? :p
 
BBA said:
It seems no one gets it. People are harping on things like games will have different gameplay with Ageia physics and those without will be unable to play...or the physics will be eye candy only, not interactive...etc...
Click to expand...

And the problem is?

Well, thats all hysteria, let me tell you what is going on here: Ageia is marketing a technology. Plain and simple, nothing more.
Click to expand...

Funny, I keep hearing/reading about their cards...

Ageia is NOT a hardware company. The hardware PPU card is only for proof of concept, to stimulate the market.
Click to expand...

"AGEIA’s products include the PhysX™ chip, the world’s first dedicated Physics Processing Unit (PPU), and the NovodeX™ Physics SDK, the first asynchronous (multithreaded) physics middleware engine with comprehensive API support for the PhysX chip."

AGEIA is both a software and hardware company.
Unless you are going to call NVIDIA a softwarecomapny too, as they don't manufactor their own cards? :)

I would bet money the 'REAL' real plan is Ageia wants to sell licensing and get royalties.
Click to expand...

Well, they are chaging an arm and a leg for it, right? :rolleyes:
Try and compare the price of an PhysX license to PC and the price of a Havock license to PC...


Once everyone supports the processing it's a done deal...and guess what, both nvidia and ATi are starting to do just that, incorporating Ageia physics programming in their future drivers.
Click to expand...

You are saying that their GPU's are 100% programmable with a memorybandwith of ~2Tbits/s (terabits per second)?
Just by a driverupdate?


What does this REALLY mean?

It means you won't need an Ageia PPU to play games that have the Ageia physics technology...all you will need is driver support for it in the video card (although it would take some graphics performance overhead and may effect playable graphics settings)
Click to expand...

I so hope not that you are suggesting that GPU's will offer the same physcis(and I don't talk about their wannabe-eyecandy-physics) processing power as the PPU...
Because then you really fell for their marketing...

Commone sense would say that yes, you will get better performance by using a dedicated PPU, but it won't be a requirement. Eventually, a hardware PPU won't be needed at all, as graphics cards become more advanced, then Ageia will be a licensing mogul (not unlike the Dolby Labs business model)
Click to expand...

Common sense would dictate that I would buy the 299$ product(PPU) that give me better preformance and REAL physics over the $500(GPU) product...
Common sense would dictate that HavcokFX will be more eyecandy, meaning just more stuff at reduced res and IQ, with no impact on gameplay physics...meaning that speaking of it like "physics" is a bad idea...
In physics every action has a reaction...this the PhysX can deliver with a preformance that leaves HavockFX in the dust...
As HavockFX does ALL gameplay physics(read REAL physics) on the CPU...

Terra - And that is what really matters here...
 
I'd like to see a setup where dual core processors used to accelerate physics. 1 core would do the normal game stuff and the other core would be dedicated to physics.
 
nobi125 said:
I'd like to see a setup where dual core processors used to accelerate physics. 1 core would do the normal game stuff and the other core would be dedicated to physics.
Click to expand...

It would still be slow as hell compared t a PPU...a CPU isn't very good at physics...like it isn't very good at graphics...

Terra...
 
Terra...read the damn post again and try to have an open mind because you don't seem to see a damn thing other than your own closed opinion.

Nevermind... it would hurt your brain or ego too much.
 
Whoa, clam down there. I dont see where you draw your conclusions from either. Ageia announced their intentions to create a physics add in card way before they acquired novodex which is their physics api. And even if this were their intentions, why create a ppu in the first place? Nvidia and ATI arent creating anything specific to ageia in any way shape or form, they arent trying to trick them into supporting them. I think you're confusing havok, a competing physics software package and their intentions to make use of the now programmable gpu's to do physics work with ageia and novodex.
 
BBA said:
Terra...read the damn post again and try to have an open mind because you don't seem to see a damn thing other than your own closed opinion.

Nevermind... it would hurt your brain or ego too much.
Click to expand...

I read every line of your post, evalueated them and replied.
But I disagree with your view, it's nothing personal.
And sorry for not lying to you, kissing you on forehead and saying: "I 100% agree with you"...but I don't lie to people :)

At least you could take the effort to point out where and why you I think I am wrong in my arguments, I did the same when writing in response to your initial post :)

Terra - Give it to me straight :)
 
BBA said:
Ageia is NOT a hardware company. The hardware PPU card is only for proof of concept, to stimulate the market.

I would bet money the 'REAL' real plan is Ageia wants to sell licensing and get royalties.
Click to expand...

They are giving their engine away for free. Obviously they are doing that to get the market ready for their cards. The cards are very much a part of their plans. Not everyone is going to have a card, which means that they are not going to be able to simply focus just on their hardware like a videocard company might be able to. Their focus has to be on the engine at this point. That makes them not a hardware company?

both nvidia and ATi are starting to do just that, incorporating Ageia physics programming in their future drivers.

What does this REALLY mean?

It means you won't need an Ageia PPU to play games that have the Ageia physics technology...all you will need is driver support for it in the video card
Click to expand...

Wow, just wow. Do you have any proof of this at all or did you just pull that right out of your ass? Nvidia is focusing on Havok FX, and ATi is doing their own thing, but for them to all of a sudden try to support the PhysX Engine somehow would be very strange. There are limitations inherent to videocards which make them unable to do certain kinds of physics and that is why Havok FX will be effects physics only while PhysX will do that as well as gameplay physics. The cards were created because there is nothing else in the computer that can perform the same tasks at the same level. Not a second CPU core, not the GPU. At the very least they would have to design new cards with AGEIA tech onboard. Not to mention you can’t just “share” the same ram with no penalty, and the AGEIA chip itself has almost as many transistors as a GPU anyway. PhysX hardware acceleration is not something that can just be “added on” to something else as an afterthought, and certainly not with just a driver update.

Commone sense would say that yes, you will get better performance by using a dedicated PPU, but it won't be a requirement. Eventually, a hardware PPU won't be needed at all, as graphics cards become more advanced, then Ageia will be a licensing mogul (not unlike the Dolby Labs business model)
Click to expand...

When a computer doesn’t have a PPU, physics acceleration is done on the CPU. Even if either ATi or Nvidia were up to the massive undertaking that would be involved in replicating the significant hardware capabilities of the PPU on their cards, it would represent a significant departure from everything they have done up to this point. Why all the talk of Havok FX if they were just planning to support PhysX engine?

Anyone can speculate, but coming on here and trying to lay it all out as fact, especially when it flies in the face of everything that we’ve read or seen up to this point helps no one. Your post is clearly inaccurate to anyone even half familiar with the product up to this point.
 
Someone explain how Dolby Labs is successful please. They don't make their own hardware, they depend on everyone else.

Ageia is doing exactly the same. They are licensing the technology to other manufacturers and they make their SDK available to 'Registered' developers.
 
GotNoRice said:
Wow, just wow. Do you have any proof of this at all or did you just pull that right out of your ass? .
Click to expand...

I don't know...I read articles like this, http://www.dailytech.com/article.aspx?newsid=1414
or

I like this:
nVidia and Havok went through a very detailed song and dance about how great they both are at GDC and demonstrated Havok FX, an add on to the Havok API that we were told will do wonders for in-game physics on nVidia hardware. ATI, anxious to respond to such claims, claims that since in-game physics relies heavily on floating point arithmetic, its R580 architecture is ideally suited to it since it features 48 pixel shaders and suggests that it has 375 GFlops per card available for such calculations. This number compares favorably to the 10 GFlops available in the fastest widely available CPUs and the 100 Ageia will offer. Other aspects of ATI architecture such as dedicated branching logic, unified shader units and a 3:1 shader/pipeline ratio also offer advantages when performing physics calculations.

Although ATI claims it will support Havok physics fully however, its real secret weapon, it seems, will be its own physics API which it plans to offer to developers soon. The first major and obvious benefit of this is that current owners of X1800 and X1900 boards will only require a driver update to be able to enjoy the benefits. ATIs own low-level physics API will offer developers the opportunity to program for its GPU directly, skipping Direct3D and OpenGL. Developers will also have the option of using D3D and OpenGL.
Click to expand...

from here: http://www.megagames.com/news/html/hardware/atiageiaandnvidia-itsphysics.shtml

Obviously ATi and nvidia have greater hardware capability than Ageia when it comes to calculating interactive physics. It also look like teh Havok physics is more eye candy than interactive physics, as in all their demonstration movies show lots of effects, but not real physics (must have been what Gabe was talking about).

I guess I am just pulling it out of my ass that all these games are signing on as Ageia supportting and now both card manufacturers will support compatible physics processing...sounds like Ageia is a licensing company for technology to me. I wonder what it takes to be a 'Registered' Ageia developer?
 
BBA said:
Obviously ATi and nvidia have greater hardware capability than Ageia when it comes to calculating interactive physics.
Click to expand...

Actually they have zero hardware capability when it comes to calculating interactive physics. The Havok FX engine is purely effects-physics only. PhysX can do effects-physics as well as Game play physics (a.k.a. interactive physics).

I guess I am just pulling it out of my ass that all these games are signing on as Ageia supportting and now both card manufacturers will support compatible physics processing...sounds like Ageia is a licensing company for technology to me.
Click to expand...

Are you trying to imply that the AGEIA PhysX Engine and the Havok FX engine are compatible? They certainly are not; No more than DirectX and OpenGL are compatible. Neither ATI or NVidia has announced anything that is compatible with PhysX. Havok FX is a competing API that is very cool, but somewhat limited in scope on a fairly basic level compared to PhysX.
 
BBA said:
Commone sense would say that yes, you will get better performance by using a dedicated PPU, but it won't be a requirement. Eventually, a hardware PPU won't be needed at all, as graphics cards become more advanced
Click to expand...

That's the same as saying that eventually you won't need a hardware GPU, when CPUs become more advanced.
A PPU is just completely different from a GPU or a CPU. And sure, in theory you could put all three of them on one physical chip, but they'd still be three separate processing units, with their own characteristics.
 
I'm not sure what is what here......but I do know from reading all the news last week that Aegia (or somebody) has put actual real cards in computers made by OEMs Dell,AW, and Falcon (to the best I can remember)....I believe the "card"was sold by BFG, but I dont know where the "Chip" came from.
The graphics card chip manufacturers ATI and nVidia have come out in support of the HavokFX brand of physics and state that their chips can indeed process some physics,too.

Overall I think this is a bad thing. Why???? Well lets say you want to play that game CellFactor......OhOh......Aegia physics......gotta have a seperate card or the game looks dull......and the stuff doesnt fly all over the place.

Now I want to play the latest from HL2 series.....OhOh.....HavokFX.....to get the full effect my graphics card will play me a slide show when stuff starts flying around.....the frame rate hit even on SLI might really suck unless I have the latest 700 dollars cards in SLI or CF.
I think the idea is interesting but it will be VHS and Beta or HD and BlueRay......and all of us will be caught holding our wallets. :eek:
 
That’s neat, never heard that ati is planning to create a proprietary api to allow direct access to its hardware. But again this isn’t specific to ageia, havok will benefit too, but I can see where your tinfoil hat is coming from. Nvidia/Ati are competing to make physics an indispensible part of a game, which can only help havok/novodex, but even with direct access to the gpu the latency kills interactive physics and gpu performance. And you're also forgetting that developers are capable of simply writing their own physics system and port it engine to engine using any advantages nvidia/ati offer.
 
GotNoRice said:
Actually they have zero hardware capability when it comes to calculating interactive physics. The Havok FX engine is purely effects-physics only. PhysX can do effects-physics as well as Game play physics (a.k.a. interactive physics).
Click to expand...

If you read from above you would have seen:

"R580 architecture is ideally suited to it since it features 48 pixel shaders and suggests that it has 375 GFlops per card available for such calculations. This number compares favorably to the 10 GFlops available in the fastest widely available CPUs and the 100 Ageia will offer"

That pretty much gives the raw physics calculation ability performance wise between Ageia and the two major card makers.

If your trying to imply the X1900 would not be doing the calculations in hardware, what do you think a GPU is?

Are you trying to imply that the AGEIA PhysX Engine and the Havok FX engine are compatible?
Click to expand...

Not at all. Havok is not real physics, only effects.
 
Lord of Shadows said:
but even with direct access to the gpu the latency kills interactive physics and gpu performance.
Click to expand...

Not sure what latency your talking about.

Are you implying access to the PPU on a PCI bus card has any less latency? I am sure the memory access speed on a video card is also greater than the memory access spped on the PPU card (after all, the memory on a PPU is made of slower chips in the first place.)

And you're also forgetting that developers are capable of simply writing their own physics system and port it engine to engine using any advantages nvidia/ati offer.
Click to expand...

True, except they are already buying into Ageia's licensing...I have seen it 'Ageia' logo's in games I already have.

I think Ageia is doing something good...they are forcing interactive calculated physics in video games. That is one big statement when you look at it, without Ageia the state of physics would not be progressing as quickly as it is now. I am saying nothing negative, in fact I am saying Ageia has a brilliant business model to be able to force it's own technology to be adopted industry wide (eventually) and they will also reap the ongoing licensing rewards with little development expense afterward.
 
Scali said:
That's the same as saying that eventually you won't need a hardware GPU, when CPUs become more advanced.
A PPU is just completely different from a GPU or a CPU. And sure, in theory you could put all three of them on one physical chip, but they'd still be three separate processing units, with their own characteristics.
Click to expand...


Where are you coming up with this? I am not saying you wont have a PPU, I am saying its functions will be incorporated into video cards, wether by a dedicated chip or by GPU makes no difference, you want the fastest means possible, and right now, 375 GFLOPS (GPU shaders) compared to 100 GFLOPS (PPU) compared to 10 GFLOPS (CPU) means a lot of physics differences.
 
magoo said:
I'm not sure what is what here......but I do know from reading all the news last week that Aegia (or somebody) has put actual real cards in computers made by OEMs Dell,AW, and Falcon (to the best I can remember)....I believe the "card"was sold by BFG, but I dont know where the "Chip" came from.
Click to expand...

Ageia did not make the hardware, they commissioned a company to fab the chips and Asus and BFG did the rest on their own.

The graphics card chip manufacturers ATI and nVidia have come out in support of the HavokFX brand of physics and state that their chips can indeed process some physics,too.
Click to expand...

ATi and nvidia are coming out with support of hardware physics calculations OTHER THAN Havok, by GPU shaders requiring only a driver update.

Overall I think this is a bad thing. Why???? Well lets say you want to play that game CellFactor......OhOh......Aegia physics......gotta have a seperate card or the game looks dull......and the stuff doesnt fly all over the place.
Click to expand...

It's not a bad thing, you have to look ahead a little, there will only be one difference in playing with or without a PPU, and that will be the graphics rendering settings may have to be reduced on the non PPU card because the graphics cards will be doing double duty (unless you have two graphics cards).

See I think your stuck on believing there will be a difference in game play becasue you will have to reduce physics without the PPU card...but I am saying the vid card makers will incorporate Ageia's physics ability in video cards, they are already working on it now while your talkin.
 
I may be wrong about the latency, but if this wasnt the case we would have already seen it happen instead of the "effects physics" we've heard so much about. And I doubt that if simply creating a new software interface to the gpu's to enable realtime physics calculations would have been overlooked by nvidia/ati for so long. I'll take it if they give it though ;o)
 
BBA said:
Where are you coming up with this? I am not saying you wont have a PPU, I am saying its functions will be incorporated into video cards
Click to expand...

That's not what I understood from your opening post, apparently.
You literally said "Eventually, a hardware PPU won't be needed at all".
Which I interpret as "no dedicated hardware, so software emulation over hardware designed for other purposes". Else it'd be dedicated hardware, wouldn't it?
Bad choice of words perhaps.
 
BBA said:
If you read from above you would have seen:

"R580 architecture is ideally suited to it since it features 48 pixel shaders and suggests that it has 375 GFlops per card available for such calculations. This number compares favorably to the 10 GFlops available in the fastest widely available CPUs and the 100 Ageia will offer"

That pretty much gives the raw physics calculation ability performance wise between Ageia and the two major card makers.
Click to expand...

Yeah, that was a quote from an ATI employee at the GDC. Given that AGEIA and Nvidia have been getting most of the physics press lately it doesn’t surprise me to see some “attention grabbing” claims coming from ATi. Beyond that, it is going to take more than vague claims from a biased employee to get me excited about the tech or believe that it is going to really be able to compete in any way. Not to mention that unless it were to support the PhysX engine it wouldn’t do you any good compared to a PPU anyway. You have said that this will be the case, but I would love to see some proof or even anything to substantiate that.

If your trying to imply the X1900 would not be doing the calculations in hardware, what do you think a GPU is?
Click to expand...

Exactly what calculations would it be doing in hardware, in a PhysX game, given that the X1900 is not compatible in any way? That’s right, none.

BBA said:
Ageia did not make the hardware, they commissioned a company to fab the chips and Asus and BFG did the rest on their own.
Click to expand...


The chips are being made by TSMC, which makes chips for a “few” other companies as well.

BBA said:
I am saying the vid card makers will incorporate Ageia's physics ability in video cards, they are already working on it now while your talkin.
Click to expand...

I’m sure that in the long term, you will be able to replicate the functions of the AGEIA card with your videocard, but who says AGEIA won’t have a new card out by then? In the short term, there are going to be a lot of games coming out that support the PhysX engine and unless you have something that supports it (a PPU) you’re going to be out of luck despite how many zomg gflops your GPU has to spare. If ATi or NVidia were to actually support PhysX, it would change things, but right now they seem to be trying to compete with PhysX.
 
without trying to start a holy war....

BBA,

You make some interesting points but I disagree with you in one key area. I would be very surprised if red or green would accelerate Novodex with their GPU's. There are several reasons for this. Its unlikely Ageia, ATI, or Nvidia would even allow it, and even less likely that it would be possible.(at least with current generation gpu's)

ATI is saying their GPU's have or will have all of the capability of Ageia's PPU and then some, but that does not mean they plan on accelerating Novodex on a GPU. They in theory could use a competing physx software that does the same thing. No one from Ageia, ATI, or Nvidia has suggested Ageia's software will be accelerated by a GPU.

Ageia does offer to sell its software to game makers. However, if the dev making a pc game agrees to offload physics functions to the PPU, the dev will be given a Novodex license for free. Ageia software has will be used in many Xbox 360 and PS3 games. Ageia's software can be used within MS's XNA. Sony inked a deal with Ageia granting Sony sub-licensing rights for Ageia's software, and it will be integrated into the SDK of the PS3. Ageia is banking heavily on the PS3, as Ageia's PPU shares a lot in common arcitectually with Sony's Cell cpu. What Ageia wants is a lot of games to be ported from the PS3 to the PC, where a PPU will be beneficial/required. Epic's Mark Rein recently touched on this.
Shack: What are your thoughts on AGEIA's PPU hardware? Any thoughts about how that's going to take off?

Mark Rein: One thing AGEIA's done that's really smart is ... they've done a really good job of optimizing their library to work well with the SPUs in the Cell processor, which means we're going to be able to get a lot of physics performance out of PlayStation 3. Also on Xbox 360 to some extent, but definitely on PS3 we're going to be able to get a lot of physics capabilities out of that. Which means that, to bring [games using those methods] to a PC, you're probably going to need the hardware. Or you could maybe scale it up even further on the PC, I believe, with their hardware. I think that bodes really well for them if developers go nuts and do really cool physics on PlayStation 3, then if people want to play it to that level on PC, they'll buy the card. So it's a matter of them coming out with great applications, great games that use it. I know Ghost Recon Advanced Warfighter will be using the hardware, and Unreal Tournament 2007 will be using the hardware. Not today, but eventually. So I think that's pretty exciting for them, and I think it's going to be pretty cool.
Click to expand...
*IF* Ageia's PPU becomes widely accepted and *IF* they put down the challenges posed by ATI and Nvidia, I could see Ageia starting to charge a bunch of money for software licenses like Havok does. This scenario is a long, long way from happening at this point.
 
GotNoRice said:
Exactly what calculations would it be doing in hardware, in a PhysX game, given that the X1900 is not compatible in any way? That’s right, none.
Click to expand...


I don't know if I'd say that becase of a feature known as programability that GPU's have had for years (well, since GPU's were called GPU's)

Thats why ATi GPU's can hardware compress video as well...a function enabled by software that reprogrammed the GPU (search a few months back and you will find it).

I am sure nvidia could do the same thing.

You have to remember the GPU is a programmable CPU with massive built in parallel processing ability. GPU's are a few evolutions ahead of Ageia's technology already and I don;t see that changing anytime soon. Especially now there is pressure in the physics realm.
 
Low Roller said:
without trying to start a holy war....

BBA,

You make some interesting points but I disagree with you in one key area. I would be very surprised if red or green would accelerate Novodex with their GPU's. There are several reasons for this. Its unlikely Ageia, ATI, or Nvidia would even allow it, and even less likely that it would be possible.(at least with current generation gpu's)
Click to expand...


Ok...look at this: Creative Labs made EAX. There wasnt much support so they said everyone can use it. Immediately every manufacturer started making EAX sound cards.

Why do you think ATi or nvidia would not do the same? Thats their key to selling more hardware...if they can emulate someones hardware and do it better than the hardware itself...and not have to pay licensing...yes, they would do it in a heartbeat.
 
BBA said:
...

Ageia is doing exactly the same. They are licensing the technology to other manufacturers and they make their SDK available to 'Registered' developers.
Click to expand...
This post sounds like you have an axe to grind with Ageia. :rolleyes: Either you don't want PC's to evolve with a great new technology, or you're upset that you have to shell out $250 for a PPU card - which one is it? :confused:

Also, you said, and I quote:
BBA said:
...
Ageia did not make the hardware, they commissioned a company to fab the chips and Asus and BFG did the rest on their own.
...
Click to expand...
So fucking what? nVidia/ATI/Matrox etc. do the same God damn thing...so according to you they're a software company too? :D These companies design the processors (the actual hardware), whether they do the manufacturing themselves or not is irrelevant.

B.t.w., your posts are a confusing mess to read through. They're incoherent so I can't make sense of them and see what the hell the point is you're trying to make. Is it "Ageia sucks"? If so, thanks for wasting my and other people's time with yet another PPU-bashing thread. If not, then_what_is_your_point?
 
BBA said:
I don't know if I'd say that becase of a feature known as programability that GPU's have had for years (well, since GPU's were called GPU's)

Thats why ATi GPU's can hardware compress video as well...a function enabled by software that reprogrammed the GPU (search a few months back and you will find it).

I am sure nvidia could do the same thing.
Click to expand...

You keep coming back to the idea that somehow they are going to embrace the PhysX engine and incorporate that into their hardware. At this point, there has been nothing said to support that while on the other hand they have been actively promoting alternative physics engines. Again do you have anything to back this up, or is it just an unsubstantiated theory? Sitting here brainstorming “what if” situations is interesting, but I’m more interested in reality.

You have to remember the GPU is a programmable CPU with massive built in parallel processing ability. GPU's are a few evolutions ahead of Ageia's technology already and I don;t see that changing anytime soon. Especially now there is pressure in the physics realm.
Click to expand...

How can you claim that GPU’s are “ahead” of the PPU when they are apples and oranges? Once again I have to wonder if you have ANYTHING to back these claims up. Havok FX is a new tech that does not appear to be specific to nvidia. If there is all this magical physics processing ability on these cards then why does Havoc FX do nothing more than effects physics? As far as ATi having an API separate from Havok FX, we really have not heard very much from them yet to the point where speculating that they might be able to whip up something that can actually compete with PhysX seems rather silly. Even if they did whip up something amazing, it would merely put them at the point AGEIA was at years ago with trying to get developers onboard. The nice thing about the PPU is that the tech is real and within a few months people will be playing PhysX games on AGEIA PPU’s. If ATi wants to step up to the plate with something concrete and not just a company representative throwing talk of GFlops out as marketing chum to get people excited, I welcome that. If we’re just talking about what might happen years from now, then how is that relevant?

BBA said:
Ok...look at this: Creative Labs made EAX. There wasnt much support so they said everyone can use it. Immediately every manufacturer started making EAX sound cards.
Click to expand...

You do know, that there has never been a single non-creative card that could do EAX in hardware, right? I thought for a second your post was April fools related even. There is software support for up to EAX 2.0 that many cards have. The PhysX Engine already has full support for running in software mode, and just as EAX2.0 in software uses the CPU to do the work, so will using the PhysX engine in software mode.

if they can emulate someones hardware and do it better than the hardware itself...and not have to pay licensing...yes, they would do it in a heartbeat.
Click to expand...

The key there being “if”, and it being a fairly large “if” at that, and one that conveniently ignores the fact that AGEIA could still just choose to not let them use the tech if they wanted to.
 
If GPU's are so capable of accelerating Novodex, why are PS3 devs using Cell to run Novodex instead of RSX? (Hint: Cell and PhysX share many similarities architecturally) :)
 
If you really wanted a comparison to what the PPU might be in a few years, think back over 10 years ago and look at the very first 3D dedicated card. There was 3D graphics before then, so why do we need a 3D card anyways... right?

Um don't ageia make their own core processors and sell them to board makers, like nvidia/ati?
 
GotNoRice said:
You do know, that there has never been a single non-creative card that could do EAX in hardware, right? I thought for a second your post was April fools related even. There is software support for up to EAX 2.0 that many cards have.
Click to expand...


nForce 2
 
hl3395 said:
Um don't ageia make their own core processors and sell them to board makers, like nvidia/ati?
Click to expand...


none of those make their own cores, they fab the core out.

ATi and nVidia both make their own finished cards (well, nvidia has backed off that in the last few years...may be a reason they have production quantity issues).

Ageia does not make anything but a core design.

All three sell the cores to third parties but in the casde of Ageia, it's the only means of selling the product.
 
BBA said:
none of those make their own cores, they fab the core out.

ATi and nVidia both make their own finished cards (well, nvidia has backed off that in the last few years...may be a reason they have production quantity issues).
Click to expand...

So the correct wording would be:
ATI make cards og design cores.
NVIDIA dosn't make cards but design cores.(they don't even make their own reference cards...)
AGEIA dosn't make cards but design cores...
So if AGEIA is not a hardware comapny, neither is NVIDIA, according to your arguments Your logic is flawed and has been ever since you "declared" AGEIA to be a "softwarecompany" only...
And with you nick and "praise" of ATI's wannabe-physics I am begining to get a impression of you that is of a person guided by brand "loyalty" over arguments...

Ageia does not make anything but a core design.
Click to expand...

Neither does NVIDIA with their videocards?

All three sell the cores to third parties but in the casde of Ageia, it's the only means of selling the product.
Click to expand...

Not understood, sorry?
What are you trying to say, please elaborate? :)

Terra...
 
Terra said:
So the correct wording would be:
ATI make cards og design cores.
NVIDIA dosn't make cards but design cores.(they don't even make their own reference cards...)
AGEIA dosn't make cards but design cores...
So if AGEIA is not a hardware comapny, neither is NVIDIA, according to your arguments Your logic is flawed and has been ever since you "declared" AGEIA to be a "softwarecompany" only...
And with you nick and "praise" of ATI's wannabe-physics I am begining to get a impression of you that is of a person guided by brand "loyalty" over arguments...



Neither does NVIDIA with their videocards?


Not understood, sorry?
What are you trying to say, please elaborate? :)

Terra...
Click to expand...


Your making arguments about stupid tangents, amusingly, in which you are wrong anyway. (Besides, nVidia DOES make it's own reference cards, btw, or do you not remember reading any tech articles where nvidia is testing their new hardware? Where do you think card and heatsink compatibility come from...all card makers getting together and saying we will use this pattern X for this card...?)

The point is ageia is only making initial cores to stimulate the public, get a fan base and license technology. Once this goes mainstream, ATi and nvidia will annie up with better support for the technology and the seperate PPU will die off.

Besides, mr know it all, how come I'm the one with an offer from Ageia to actually buy one (got the offer last month) not you, yet your the one who is hooked?

You need to see life in more realistic terms and in the business world, thats dollar terms, not !!!!!! terms about technology points that are debateable.


I am explaining how the business model will succeed, your arguing about tech points your not even right about.
 
Regardless of statements made by ATI/Nvidia, Gflops penis contest statements don't mean much in describing the ability of ATI/Nvidia to accelerate Novodex. Its pretty obvious current GPU's can't do what Ageia's physX is can do. If they could, devs would be using RSX in the PS3 to process Novodex instead of Cell. Cell shares more in common with Ageia's PPU than it does with a gpu, and, not suprisingly, its superior in running Ageia's software.

Who knows, red or green could buy up Ageia and slap the chips on their boards. Both companies could make their own. Either is a long, long way from happening, and never has ATI, Nvidia, or Ageia stated Novodex could even be run on a GPU, let alone perfom as well as a PPU with regards to running Novodex.

You insist they can, and because of it, you think Ageia is bound to be a software licensing company only. Time will tell. Myself and many others here obviously disagree with you.
 
Low Roller said:
Regardless of statements made by ATI/Nvidia, Gflops penis contest statements don't mean much in describing the ability of ATI/Nvidia to accelerate Novodex.
Click to expand...

Indeed. If gflops were all that mattered, we wouldn't be using CPUs anymore, because GPUs have a lot more gflops than CPUs, so let's run all our applications on them, and make them go much faster?

Gflops just give an indication of the number of operations that can be performed in a given timespan. That still doesn't say what kind of operations these are, and in what kind of circumstances these operations can be performed.
A GPU is a very long pipelined system, and while it can perform a lot of maths along the way, it's very hard to do much with it, other than graphics, because everything is pretty much hardwired to go from a vertexbuffer to a framebuffer. The only way to do 'regular' calculations like a CPU does, is to set up your vertexbuffers with the proper info, render a frame, and then read back the framebuffer.
A CPU is far more flexible, and has far less overhead, so you can easily calc 1+1 and get the result immediately. A GPU with all its gflops would take a LONG time to just calc 1+1, because of the extreme overhead, caused by its deep pipelined and hardwired design.
A PPU is somewhere between a CPU and GPU. It is hardwired to a certain extent, so it can mainly do physics and related operations, but it doesn't have a deep pipeline like a GPU, so there's much less overhead in doing operations and getting their results back.
The result is that the PPU is orders of magnitude faster in calculating physics than a CPU or a GPU. Just like the GPU is orders faster in calculating 3d graphics than a CPU and PPU, and the CPU is still the best way to go for most stuff that we don't have dedicated hardware for.
 
hl3395 said:
If you really wanted a comparison to what the PPU might be in a few years, think back over 10 years ago and look at the very first 3D dedicated card. There was 3D graphics before then, so why do we need a 3D card anyways... right?

Um don't ageia make their own core processors and sell them to board makers, like nvidia/ati?
Click to expand...

Ageia did the design, subcontracted actual chip fabrication out to TSMC, and has third parties (at this point, BFG and AsusTek initially) handling resale (in short, the nVidia model, to a T). ATI differs from nVidia in that they also build cards themselves to sell under their own *brand*.

And you brought up a name near and dear to my heart at least (and one that should be familiar to the history of 3D graphics on the PC): 3dfx. The first two Voodoo card models were *add-in* cards that still required a regular 2D (or even 2D/3D card) to be connected in series with it. I bought both a Voodoo Graphics original (in fact, I later bought a Voodoo 2 as well, which I still have). Both were, in the 3dfx heyday, paired with ATI cards. While 3dfx' cards were primarily known for the Glide API (their preferred API), they could also (much to the surprise of just about everybody else) accelerate not merely Glide and OpenGL gaming, but even Direct3D (you know, that other API developed by a West Coast software company). My three favorite games of that time (outside of Doom and the original Quake/Quake II), while they took serious advantage of the Voodoo Graphics and Voodoo2, were *not* written to the Glide API, or even OpenGL (though the demo of one *was* written in a Glide variant; the finished game didn't support it, however). All three were Direct3D titles (and one was even from Microsoft). Segue to today, and the Ageia PPU. Once again, there is Microsoft, with a title that supports it. What hasn't seemed to cross anyone's mind yet (despite the fact that Ageia has been pointing it out in their press releases) is that the Ageia PPU is both programmable and software-updatable. Why *couldn't* a future update for the Ageia PPU support HavokFX? (Remember, 3dfx also accelerated Direct3D.) It's far from impossible.
 
BBA said:
nForce 2
Click to expand...

nForce 2 is the only *chipset* that supports EAX 2.0 in hardware using no Creative-manufactured parts. The only *sound card* that ever did that was (amusingly) Aureal (as proven by the Diamond Multimedia MonsterSound MX-300; however, that didn't stop the card, or Aureal itself, from dying an ignominious death). I wonder why nVidia never stressed that sort of cross-compatibility.
 
PGHammer said:
Why *couldn't* a future update for the Ageia PPU support HavokFX? (Remember, 3dfx also accelerated Direct3D.) It's far from impossible.
Click to expand...

Great points :)

If HavokFX follows Havok's developer pricing system (250k a pop) I don't see Ageia jumping on board, if only because they will (naturally) take the view that their SDK and API do everything that Havok's does and on top of that is free - so why not just use theirs?

Roll in the other option: open standards.

I've read two articles and many posts attacking Ageia about promoting their physics package, with most claiming that their "closed" PhysX acceleration standard is sub-par and will cause the demise of in-game physics. You'll recall this was said about Glide so many times it just wasn't funny - read on for the punch-line...

The only open physics standard right now is called Collada (currently version 1.4), and they also provide (or will) an interpretation framework to convert one physics API into another.

As far as I am aware Ageia (and others) has/have been forthcoming with details for this. Havok has not :(

I'm sure GPU "physics", while a fantastic idea, will do more to kill off any future open standards in physics than Ageia's PPU will to promote them.
 
Low Roller said:
Regardless of statements made by ATI/Nvidia, Gflops penis contest statements don't mean much in describing the ability of ATI/Nvidia to accelerate Novodex. Its pretty obvious current GPU's can't do what Ageia's physX is can do. If they could, devs would be using RSX in the PS3 to process Novodex instead of Cell. Cell shares more in common with Ageia's PPU than it does with a gpu, and, not suprisingly, its superior in running Ageia's software.

Who knows, red or green could buy up Ageia and slap the chips on their boards. Both companies could make their own. Either is a long, long way from happening, and never has ATI, Nvidia, or Ageia stated Novodex could even be run on a GPU, let alone perfom as well as a PPU with regards to running Novodex.

You insist they can, and because of it, you think Ageia is bound to be a software licensing company only. Time will tell. Myself and many others here obviously disagree with you.
Click to expand...

He is seeing the contest as being HavokFX vs. Novodex (an API battle). There is nothing stopping Ageia from reprogramming Ageia PPUs to support HavokFX except that HavokFX wants (naturally) licensing fees to enable the PhysX GPU to accelerate it. In what way is this scenario implausible?
 
You must log in or register to reply here.
Back
Top