Ageia PPU pricing announced

You are entirely wrong as horse and buggys were limited by the power of the horse, and the people riding the horse and buggy on the cobblestone road did not have a smooth ride. As such there was a need to develope something that could overcome these hurdles, because the technology at the time could not keep up with the need.
Click to expand...

Actually it doesn't matter how smooth the ride is. I CERTAINLY wouldn't drive a ferrari down a cobblestone/dirt/rutter road. Would bottom out and tear the car up. Plus I wouldn't get there anywhere near as fast.

So we could stick with the dirt roads and ride buggies or we could build a nice road and drive ferraris. I know where I want to go :)
 
aringail said:
So we could stick with the dirt roads and ride buggies or we could build a nice road and drive ferraris. I know where I want to go :)
Click to expand...

i'll take a flying car :p
 
I thought they were going to have differ models. Maybe this is just the high end model with a 150$ mid model and a 75 low model.
 
Cause if you do I have a lighting addon card, an AI addon card, and a water ripple math processor card, that I am looking to sell.
Click to expand...

Honestly if you did you would probably make a mint. Especially with the AI card :)

And I do believe his comparison between a buggy and a physics engine is sorta valid.

Once you have ridden in a fast car with good suspension you really don't know what it's like in something other than a buggy
 
aringail said:
Honestly if you did you would probably make a mint. Especially with the AI card :)

And I do believe his comparison between a buggy and a physics engine is sorta valid.

Once you have ridden in a fast car with good suspension you really don't know what it's like in something other than a buggy
Click to expand...

Well then this thread is pointless as comparing a cobblestone horse and buggy to a physics add-on card is asinine at best.
 
I know this is radical, and probably not in the best interests of the computing industry, but I think they should ditch the x86 architechture and go with something far better.

x86 has been around since the late 1970's. Sure, it must be a good archtechture because it has lasted this long, but surely there is a better alternative.

Even though they are not as good for general computing, look at the new consoles coming out. For a sub $400 price range, they are able to compete (and probably currently surpass) the high-end PC in terms of gaming. They absolutely crush midrange PC systems.

It is time for change.

The need to modify the current PC system in order to keep up with consoles (in terms of gaming) and other non-x86 platforms doesn't make sense.

Although there are great performance benefits, not many people want to shell the same amount of money on a PPU as a mid to high-end GPU.

Just my $.02.

If there is anything flawed with my reasoning, let me know.
 
most current software is based on x86

so if some new kind of CPU was designed, wouldn't software need some kind of emulation or something to keep it all backwards compatible?
 
The point behind the analogy (It's an analogy not a comparison). Is that we don't really know how much we are missing by having approximated physics as opposed to something this card can provide. The developers do their best to hide this from us. It's very probable that this is the next big thing as GPUs were the next big thing for graphics when the gForce first came out.
 
Brent_Justice said:
most current software is based on x86

so if some new kind of CPU was designed, wouldn't software need some kind of emulation or something to keep it all backwards compatible?
Click to expand...

Yeah, that is true. I did not take that into account. What if they devised a way so that the software did work on the new hardware?

Also, isn't .NET and java multiplatform? I know that microsoft is pushing this, which is an excellent move on their part.

Current software could also be ported to whatever the new platform is, too. But yeah, this is not an easy task...
 
until we reach photorealism and completely life like environments there is always room for improvement which hardware can provide

there is always room for advancement

we are not at lifelike realism yet
 
My biggest question is that of adoption: How many developers, already crunched for deadlines, are going to spend extra time adding support for a feature that MAYBE 1% of users will actually have? Not everyone has the luxury of time wrt to programming in 'bonus features' such as support for optional hardware. After all, look at how few games support force feedback, or even widescreen displays?

I just don't see the market share being there for this product, despite its potential benefits.

PS: Epic said that their new physics engine can handle 600 objects at a time, so I really don't see too much need for this device right now. Will one be needed in the future? Perhaps, but the timing isn't right for this to be successful.
 
By adding a physics API it should take a shorter amount of time. Just like it's alot quicker to use Direct3D or OpenGL than doing it yourself.
 
Time to start looking for a water block for my overclocked PPU to add to the loop. :D
 
Jonsey said:
Time to start looking for a water block for my overclocked PPU to add to the loop. :D
Click to expand...
Drat. I was going to make that comment. Blech. Course if you overclock it you'll just end up with grass that waves faster :)
 
BO(V)BZ said:
PS: Epic said that their new physics engine can handle 600 objects at a time, so I really don't see too much need for this device right now. Will one be needed in the future? Perhaps, but the timing isn't right for this to be successful.
Click to expand...

but can the hardware handle 600 objects and still perform to playable levels.

afik all engines (physics, graphics, whatever) are able to handle more than the current fastest hardware. watch the raw tech demos, then look at how those engines are used in games. we only get a fraction of whats "possible" in order to prevent games from turning into slide shows.
 
BO(V)BZ said:
My biggest question is that of adoption: How many developers, already crunched for deadlines, are going to spend extra time adding support for a feature that MAYBE 1% of users will actually have? Not everyone has the luxury of time wrt to programming in 'bonus features' such as support for optional hardware. After all, look at how few games support force feedback, or even widescreen displays?

I just don't see the market share being there for this product, despite its potential benefits.

PS: Epic said that their new physics engine can handle 600 objects at a time, so I really don't see too much need for this device right now. Will one be needed in the future? Perhaps, but the timing isn't right for this to be successful.
Click to expand...

The time may not be right for it to be 100% successful right now..but does that mean give up and go home? Pack it up folks...640k out to be enough for anybody. The cell phone wasn't an instant success overnight...but it did grow...and consistantly grow. If the idea is good enough....people will use it. Just like they added HDR to Farcry. Did we need it...no, was it cool that they desided to start using it....yes, will it become widely used in the future....probably. Should they give up because it wasn't 100% adopted day 1...according to your logic...the answer is yes.

One of our engineers at work took a gander at the specs. He figured if the card could do 1/10 of what it claims that he could reduce his modelling time by an order of magnitude. What would take a day to simulate would take an hour or so. With engineers being billed out at $100-200/hour...any boss with 1/10th a brain would buy these ASAP. Any engineering software company who does dynamic analysis would be nuts not to incorporate this. It is just like when the FPU was added to the PC....in engineering land..it was great. I can just imagine being able to model a building swaying in the wind in near real time for maybe a $10k. Hell...they could sell that $10k box for a million and easily get away with it.

I think we need to stop looking at just the game market and at markets that can net them much more money. The gamers just might be the pilot for a much bigger idea.

In my PC at work, I have an add-on card that has 4 Analog Devices DSP's on it that I use to do certain filter functions in real-time. Compared to pure software...it saves me about 80% of my time. We paid $2k for the card...it has since saved about 400 hours of work. Do the math.

-tReP
 
aringail said:
Actually it doesn't matter how smooth the ride is. I CERTAINLY wouldn't drive a ferrari down a cobblestone/dirt/rutter road. Would bottom out and tear the car up. Plus I wouldn't get there anywhere near as fast.

So we could stick with the dirt roads and ride buggies or we could build a nice road and drive ferraris. I know where I want to go :)
Click to expand...
Exactly my point. Though it was deleted as off topic. Nice narrow focus we have, I guess.
 
bonkrowave said:
Well then this thread is pointless as comparing a cobblestone horse and buggy to a physics add-on card is asinine at best.
Click to expand...
No it's not. And I'm deeply saddened at how much some of you lack creativity here. If you never thought you would ever ride in something better than a horse and buggie, you would never have to build anything better than a dirt or cobblestone road.

If you never expect you will ever have better graphics or better physics in games than you have now, you would never expect to need a PPU. What you think of as the top of the line physics model in games today is naturally limited by the power available to perform physics modeling in game engines and in GPUs and CPUs. Once you have a dedicated PPU, the degree and detail of a physics model could increase exponentially for all we know. To say that because physics models are fine on today's technologcy is no different than saying transporation in 1900 was just fine on dirt roads and there would be no need to improve the infrastructure.

Is THAT off topic ?
 
All I can say is what in the heck... do some of the people in this thread not understand basic economics and the history of all technology that has ever existed? PPU's will be great :)!
 
GoldenTiger said:
All I can say is what in the heck... do some of the people in this thread not understand basic economics and the history of all technology that has ever existed? PPU's will be great :)!
Click to expand...
True enough. The market will set the price anyhow. If they price it at $250 and nobody buys it, either it goes the way of the Edsal, or the price drops. Therein lies the beauty of capitalism.
 
GoldenTiger said:
All I can say is what in the heck... do some of the people in this thread not understand basic economics and the history of all technology that has ever existed? PPU's will be great :)!
Click to expand...

Can you imagine how good vritual pr0n could get? Give a person a VR helmet and some gryo sensed gloves and you could have the best VR porno on the face of the planet.
Squeeze a booby and see it actually change shape in realtime :) Boy am I a sick bastard....

-tReP
 
They really need a killer app..a killer game as it were, to demonstrate this card. Some kind of FPS or something with crazy, fully destructable environments. Otherwise, I don't think there will be any real reason to buy one of these, especially for the prices they're charging.

Also, I wonder if you can overclock these ;) D00D! I G()7S /\/\Y F1ZZ1X P|20C3SS()|2 {_}P 70 300 MHz!!
 
Wow, the videos are impressive.

A first-implementation PPU, guys. Maybe not even the 3DFX Voodoo of physics processing--perhaps it's jus the EDGE. Regardless, someone has to make the first in-roads.

Maybe my money tree will bloom soon..
 
Just another price to add to PC gaming. Hopefully the next gen consoles will do me justice.
 
Brent_Justice said:
wow, that second video with the last scene, tons of destruction, neat
Click to expand...

No kidding. I gotta get me one of these so I can watch a house blow up all day...the truly sad part is I'm not being sarcastic....*sigh*

If they can even get halfway between the two in gaming I'll buy
 
Trepidati0n said:
One of our engineers at work took a gander at the specs. He figured if the card could do 1/10 of what it claims that he could reduce his modelling time by an order of magnitude. What would take a day to simulate would take an hour or so. With engineers being billed out at $100-200/hour...any boss with 1/10th a brain would buy these ASAP. Any engineering software company who does dynamic analysis would be nuts not to incorporate this. It is just like when the FPU was added to the PC....in engineering land..it was great. I can just imagine being able to model a building swaying in the wind in near real time for maybe a $10k. Hell...they could sell that $10k box for a million and easily get away with it.
-tReP
Click to expand...


This was exactly my point on page 1 of this thread. This has the potential to be an enourmous boon to the scientific and engineering community. Thus, if it becomes a needed thing in the business world (very possible), then it will definitely become a commodity market. which means prices will fall, and that's a good thing. :)
 
Nice find with the videos. The first, smaller one was damn impressive, and the second one really didn't seem like anything that can't be done now.

I definitely see the use for a PPU for a scientific standpoint, but I still question its retail viability. Its marketshare potential is so small, I just wonder how much use it will get. Do you really think that game designers will expend the effort to create effects such as the ones found in the first video when only about a tenth of a percent of computers AT MOST will support them?

On a different tangent, I'm assuming that physics calculations of this type most likely consist of a lot of vector operations in parallel, similar to how graphics is an essentially infinitely parallel workload. What's to say that nextgen, SM4.0 graphics cards couldn't be highjacked to process this sort of info? I could be way off base in the type of processing needed for physics, but vector ops sounds about right to me. After all, once SM4.0 hardware comes out, it's essentially a big pool of vector units that can be fully configured. Note also how a company has designed software that can use NV40's to process audio.. I'm going to guess that SM4 will make this much more commonplace.
 
Yeah, but how will it make the Sims, Deer Hunter and Roller Coaster Tycoon look.

No seriously,

I think the trend from game developers has been to cater to the low end of hardware. I was frankly shocked at how well D3 ran on relatively old stuff...HL2 as well (look at the hardware poll).
They increase their sales base that way...spending a lot of manhours to explicitly support hardware that only a handful of 'enthusiasts' purchase is back burner.

Yeah the comparisons to Voodoo1 and T&L and such hold water, but the gaming industry *is* different now. The small devs have been gobbled up or destroyed. Current games are simpler and dumber, for a 'broader appeal' (and IMHO, they are mostly derivative and uninspired). Release cycle is king. The publishers are going to push console use to better control their product.
 
and the second one really didn't seem like anything that can't be done now.
Click to expand...
Yeah the first part was kind of lame. And they spent FOREVER doing it. But the last part, where the plane crashes into the barrels and knocks things everywhere was pretty cool. Of course what wasn't cool was the plane failing to clip properly with the ground. That's kind of lame. If you have a demo of physics at least make sure it doesn't fail basic collision rules.

The small devs have been gobbled up or destroyed. Current games are simpler and dumber, for a 'broader appeal' (and IMHO, they are mostly derivative and uninspired). Release cycle is king.
Click to expand...
Sadly I agree. But what I am thinking is that if it's possible for the engines (Unreal/Source/Quake) to incorporate the physics api that Ageia uses then the development houses will get the physics boost for free. Just a hopefull wish :)
 
aringail:

That UE3 or the Doom3 engine incorporate Novadex is the only thing I can see saving this first iteration of the PPU, but each already have their own, damn fine physics engines, so I don't see Sweeney or Carmack just throwing out all their hard work.

Also, going along with the reduced number of gaming studios is the increased cost of developing games. People are saying that next generation games will have to have budgets of 10million. At that point, publishers are not going to stand by and watch their cash cows get delayed so that optional features are added. While it might be easy to just drop in their physics engine and let the PPU do all the heavy lifting, this won't make games look better, just run a little faster. It's going to take a huge amount of extra work by the artists to actually take advantage of what the PPU offers.

Think of it this way: What if the doom3 engine's fancy lighting only worked on Geforce 6800 Ultras? Think it would even exist? That's the thrust of my argument...if all next-gen games suddenly added in REAL support with features as cool as that house-demo video, I'd buy a PPU right now, but they won't, and they might not ever, until CPU's possess the power to do the work themselves.
 
spugnor said:
This was exactly my point on page 1 of this thread. This has the potential to be an enourmous boon to the scientific and engineering community. Thus, if it becomes a needed thing in the business world (very possible), then it will definitely become a commodity market. which means prices will fall, and that's a good thing. :)
Click to expand...
I totally agree! Why does everybody think this card was developed around gaming? We aren't the only market for a physics card you know. This is probably highly priced out of the chute for Engineering firms Movie Studios, and Game Development studios. The Average consumer won't be targeted until those sales are done. Sure they will get some of the Hard Core Gamers. But They aren't the main target right now.
 
Brent_Justice said:
wow, that second video with the last scene, tons of destruction, neat
Click to expand...

Very impressive, indeed... I just watched it :)! I also liked the plane crashing into the globe as you could see it crumple up realistically and deform as it plunged into it while exploding.
 
What if the doom3 engine's fancy lighting only worked on Geforce 6800 Ultras
Click to expand...
But that's just the point. Not all cards CAN do all of the fancy lighting. And not all cards can do the most recent shader tech, but by using and API (OpenGL or Direct3D) you can tell the API to do something. If it doesn't have the hardware then it will do something less in software. Or ignore it altogether.

It's the same thing with enabling EAX of A3D for sound.

If Agiea can get there API to be popular in the market then push the card it will be used alot more. To do this they need to release the API to the public and get/take feedback. The software version of the API should be free so it's easier to adopt. Just like OpenGL.
 
If there's a FUN game that can take advantage of the PPU, i'll buy them both. But the games have to be fun too, i dont' want a souped up version of Pong with improved bounciness.
 
I am in the boat that a physics processor is only worth about $50-$100 bucks. Even then, in order for me to spring for it, most every game needs to be programed to support it. My question is, what sort of revolutionary effects are we going to see added? I mean, I already have hardware soundcard, hardware video, after that, I always read about having x amount of free CPU cycles to work things like physics and AI. Shit, we take the offload of physics, and what is my main CPU going to do? Play the MP3 files I listen too while playing, and uncompressing game data...?
 
bonkrowave said:
My 2500+ is nowhere near top of the line, and if it can handle the current generation physics engines, well then ... what realy is the point of a PPU at this point in time?
Click to expand...

Probably beaten, but ...

I believe Tim Sweeney said that current games have, at most, about 50 "physics objects" on screen at any one time, because that's about all today's CPUs can handle. He said a dedicated physics chip could handle 10,000 physics objects simultaneously.
 
You must log in or register to reply here.
Back
Top