Ageia responds to GPU physics

how many game developers are going to waste timing programming for one of these specialty add in cards, when in another year almost all CPU's will have an extra core sitting idle during most games.

If this had been launched 3 years ago, it probably would have succeded, we'd all have graphics, sound & physics cards. But with CPU's going dual core. I think we'll see the current trend of graphics cards ONLY with software physics done on the CPU and onboard network & sound (onboard = mostly software)
 
We really just have to wait and see how the software side (Havok and DirectPhysics) optimize for dual core CPUs. That isn't going to happen until dual core becomes more standard in the user base. Although the same can be said for PPUs. Whoever gets the good implementation first is gonna win in the long term because of this, although software has the edge as they aren't asking people to buy more hardware, they are simply using something that all users will eventually have. You could probably say PPUs have to get large acceptance by gamers before dual core becomes standard, assuming dual core will be the answer to software-based physics (arguably a big assumption, but I personally don't think so).
 
chrisf6969 said:
how many game developers are going to waste timing programming for one of these specialty add in cards, when in another year almost all CPU's will have an extra core sitting idle during most games.

If this had been launched 3 years ago, it probably would have succeded, we'd all have graphics, sound & physics cards. But with CPU's going dual core. I think we'll see the current trend of graphics cards ONLY with software physics done on the CPU and onboard network & sound (onboard = mostly software)
Click to expand...

A PPU is still ~20.000 times faster at physcis than a second CPU-core...

Terra
 
Terra said:
A PPU is still ~20.000 times faster at physcis than a second CPU-core...

Terra
Click to expand...

Source ? Where's the Bibliography page to your report?

BTW, how much faster is a GPU than a CPU while you're at it?

Just guessing its more than 20x since lets say the 3dMark2005 CPU tests run at 3fps for CPU (at lower resolution/quality) vs 20 fps for the GPU.

thats = 7.0 x + resolution & quality differences = ___ ?
 
chrisf6969 said:
Source ? Where's the Bibliography page to your report?

BTW, how much faster is a GPU than a CPU while you're at it?

Just guessing its more than 20x since lets say the 3dMark2005 CPU tests run at 3fps for CPU (at lower resolution/quality) vs 20 fps for the GPU.

thats = 7.0 x + resolution & quality differences = ___ ?
Click to expand...

1. thats a , not a . not ever one uses , soo think 20,000

2. its works with middle ware like Havoc or anything else

3. id bet money that Direct Physics can use a PPU GPU or CPU to do physics then its out

4. go to there site theres ALOT of big names backing this thing
 
its works with middle ware like Havoc or anything else
Click to expand...

Wrong. It works with the engine they bought, Novodex. It doesn't work with Havok, nor has either party expressed specific interest in working together. The card has to be designed to work with those products, just like graphics cards must specifically support Direct3D and OpenGL.
 
jdLordHelmet said:
Give me an optimized software solution to do it.
Click to expand...

Havok.. but it would choke and die trying.

Edit: How well does it deal with the same kind of thing that the NovodexRocket does? I mean it should really be as optimized as possible.. Can it do 16000 bricks? And if so, how many FPS would it get in something similar to the "big bang" simulation?

==>Lazn
 
Lazn_Work said:
Havok.. but it would choke and die trying.
Click to expand...

Would it? I have yet to see what should be very simple: A demo, with the following criteria:

1) Full optimization for both solution types. Let us say Havok for software, PhysX for hardware.

2) Full, high detail modern graphics. No wireframe, no 1998-quality jenga towers. I want soft shadows, bump mapping, all of it.

3) High number of high-poly objects interacting through physics.

Both modes of the demo should be the exact same, with one making calls to Havok and one making calls to PhysX. Let's see how they both measure up then (I predict both choke). So far, I have yet to see or hear of such a demo. Until it exists, all is theory.
 
jdLordHelmet said:
Give me an optimized software solution to do it.
Click to expand...

http://www.hardocp.com/news.html?news=MTUzOTcsT2N0b2JlciAgLDIwMDU=

software 3d renders I think are giving performance about equal to Intel's GMA (which actually has some hardware on the NB), albeit lowend

But I would think the physics demands would be "low end" at least intially so they're wouldn't be a HUGE difference in coding from the standard physics path (cpu/gpu) vs the ppu.
 
chrisf6969 said:
Source ? Where's the Bibliography page to your report?

BTW, how much faster is a GPU than a CPU while you're at it?

Just guessing its more than 20x since lets say the 3dMark2005 CPU tests run at 3fps for CPU (at lower resolution/quality) vs 20 fps for the GPU.

thats = 7.0 x + resolution & quality differences = ___ ?
Click to expand...



You can't compare a GPU to a CPU and think the exact equasion will fit a CPU vs a PPU...
Look a the Novodex bigbang demo
16.000 obejcts:
CPU 0-4 FPS
PPU +60 FPS

Terra...
 
That demo is of no importance. As I have said, you can't look at a demo using the PhysX API and expect it to run well without a PhysX card. If it did, they wouldn't sell any cards.
 
Is there any word on the PCIe x4 variety as well as the dual ppu boards. Can't wait till those come out. Now that is going to rule
 
jdLordHelmet said:
Wrong. It works with the engine they bought, Novodex. It doesn't work with Havok, nor has either party expressed specific interest in working together. The card has to be designed to work with those products, just like graphics cards must specifically support Direct3D and OpenGL.
Click to expand...


i meant

the PPU works with Middle ware as in there no real programing involved other then grafting in the the middle ware the same way you would now with Havoc

i know it dont work with it i was giveing Havoc as an ex. of middle ware and it not being hard to make a game that uses the PPU
 
chrisf6969 said:
how many game developers are going to waste timing programming for one of these specialty add in cards, when in another year almost all CPU's will have an extra core sitting idle during most games.
Click to expand...

That's what DirectX should resolve if it would support it. If a hardware function was unavailable, it would do it in software, that's how it worked when 3D cards with different featuresets when DirectX was new. I suppose that's why MS is developing their own physics.
 
heres something for u guys argueing about how thers not enough uses for appu, and how the current uses only go so far. one thing id like to see , is the water in games, really be like a fluid....in hl2, the water is really just a diffrently colored part of the map, with lower gravity, and sprites when something enters or exits. thus in garry's mod u cant make a true submarine, or make a boat and have it float the way it should. do u call this fluid dynamics? not sure but things like that would be nice, or like have puddles build up on streets when it rains, and then have all that water on the ground flow into storm drains.
 
x_dreamer said:
heres something for u guys argueing about how thers not enough uses for appu, and how the current uses only go so far. one thing id like to see , is the water in games, really be like a fluid....in hl2, the water is really just a diffrently colored part of the map, with lower gravity, and sprites when something enters or exits. thus in garry's mod u cant make a true submarine, or make a boat and have it float the way it should. do u call this fluid dynamics? not sure but things like that would be nice, or like have puddles build up on streets when it rains, and then have all that water on the ground flow into storm drains.
Click to expand...


i see flight sim and race sims realy geting the most out of this thing :D
 
Lazn_Work said:
Seriosly, do you know? Did you realize the Power IBM chips that Apple uses are less RISC than P4's?

Pretty much all CPUs are "post RISC" these days. And the PPU is more of a SIMD streaming CPU than a RISC CPU.

http://arstechnica.com/cpu/4q99/risc-cisc/rvc-1.html

==>Lazn
Click to expand...

yes, I do. He wrote that a RISC machine can't "execute a wide variety of instructions." You want me to read endless paragraphs of speculation, excuse me, BS, after that?

agree on the SIMD point.
 
You completely misread my post. First of all, I never said a RISC processor can't execute a wide variety of instructions. I wasn't talking about RISC processing literally so much as the "RISC" factor someone else attributed to GPUs on another thread. The idea that by reducing the instructions, i.e. complexity of the process, the process will perform those instructions faster. It wasn't meant to be as literal as you are taking it. I AM a programmer, so I'm not trying to get to a sub-engineering level of detail.
 
Rabidfox said:
I imagine a world where blowing the hinges off of a door to breach is a reality. You're simply not blowing the hinges off, you're splintering the wood around the door and realisticly warping the steel hinges, perhaps bending the door in with a well placed kick by the knob.

Instead of bullets dealing damage, they have actual ballistic characteristics, and your body has the consistancy of flesh and bone, and said bullet actually tears bits away as it hits you, or cracks your plate, if you're wearing armour. Maybe taking a headshot with a helmet will actualy send you ass over tea-kettle, knocking over anything in your path.

MIllions of particles of dust are kicked up behind you as you dash through the streets of some god forsaken country, when you stumble through the next archway, a street dealers cart is pushed in front of you, which you immediately tumble over. The car is tipped over and out spill the 250 apples and oranges and taiwanese trinkets he was going to sell. Lucky for you, you fell just as the archway above you has a dozen 7.62 mm holes punched through it by automatic weapons fire from the third world militants behind you. Light streams through the holes and refracts off of the dust as you scramble for cover.

Yeah, so you can quit trying to stop progress because you're an amateur graphics programmer, if the coders for UE3 think it warrants inclusion, and so does Sony(ps3) and Nintendo(revolution (x360 got to market to quick for PPU inclusion)) then I think it could be a worthy inclusion. It's just like saying "Hey, this cpu runs wordpad, why would I ever need anything faster?"
Click to expand...

just thought i would point out the part where you talk about a headshot tossing someone end over end would not be a realistic response.in real life the person would just fall with at most a small backwards motion or forwards motion. check out Mythbusters sometime, they have tested this and found that even a .50 weapon did not toss a body backward more than an inch due to mass. the bullet is moving fast and has momentum but lacks the physical mass required to actually effect the human body as you describe.
 
acidfr33k said:
yes, I do. He wrote that a RISC machine can't "execute a wide variety of instructions." You want me to read endless paragraphs of speculation, excuse me, BS, after that?

agree on the SIMD point.
Click to expand...

Oh give him a break, what he said was technically true, in that the concept of RISC was to have a few simple "equal complexity" instructions.. Now days the "RISC" processors are no longer truly RISC, and as you know do execute a wide variety of instructions, but initially (say with the 603e) they were simple by design.

But I think jdLordHelmet is underestimating the PPU, while the rest of us are a little too excited about it, all things considered.

==>Lazn
 
Nomikal said:
Supposedly in the next month or so...
Click to expand...
Manju Hegde said in the interview that "At this time, we are evaluating the titles in development and are assessing our launch schedule accordingly.", so maybe the rumours about a Q2 2006 launch could be true then. Who knows. Let's wait and see.

- - - - - - - - - - - - - - - -
AGEIA PhysX PPU Links & Info Page
 
jdLordHelmet said:
I say hardware physics will never be needed.
Click to expand...

As the person below your posts pointed out, statements like that have been heard before.. yet, for some reason.. the products still come about.

Will ageia be a success? Maybe, maybe not.. But even if their product is a complete failure, that does not mean that there won't be some programmer out there designing support for it.

Years ago, Dual processing was "only for workstation / rendering environments"

Bill Gates said:
640K out to be enough for anybody
Click to expand...


Times change my good man, times change.

Hell, I've got a faster internet connection in my home than the college I went to had for the entire campus when I went there.

(11.0 Megabit for me.. they had a t1)
 
LoneWolf said:
I'd like to see a company like Ageia produce a less expensive product, or PhysX processors that could be added to a graphics card, rather than a separate product.
Click to expand...

See, I don't like this idea.. Want to know why?

I'm going to use arbitrary numbers, but lets just say that you are joe computer gamer.

You decide to buy a new video card.. say an x800pro for $200.

You can buy the x800Pro-PhysX for $329.

Ok.. so you do that, and you are happy. 8 months later, HalfLife 3 comes out, and it needs an x2600xl to run well. Price of the x2600Xl - $299.

Oh, wait.. You're going to want the PhysX processor too.. Make that $428.

With an integrated device, you are going to have to re-buy the same thing over and over again. This is why I am opposed to the All-In-Wonder video cards from ati.. Everytime you go from AIW-9000 to AIW 9600 to AIW 9800Pro to AIW X700 to AIW X1600 you are purchasing ANOTHER tv tuner. If you just had a tv tuner [like the hauppauge] you wouldn't have to rebuy the same part over and over again.

Also, putting the physX on the same card means that they have to share the same I/O slot, rather than having a seperate I/O slot for operation.



The thing that will make PPU tech really worthwhile for the end user (and game programmer) is some sort of official support for it from a big-boy.. Say microsoft implementing "External Physics Calls" into directX 10 or something. However, I'm sure that Ageia probably wants their own tech to stay in their own tech. From a programmer standpoint, if it's a directX feature, then it's much more straightforward to program for than having to design your own interface for some "small" upcoming technology.
 
jdLordHelmet said:
That's just not true. We've had physics to do what you describe - shooting a certain part of a bridge to make it collapse due to physics - even before we had the Havok engine (Max Payne 2 even had stuff like that). Do I even need to mention some of the physics-based traps and such from HL 2?
Click to expand...

Uhm, wasnt the havok engine in 3dmark2001 ?
Maybe i'm thinking 03.. but I swear I saw "Havok" in 2001
 
CodeWaste said:
Im wondering how sustainable this will be? Right now there is A LOT of room for improvement in graphics to get to photorealism. But say you're "rendering" gravity on a PPU,
the accelaration rate is always going to be 9.8 meters per second at 1G. even in 50 years when computers are doing things we cant even comprehend right now, accelaration will always be 9.8 meters per second at 1G.

How much room for improvement is there really going to be? is this a one trick pony?
Click to expand...

Well, there is the concept of alternate environments - What's your acceleration rate when you've got drag?, or are on the moon. - What about the adjustments that will occur if, in game, someone shoots your parachute full of holes in BattleField 4 ?

A PPU will be able to calculate the speed adjustment that has to occur while your video card is free to render the impending splat..
 
MartinX said:
Why is being able to realistically blast through walls a good thing?
Click to expand...
Cause it makes more sense. It is more realistic if, in a game, when you fire 500 rounds at opponents standing in front of a balsa-wood door that it actually gets holes in it. Imagine the fun that we could have had in Wolfenstein 3d if we could have killed a room full of nazi's by collapsing the roof in on them instead of just shooting them.

Imagine the feel you'd get from "seeing" the inertia shift of items in your cockpit in a flight simulator, or being able to visibly "Feel" the shockwave of a sonic boom. The concussion grenades in battlefield2 are pretty slick, as a graphical effect, however, after about 5 seconds, they are just "fog"..

Why is grass blowing realistically in the wind a good thing?
Click to expand...
Well, maybe it's unimportant for a first person shooter.. But Sims3 where your virtual family is having a picnic is nice.
TigerWoods Golf 2007 where the wind is blowing the grass at about 5mph to SxSW direction meaning you've got to adjust your hook a little..

Why is being able to model the trajectories of stones in an avalanche a good thing?
Click to expand...
Back to the concept of "Crush your enemies with the environment" - It's used already in games like redfaction, - and HalfLife 2's use of the gravity gun (picking up opponents and flinging them away..)

And the over-arching "trick" question you should bear in mind,: Why do we need a PPU to do them?
Click to expand...
Cause its Physics, not Graphics that calculates how the graphics should render it.

10 Years ago, when we had the 386SX processor (that didn't have a math unit) there were math co-processors. - But just because you didn't have one, didn't mean your computer couldn't do math.... - It just meant it couldn't do it effficiently.

GPU's are not physics processors. They don't have subroutines for calculating the coefficient of drag. All items in a GPU would be currently calculated as if they are in a vacuum. - A radeon doesn't understand the difference between a feather and a brick falling. That's completely handled by software (physics programming within the game engine) . Currently, the process of calculating how a brick would fall differently than a feather would, (if it's not completely pre-scripted) is offloaded to the CPU, then the results are handed back to the GPU to process into polygons. The cpu, while they of course, are getting faster and faster every day, already has work to do. Offloading that task to a seperate device can only improve things.

Yes, the cpu can do everything itself. - Just look at how well though.

Run 3dmark05 and see the framerates you get from the "Dedicated 3d Graphics Processor" (GPU) vs the "General Purpose" processor (CPU) - My athlon64 does not give anywhere near the image quality, or framerate as what is done by my more specifically purposed graphics device.

The PPU is an add on to that.

Do we "NEED" it? No.. because there isn't any support for it .. yet
But do you "NEED* a 3d accelerator? NO. You can run Battlefield2, Fear, and and Quake4 on an onboard, 2d only video card if you want.. But if you want it to run FAST... you'll need that 7800GT or whichever.

This is the very same argument that occured when the Voodoo2's came out. "Who needs a seperate 3d Rendering card?" - Well, apparently quite a few people did.
 
jebo_4jc said:
I think you are right.
When I worked at Best Buy (probably 2 years ago...9800XT and 5900 Ultra were the top-end video cards at the time), I sold hundreds of GF4MX's, and 5200's, and etc. to pissed off people who took a copy of a game home for their kids and found out that their Intel Extreme graphics weren't so extreme. So now they are coming to me with an attitude saying things like "you want me to spend another $100 to play this stupid game?"
Now, if they take a game home, and find out that not only do they need to upgrade their graphics, but they need a new PPU too, they will be doubly pissed.
Click to expand...

Yes, but it's not your fault they can't read the hardware requirements on the side of the box.
 
CodeWaste said:
Look i was making a generalzation about accelaration, its really not the point. the point is, physics dosen't change. its not like its gonna be 9.8 m/s^2 today, but next week its gonna be 3.4m/s^2, no its always going to be 9.8.
Click to expand...

As I make mention of.. - The laws of physics may not change.. but you are being very blind about the concept that maybe not every game in the world is going to be "standing on terra firma at sea level"
 
CodeWaste said:
Anyways the whole idea soon became irrelevant due to the fact that faster CPU's were developed.
Click to expand...

No, actually, it became irrelevant due to the fact that the math coprocessors were integrated into the cpu itself.
 
chrisf6969 said:
how many game developers are going to waste timing programming for one of these specialty add in cards, when in another year almost all CPU's will have an extra core sitting idle during most games.
Click to expand...

You assume everyone in the world is just going to ditch their P4 3.6's and athlon 3800+'s for dual core.
 
Malogato said:
Originally Posted by Bill Gates, 1984
640K out to be enough for anybody
Click to expand...


That statement is a myth.
He never said that, and I dare you to find the original source :)
Had to point it out, sorry *S*

Terra - Bill Gates is very popular in urban myths...but myths are no truth ;)
 
Malogato said:
This is the very same argument that occured when the Voodoo2's came out. "Who needs a seperate 3d Rendering card?" - Well, apparently quite a few people did.
Click to expand...

Exactly. Anything function offloaded from the CPU will have more capabilities that are done more efficiently. Just look at 3d accelerators and sound cards.
 
DaLurker said:
Exactly. Anything function offloaded from the CPU will have more capabilities that are done more efficiently. Just look at 3d accelerators and sound cards.
Click to expand...

Or NIC's
Or SATA/SCSI controllers

Terra -The CPU got enough to do already, any offload will be a plus :)
 
Terra said:
That statement is a myth.
He never said that, and I dare you to find the original source :)
Had to point it out, sorry *S*

Terra - Bill Gates is very popular in urban myths...but myths are no truth ;)
Click to expand...

Have you actually read any of his books?

While he does not say that in his book, you can see that there was a time where he planned for the future, "within limits" -- At the time he supposedly said it, 640k was a TREMENDOUS amount of memory (64k being the norm), so it would not be unusual for him to inflate something 10x (and to the cpu limitation... ) and say "That's enough for now"

He didn't say "That should be enough for anybody forever" It was based on the fact that the limitations imposed by hardware didn't need to be overcome for quite a while.
 
You must log in or register to reply here.
Back
Top