Nvidia Physx strikes again: Cryostasis tech demo

ON the subject of HDR and the fact that it was more difficult than most HDR was that a lot of games used openexr HDR or something like that which required the same hardware abilities as SM3. Valve specifically worked on it's own solution which allowed all sm2 or sm2.b i think hardware to run it thus making a lot of supposedly out of date hardware compatible with it's games.
 
necrophile said:
ON the subject of HDR and the fact that it was more difficult than most HDR was that a lot of games used openexr HDR or something like that which required the same hardware abilities as SM3. Valve specifically worked on it's own solution which allowed all sm2 or sm2.b i think hardware to run it thus making a lot of supposedly out of date hardware compatible with it's games.
Click to expand...

Pfft, I implemented HDR bloom effects on a GeForce2: http://bohemiq.scali.eu.org/forum/viewtopic.php?t=36

What Valve did is no big deal.
By the way, OpenEXR is a file-format, not a rendering technique.
 
Scali2 said:
Ironically enough I myself was rather late into the 3d accelerated arena... Then again I was a software rendering specialist. But I rehabilitated and do OGL and D3D aswell now :)
Click to expand...

I was too busy with SNES and N64 games to care much about PC games, then I got my first PC with a real video card (3dfx Voodoo III 3000). Ultimately I was playing more N64 games emulated on my PC, though :p

Now that I'm a game developer I still don't play many games... but I guess that's why I'm making them instead of playing ;)
 
Silus said:
And that's basically what happens now, in games like Crysis for example. You can interact with vegetation and when you do, passing through them not only makes them move according to you, but it also makes a sound, that can trigger the attention of some KPA or Alien near you. By toning the physics down, you don't see the plant move, but you HEAR the same sound, which will convey the same basic effect, but without it being realistic. The gameplay mechanic still works, because you know the sound may alert enemy forces and so you try to avoid interacting with plants, when enemies are near you, but you just don't see the proper effect.
Click to expand...

But I would guess in your example, the flappy leaf dynamics are simply eye candy, and the sound trigger handled separately, even if the engine's physics subsystem is used to do the initial collision detection. That's fundamentally different to the previous example where individual, physically-simulated grenade fragments were directly responsible for secondary effects. Flappy leaves still have no effect on gameplay, so it doesn't matter how realistically they're simulated, if at all.
 
Let's go back to text-based games, then? After all, graphics are only useless eyecandy. Nobody needs graphics for realistic gameplay. In fact, I just heard of this one new MUD which seems really great...


:)
 
Elledan said:
Let's go back to text-based games, then? After all, graphics are only useless eyecandy. Nobody needs graphics for realistic gameplay. In fact, I just heard of this one new MUD which seems really great...
:)
Click to expand...

Where do you find an ascii accelerator card? Really - Elledan has a point - it is all eye candy when you think about it. The more the better. What's so wrong with realism? I don't care what method and technology is used - let's get it coded into the games.
 
Mega666 said:
Where do you find an ascii accelerator card? Really - Elledan has a point - it is all eye candy when you think about it. The more the better. What's so wrong with realism? I don't care what method and technology is used - let's get it coded into the games.
Click to expand...

QFT. I don't see anybody else producing the same results or even making an attempt to bring more life-like environments like physx is. If it's specifically designed for nvidia, that's the card I will get. You don't see people complaining about console games only being able to be played on a ps3 or only on a 360(Well, sometimes you do :p )
 
For me, I don't care if only Nvidia does something, but I really want physics to take off and be successful (I know that it will no matter what). I just don't see it being THAT successful until there is an open standard. How many times have we seen one-brand proprietary video effects not succeed? Glide? TruForm? Hack-HDR?

Until there is a standard, I believe that it will be mired in a mixed-bag of graphic-only physics. And I don't think there will be standard until it's available to all cards. We'll just continue to get a mix of games that do Havok and Physx.
 
Whom holds the cards for a physics api? I hate to say it but - Microsoft. Once MS gets/adapts a physics api (whatever that may be) into DX - we are there. Until then everyone is outside looking in.
 
Mega666 said:
Whom holds the cards for a physics api? I hate to say it but - Microsoft. Once MS gets/adapts a physics api (whatever that may be) into DX - we are there.
Click to expand...

Why would MS put a physics API into DX?
DirectX isn't middleware, and shouldn't be.
Microsoft has never developed middleware, and I don't see why they would start now.
They most certainly have not announced anything in that direction so far.

So why do people put their hopes on Microsoft?
 
Scali2 said:
Why would MS put a physics API into DX?
DirectX isn't middleware, and shouldn't be.
Microsoft has never developed middleware, and I don't see why they would start now.
They most certainly have not announced anything in that direction so far.

So why do people put their hopes on Microsoft?
Click to expand...

I NEVER said develop. I said accepted as a standard in dx. Comprende?
 
wow, this thread has become quite interesting. appreciate all the constructive comments, whether positive or negative in regards to the technology in general. i imagine polarization is to be expected on a topic such as this. i'm sure everyone would agree that more realistic and immersive physics is the next necessary evolution in gaming. it is starting to grow now, and i wonder how much potential it has to explode if the next generation of consoles start to support hardware-accelerated physx instead of through software with this current generation. but getting back to cryostatis, what do you guys think of the game so far (well the tech demo, at least)?
 
Mega666 said:
I NEVER said develop. I said accepted as a standard in dx. Comprende?
Click to expand...

MS doesn't 'accept' anything. DirectX is what they develop.
So either they develop a physics API, or there will be no physics in DirectX.
The Not-Invented-Here mentality is strong within Microsoft.
 
Scali2 said:
MS doesn't 'accept' anything. DirectX is what they develop.
So either they develop a physics API, or there will be no physics in DirectX.
The Not-Invented-Here mentality is strong within Microsoft.
Click to expand...

I'd rather talk about concepts and ideas than definitions. This is a concept of physics being integrated into an/some API in some form or manner. Possibly (hopefully) - making it "standard" so to speak. It's not like M$ has written the "laws of physics" : )

Was "accept" to specific for you? I thought " (whatever that may be) " was as vague as I could get.
 
Well, I think you're trying to say that you hope Microsoft will integrate some kind of physics support into DirectX.
I'm saying that the chances of that are very slim. Physics are a level above what DirectX is aiming for.
Aside from that... Microsoft would probably have added physics a long time ago if they had any plans of doing so... Havok has been around for about 10 years now?

Bottom line: It's not going to happen.
You probably understood me already, but were hoping my answer would change if you rephrased the question. It doesn't.
 
Scali2 said:
Well, I think you're trying to say that you hope Microsoft will integrate some kind of physics support into DirectX.
I'm saying that the chances of that are very slim. Physics are a level above what DirectX is aiming for.
Aside from that... Microsoft would probably have added physics a long time ago if they had any plans of doing so... Havok has been around for about 10 years now?

Bottom line: It's not going to happen.
You probably understood me already, but were hoping my answer would change if you rephrased the question. It doesn't.
Click to expand...

You're the one making huge assumptions.
 
Just to throw this out there - Physix on OpenCL?

CUDA and OpenCL are just two camps of GPGPU where CUDA is closed. By the looks of things Intel will put Havok on OpenCl on Larabee. And AMD will do the same. NV say they will make CUDA compatible with OpenCL making CUDA its superset. So it is possible with InteL/AMD + apple combining weight pushing OpenCL NV will have to put physix on OpenCL. Eventually giving up on CUDA. (slim chance I know)

With OpenCL available to every GPU sold from here on devs may also get third options where they make their own HW physics with openCL. Maybe even cryengine3.
 
Intel will put Havok on Larrabee, no doubt... But why would they use OpenCL when they can do it natively? In fact, chances are that the first generation of Larrabee is not as fast as the AMD and nVidia GPUs of that time, so making it run on OpenCL means the competitors will get better Havok performance for free.

Likewise, why would nVidia change PhysX to OpenCL? It already works fine with Cuda, there's nothing to gain from moving it to OpenCL.

It doesn't make sense for either company to use OpenCL in their physics libraries.
 
Scali2 said:
Intel will put Havok on Larrabee, no doubt... But why would they use OpenCL when they can do it natively? In fact, chances are that the first generation of Larrabee is not as fast as the AMD and nVidia GPUs of that time, so making it run on OpenCL means the competitors will get better Havok performance for free.

Likewise, why would nVidia change PhysX to OpenCL? It already works fine with Cuda, there's nothing to gain from moving it to OpenCL.

It doesn't make sense for either company to use OpenCL in their physics libraries.
Click to expand...

Simply an hopeful extrapolation on OpenCL claiming to be... well... open!:p
 
Scali2 said:
Yea, open to all competitors :)
Manufacturers hate fair competition :)
Click to expand...

Well Adobe is going to be the king maker. Currently PS is CUDA so if they switch to OpenCL Intel will probably follow. NV may have gotten EA but AMD has Blizzard too. So it's business/scientific computing that is the real battle ground.
 
Well, OpenCL support from all GPU makers which matter (nVidia, AMD) will take until 2010 (yay for AMD dragging its heels...), so that's still a while off. Enough time for CUDA and PhysX to continue its conquering of the hearts & minds of developers.

My company has settled on PhysX and CUDA simply because they're the best options for us. The fact that you really need an nVidia card for them doesn't really matter. Businesses like nVidia cards more in general anyway (personal observation, before you all dive on me), so in the end it all works out, especially when developing multi-thousand Euro applications to be sold to big companies.
 
There are companies which have settled on Havok, that doesn't mean anything actually.
 
Elledan reminds me of the guy who said he had worked in IT for 15 years, then decided to educate me on DDR ram by saying that DDR 400 actually ran at 400mhz and was doubled effectively to 800mhz so it could run in 1:1 ratio with a 800fsb.

Had a vague idea, but not a scooby of the details.

You get what i am saying here.

P.S Does this count as a personal insult?
 
alg7_munif said:
There are companies which have settled on Havok, that doesn't mean anything actually.
Click to expand...

Oh sure, if they don't feel that hardware accelerated physics is important to them, then they're more than welcome to go with Havok. We felt it matters, so we didn't go with Havok.
 
As a customer, for me the content is far more important than technology. There a lot of games out there which use a high end technology with a moderate content and the sales are not as good as games with a great content but only use a moderate technology.
 
alg7_munif said:
As a customer, for me the content is far more important than technology. There a lot of games out there which use a high end technology with a moderate content and the sales are not as good as games with a great content but only use a moderate technology.
Click to expand...

As a coustmer innovation and hardqware support is the most important.
"the grapes are sour" spring to mind everytime you post.

Hell you even got a developer telling you that Havok just don't cut it.
But do you hear?
No, one place you are all up in arms about simple collision physics (it's good that you care about physcis though).
But when you get told told "Havok just don't cut it", you move the golepost, and suddenly you are playing "god of the gaps" in order to avoid supporting hardware physcis (because it dosn't run on you X2 card?)

Never mind, read enough trolling from you...iggy time.
 
A DEVELOPER said that and the world is now a completely different place? ;)

If hardware support and innovation is the most important thing for a consumer, I guess that the majority of computer user will have a highend CPU and GPU in their system and nobody will buy the mainstream hardwares anymore. :eek:
 
alg7_munif said:
As a customer, for me the content is far more important than technology.
Click to expand...

Of course, content is always far more important than technology. But we don't differentiate hardware based on content, we differentiate games based on that. So therefore, in a technology discussion content is irrelevant as the same content is available on everyone's hardware. That is, unless that content is tightly coupled to hardware accelerated physics effects. Which brings us back to where we started ;)

If what you're saying is that games with more advanced technology will automatically neglect content compared to less technologically advanced titles then that's simply a baseless claim. And the two aren't as distinct as you make it seem. You need technological advancement to bring that content to life....the best gameplay and storyline will fall flat on its face today if it's presented with 2002 graphics technology.
 
The GPGPU war has just begun, it will be a good 2 years before the clear winner is decided. CUDA will either become the Dreamcast or the XBOX360.
 
Hypernova said:
By the looks of things Intel will put Havok on OpenCl on Larabee.
Click to expand...

That's doubtful. Intel still wants to push its CPUs and offloading physics to the GPU isn't going to help in that regard. And if they do allow Havok to run on Larrabee it most likely will be a native implementation and not OpenCL.

With OpenCL available to every GPU sold from here on devs may also get third options where they make their own HW physics with openCL. Maybe even cryengine3.
Click to expand...

Yeah I think this is definitely a possibility. I won't be surprised if one of the big dev houses incorporates an OpenCL based hardware physics module in their next engine release.
 
alg7_munif said:
If hardware support and innovation is the most important thing for a consumer, I guess that the majority of computer user will have a highend CPU and GPU in their system and nobody will buy the mainstream hardwares anymore. :eek:
Click to expand...

In a way that's true. Most games require a highend CPU and GPU to be run at the highest detail settings and high resolutions + AA/AF.
You can run games on mainstream hardware, but you will have to turn down the detail level.
Having less detailed physics because you don't have powerful enough hardware is just part of that.
 
trinibwoy said:
Of course, content is always far more important than technology. But we don't differentiate hardware based on content, we differentiate games based on that. So therefore, in a technology discussion content is irrelevant as the same content is available on everyone's hardware. That is, unless that content is tightly coupled to hardware accelerated physics effects. Which brings us back to where we started ;)

If what you're saying is that games with more advanced technology will automatically neglect content compared to less technologically advanced titles then that's simply a baseless claim. And the two aren't as distinct as you make it seem. You need technological advancement to bring that content to life....the best gameplay and storyline will fall flat on its face today if it's presented with 2002 graphics technology.
Click to expand...

I'm not saying that games with more advanced technology will automatically neglect content compared to less technologically advanced titles, I'm just reminding Elledan that develepors should put more efforts in the content first before wasting time implementing a new technology in their games.
 
alg7_munif said:
I'm not saying that games with more advanced technology will automatically neglect content compared to less technologically advanced titles, I'm just reminding Elledan that develepors should put more efforts in the content first before wasting time implementing a new technology in their games.
Click to expand...

Wasting time? Perhaps you don't realize that every (3D) game engine needs a physics engine of some sorts to provide a semblance of realism (you can't pre-animate everything). We needed a physics engine for our game engine, which we ultimately wrote ourselves to make it fit with our custom game engine. We chose PhysX as the underlying API as it provided the best set of features, tools and documentation.

Having PhysX integrated into our game engine enables us to make even better content. Yes, more eye candy mostly, but isn't being able to crank everything up to 11 and feeling completely immersed in a game by more than just the gameplay a great thing?

Truth to be told, my favourite game is still Lufia II on the SNES. I absolutely love the story of that game to death. The graphics look very decent still (for a SNES game...), but it lacks virtually all of the things used to market current games (realism, physics, sound-effects, &c). Still, I'd place this game over any other game that has ever been released, so I think I understand what you mean by 'content' :)
 
Hypernova said:
Well Adobe is going to be the king maker. Currently PS is CUDA so if they switch to OpenCL Intel will probably follow. NV may have gotten EA but AMD has Blizzard too. So it's business/scientific computing that is the real battle ground.
Click to expand...

AMD has Blizzard for the GITG program. That has nothing to do with Physics APIs...
 
Hypernova said:
Just to throw this out there - Physix on OpenCL?

CUDA and OpenCL are just two camps of GPGPU where CUDA is closed. By the looks of things Intel will put Havok on OpenCl on Larabee. And AMD will do the same. NV say they will make CUDA compatible with OpenCL making CUDA its superset. So it is possible with InteL/AMD + apple combining weight pushing OpenCL NV will have to put physix on OpenCL. Eventually giving up on CUDA. (slim chance I know)

With OpenCL available to every GPU sold from here on devs may also get third options where they make their own HW physics with openCL. Maybe even cryengine3.
Click to expand...

They don't need to give up CUDA. OpenCL will be supported through CUDA.
Also Larrabee is a set of x86 cores and Havok will run on them just as it runs now, n current CPUs. That leaves AMD with Havok...on their CPUs aswell. Unless they can make a "Larrabee" like product, they'll be stuck with their CPUs to compute it.

Also, Apple is the main force behind OpenCL and NVIDIA was chosen by Apple for their new products. You take your conclusions from that.
 
Elledan said:
Wasting time? Perhaps you don't realize that every (3D) game engine needs a physics engine of some sorts to provide a semblance of realism (you can't pre-animate everything). We needed a physics engine for our game engine, which we ultimately wrote ourselves to make it fit with our custom game engine. We chose PhysX as the underlying API as it provided the best set of features, tools and documentation.

Having PhysX integrated into our game engine enables us to make even better content. Yes, more eye candy mostly, but isn't being able to crank everything up to 11 and feeling completely immersed in a game by more than just the gameplay a great thing?

Truth to be told, my favourite game is still Lufia II on the SNES. I absolutely love the story of that game to death. The graphics look very decent still (for a SNES game...), but it lacks virtually all of the things used to market current games (realism, physics, sound-effects, &c). Still, I'd place this game over any other game that has ever been released, so I think I understand what you mean by 'content' :)
Click to expand...

Thanks for understanding what I meant. I guess that what I want to say is, eventhough a developer doesn't think that hardware physics is important for them, it doesn't mean that they can't create a great game without it and eventhough a developer thinks that hardware physics is important to be used in their game, it doesn't mean that they will create a good game with it.
 
You must log in or register to reply here.
Back
Top