Meh, there will be a crossover point where precanned fake effects don't cut it any more. Before that point people will say that real-time simulation is just hype, lame, etc etc. I'm just glad somebody is taking steps to move the technology forward so when we get to that point all of us - including the current naysayers - will have better gaming experiences. 3D graphics was just hype at one point too.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Your Opinion on BFBC2 physics
- Thread starter Brackle
- Start date
Where have you proved me wrong? Nowhere is where. You state no $50 power hungry card can supply physx from nvidia. I'll be more than happy to bust out the links showing GT240s and 9600GTs being more than adaquet to perform the task busting that thory to shreds. You then change your story to no 8800GT/9800GTX can render and do PhysX at the same time. You are doing just as any fanATIc does, when busted out, you change the goal posts to suit your needs. I'm still waiting for the power hungry $50 card.
http://www.firingsquad.com/hardware/mirrors_edge_physx_performance/page7.asp
http://translate.google.com/transla...a_physx_ueberblick/&hl=en&ie=UTF8&sl=de&tl=en
Above is proof that a 9600GT is more than enough to run PhysX calls in games, I'll post more as I find them.
http://www.tomshardware.com/reviews/batman-arkham-asylum,2465-8.html
Again, proof of what I have said.
And FYI, one of the links I provided shows 8800GT/9800GTs doing rendering and PhysX just fine by themselves.
Last edited:
Do you think that everyone's CPU can handle computations for the game (esp. a multi-player online game) and CPU-processed physics?
You quote a relatively modest card, by today's standards, as being incapable of handling the extra physics load but seem to feel that there's oodles of CPU time sitting unused.
If you have a low end GPU - there's less likely to be spare cycles for GPU physics.
If you have a weak CPU - there's less likely to be spare cycles for CPU physics.
If you have low end both, you might as well turn physics off if you can run the game at all.
Again, everyone here expects something for nothing, and that non-PhysX physics implementations somehow consume no resources while PhysX is somehow always a hog....
Neither of those are $50 cards and even then both have been shown to be inadequate to run the highest PhysX settings in games, so he's right. And he's also right that no 8800GT/9800GTX can provide a good gaming experience while rendering and running PhysX. You're the one that has to prove differently, not him.
Thank you for the points you've provided.
Cellfactor was boring. GRAW2 I played in SP mode, no online. and I prefer reactionary PhysX to preprogrammed crap.
You dear sir need to look at the links I have provided. They proof not only your statements wrong, but also his.
necrophile
[H]ard|Gawd
- Joined
- Sep 6, 2008
- Messages
- 1,243
What is all this reactionary stuff. What do people mean by that?
Are you meaning something which is not scripted?
I really think some people have no thoughts of there own and only actually think they care what "physics" are in a game because they have been told to think that.
Sorry, but if a game developer wanted to, they could fool you into thinking physx or any type of physics effects was being done properly even though there was no calculations being done in the background. As all physics effects are just graphics, there is no way to tell what is "real" or "accurate" and if you think you could tell it apart from normal smoke, or normal "particle effects" then you must be gullible as hell
Sorry for so much quotation marks but so much of those words and the way they are used is pure BS.
Are you meaning something which is not scripted?
I really think some people have no thoughts of there own and only actually think they care what "physics" are in a game because they have been told to think that.
Sorry, but if a game developer wanted to, they could fool you into thinking physx or any type of physics effects was being done properly even though there was no calculations being done in the background. As all physics effects are just graphics, there is no way to tell what is "real" or "accurate" and if you think you could tell it apart from normal smoke, or normal "particle effects" then you must be gullible as hell
Sorry for so much quotation marks but so much of those words and the way they are used is pure BS.
SixtyWattMan
Supreme [H]ardness
- Joined
- Aug 4, 2008
- Messages
- 4,466
Destructable environments > tacked on effects
In your oppion.
I'd say if you can't tell the difference, you'd be the fool. Just as people can tell teh difference between 30, 60 and 120fps, there are those who can tell the difference between scripted, preprogrammed crap and stuff that acts realisticly.
Some good videos showing the Physics and destruction in BFBC2:
The Physics and Destruction of Battlefield Bad Company 2
Entire Buildings Collapse in Battlefield Bad Company 2 With Destruction 2.0
The Physics and Destruction of Battlefield Bad Company 2
Entire Buildings Collapse in Battlefield Bad Company 2 With Destruction 2.0
Having watched that, you can see that the 'destruction' is all very neatly organised into predefined holes in a structure. The explosions themselves are just a bit of shrapnel that disappears in a few seconds. It's very much smoke and mirrors as in-game you don't really notice how neat it all is when you're crouching behind a bit of crumbly wall with loud explosions going off and hot wee running down your leg.
SixtyWattMan
Supreme [H]ardness
- Joined
- Aug 4, 2008
- Messages
- 4,466
Bad Company 2 is fun though. As you can see, fun can be done on the CPU.
Its fricken scripted, there is no real physics going on. its all pre rendered to do exactly that after x amount of damage. How is that physics? it isn't.
Physics would be if you destroyed one side of the building and it colasped in that direction or if you took out the corner supporting structure then then another and it fell down according to what you did. Hell, call and air raid on the building and I'll bet it still falls in on itself instead of blowing up and out like it should with whatever is left falling.
SixtyWattMan
Supreme [H]ardness
- Joined
- Aug 4, 2008
- Messages
- 4,466
No all building destruction in BC2 is ray traced in real time using skynet.
Right and you believe it.
It takes a render farm the size of your home to render at 30FPS in raytracing, you honestly believe they somehow managed to get a complex "physics" scene, to render via raytracing on yours and everyone elses CPU, even those running poltry celerons, at the exact same speed so EVERYONE can enjoy the same fluid scene.GET REAL! Its a preprogrammed renderation of a building colaspe. They "MAY HAVE" rendered it using ray tracing, but there is no way in hell your machine is ray tracing that part of the scene when it happens, your frame rate would tank to single digits even if you had a skulltrail setup with your 5870.
I hope you're joking because it's very easy to tell the difference between a precanned graphical effect and a simulated one. Unless you don't know what real life looks like, then you're excused
This discussion is silly. BFBC2 is awesome but it's not due to awesome interactive physics, it's due to developer creativity as is usually the case.
I hope you're joking because it's very easy to tell the difference between a precanned graphical effect and a simulated one. Unless you don't know what real life looks like, then you're excused
This discussion is silly. BFBC2 is awesome but it's not due to awesome interactive physics, it's due to developer creativity as is usually the case.
Thank god there is a voice of reason in this discussion.
SixtyWattMan
Supreme [H]ardness
- Joined
- Aug 4, 2008
- Messages
- 4,466
Right and you believe it.
GET REAL! Its a preprogrammed renderation of a building colaspe. They "MAY HAVE" rendered it using ray tracing, but there is no way in hell your machine is ray tracing that part of the scene when it happens, your frame rate would tank to single digits even if you had a skulltrail setup with your 5870.
The Frostbite engine is fully ray traced and everything is completely destructible and it's all done in real time and can be done a dual core CPU
LanPirate1
Limp Gawd
- Joined
- Nov 7, 2004
- Messages
- 362
I don't think he got the skynet reference, heh.
As for the "scripted" holes in walls, at least its a step up compared to the COD series. Should make for an interesting game, especially if we can have private servers.
As for the "scripted" holes in walls, at least its a step up compared to the COD series. Should make for an interesting game, especially if we can have private servers.
Again, its not something that is capable of happening on todays hardware. raytracing requires huge farms to get smooth frame rates, you truely are guilable.
I have a feeling you just completely made this up
Edit - it's never going to be as fast on the cpu as the gpu, but don't kid yourself thinking they didn't remove all the effects to bundle into "Nvidia Phys-X".
necrophile
[H]ard|Gawd
- Joined
- Sep 6, 2008
- Messages
- 1,243
Seeing as there is no real physics needed for nvidia's graphical effects I am right on you would not know the difference.
I have never seen anything done which could not be done as purely a graphical effect. Again, you are beingf very gullible to believe that a bit of smoke and whatnot could not be done more "accurately" if a game developer thought it was worth the time and effort. It isn't.
If you want smoke to drift in the wind, then all you need is a seperate animation for a different wind direction, that is not physics.
Perhaps I am wrong to call it gullibility. It is clearly pure ego driven in that you people like to think you are smarter than most in understanding this "new fangled" physics PR.
I have never seen anything done which could not be done as purely a graphical effect. Again, you are beingf very gullible to believe that a bit of smoke and whatnot could not be done more "accurately" if a game developer thought it was worth the time and effort. It isn't.
If you want smoke to drift in the wind, then all you need is a seperate animation for a different wind direction, that is not physics.
Perhaps I am wrong to call it gullibility. It is clearly pure ego driven in that you people like to think you are smarter than most in understanding this "new fangled" physics PR.
Why would you write that out when it makes it clear you aren't understanding what physics are at all. Physics are the interactions between objects--how collisions are handled. Animations aren't physics, and PhysX is not simply animations.
That's the circular reasoning you guys keep employing that makes no sense: PhysX effects are simple and boring and therefore they are insignificant. They could be run anywhere because they are insignificant.
In reality, you're just wrong about whether the calculations are insignificant. You may not personally care that objects have true collision in a PhysX game, but that doesn't make it insignificant. Yes, Crysis has true collision without PhysX, but it also requires a ton of horsepower to play it in that mode and even more to play it in that mode online. The physics in Crysis blow away BFBC2 but there's a huge performance difference, too.
I don't see how anyone who is looking at their game with both eyes open with working depth-perception would argue that true collision is not qualitatively and visibly better compared to non-collision destruction.
That's the circular reasoning you guys keep employing that makes no sense: PhysX effects are simple and boring and therefore they are insignificant. They could be run anywhere because they are insignificant.
In reality, you're just wrong about whether the calculations are insignificant. You may not personally care that objects have true collision in a PhysX game, but that doesn't make it insignificant. Yes, Crysis has true collision without PhysX, but it also requires a ton of horsepower to play it in that mode and even more to play it in that mode online. The physics in Crysis blow away BFBC2 but there's a huge performance difference, too.
I don't see how anyone who is looking at their game with both eyes open with working depth-perception would argue that true collision is not qualitatively and visibly better compared to non-collision destruction.
At this stage you are correct. But I dream of a day where real time accurate physics is in games. A game like BF where the smoke drifts on the wind and a truck drives through it whipping up von karman vortex streets in it, a boat rushes through the water creating waves, a plane flies through the air and you can see the induced vorticies of its wing tips as it flies through the black smoke cloud of a burning jeep. Then you fire at a building with an assault rifle and tear chunks out of it realistically, fire at a fence or a tree and the wood splinters. A rocket flies into a building and the building sees destruction according to the actual stresses in the material and not just the same blocks falling off every time it happens. Cloth flutters in the wind without clipping like it does in most physx demos I've seen.
When game physics are that good, it'll be awesome and for someone like me, I'll most definately notice the difference compared to pre made animations.
Of course that's a long way off and requires a hell of a lot more computing power than we have now, but its just a dream, and games have to take baby steps with actually implementing some physics engines (opposed to animations) before my dream can come true.
personally its a feature that doesn't exist in the world of games now. The original co lost butt loads of money and never sold anything. Nvidea picked it up for peanuts and hyping it along.
1 out of 10
a few wow and ohhh on some demos doesn't mean anything.
1 out of 10
a few wow and ohhh on some demos doesn't mean anything.
How do you think the technology will ever progress if no one starts it along its way? If no one started 3D acceleration because all that existed was "wow oooh wow" demos for the Mach32 or Voodoo then we wouldn't be anywhere.
Think of nV accelerated PhysX as the Mach64 of PhysX, if that helps you.
On the topic of scripted vs. real effects, it's often pretty easy to tell. I'm not sure why people disagree. Some devs are better at hiding it than others.
necrophile
[H]ard|Gawd
- Joined
- Sep 6, 2008
- Messages
- 1,243
Mope54, you have missed the point entirely.
In fact I think most people miss the point over physics. It is not for fancy effects, that is a waste of resources. This is about efficiency. It is about getting the most impact out of the least performance.
Wasting resources on smoke and mirrors is not what physics is about, it should be more about the rocket flying into a building than wind blowing smoke, again you do not need physics to do any vortices or waves through water. That can all be done cheaply through gfx and you won't notice one bit of difference than if the dev's like and say it uses physics calculations to do it.
Ask yourself this, if the dev's of batman only said it was using physx and in actual fact it wasn't. How would you know?
In fact I think most people miss the point over physics. It is not for fancy effects, that is a waste of resources. This is about efficiency. It is about getting the most impact out of the least performance.
Wasting resources on smoke and mirrors is not what physics is about, it should be more about the rocket flying into a building than wind blowing smoke, again you do not need physics to do any vortices or waves through water. That can all be done cheaply through gfx and you won't notice one bit of difference than if the dev's like and say it uses physics calculations to do it.
Ask yourself this, if the dev's of batman only said it was using physx and in actual fact it wasn't. How would you know?
Please, do go into detail. Perhaps you have a research paper demonstrating this technique? I'd love to know how you produce static animations for every possible combination of wind direction and force. Vortexes should be fun too.
The rest of us fortunately do not suffer from your lack of imagination. The OpenGL/DirecX pipelines are all about solving a physics problem - the interaction between light and the environment. Why do developers waste resources on real-time dynamic shadows when they can give everything a static shadow like in the olden days? Let me guess, dynamic shadows are important but a dynamic environment isn't because necrophile says so?
You really need to stop with the "XYZ can be done cheaply with graphics" unless you actually have some knowledge on the matter. You can fake waves and smoke and any precanned animation but you can't fake the interaction of those phenomenon with other objects in the environment - that's the point that you seem to be intentionally missing. Ever tried walking through water in HL2?
necrophile
[H]ard|Gawd
- Joined
- Sep 6, 2008
- Messages
- 1,243
Lol, the useless arguments of trinibwoy.
You do not need to do a seperate animation for every instance. Only enough to make it look passable. The basis of computer games is fooling people into thinking something is when it isn't.
You keep up with this notion that everything has to be perfect yet we are happy to accept 2-d images as 3-d and play games fine enough. No doubt you will bring in "true" 3-d gaming into this conversation soon enough.
I also use the term cheaply as you do not need a seperate anything to be able to do the graphics that apparently use physics.
I fail to see how you cannot simulate an interaction with smoke and water the same as you simulate interaction with any other object in a game. It may not be completely accurate to do it that way but you will not know the difference if it looked alright. If you believe so then you are kidding yourself.
Again, if batman didn't use physx but you were told it did, how would you know?
You do not need to do a seperate animation for every instance. Only enough to make it look passable. The basis of computer games is fooling people into thinking something is when it isn't.
You keep up with this notion that everything has to be perfect yet we are happy to accept 2-d images as 3-d and play games fine enough. No doubt you will bring in "true" 3-d gaming into this conversation soon enough.
I also use the term cheaply as you do not need a seperate anything to be able to do the graphics that apparently use physics.
I fail to see how you cannot simulate an interaction with smoke and water the same as you simulate interaction with any other object in a game. It may not be completely accurate to do it that way but you will not know the difference if it looked alright. If you believe so then you are kidding yourself.
Again, if batman didn't use physx but you were told it did, how would you know?
Yes, and what was passable in the 80's isn't passable today. You're basically trying to convince yourself and others that we don't need progress beyond what we have today. I'm sure we can all agree that position is complete nonsense. It doesn't have to be perfect but we have a long, long way to go before we can sit back and say we don't need to keep progressing.
You fail to see it because you don't understand what interaction means. Rigid body interaction is done using very coarse bounding boxes. It's been the same way for a decade and is nothing computationally intensive. Interaction with a fluid requires you to simulate the interaction between 10's of thousands of particles and the environment including forces due to gravity, wind etc. If you can't grasp that simple difference then I will leave you to be happy with whatever it is you deem to be good enough while the rest of us move on. I'm not sure how your inability to perceive simple things translates to me "kidding myself".
Is that a serious question? How would you know that any game uses anything if developers didn't tell you? How would you know that AvP uses tessellation or Stalker:COP implemented contact hardening shadows?
necrophile
[H]ard|Gawd
- Joined
- Sep 6, 2008
- Messages
- 1,243
You don't need 10's of thousands of particles though. You need that for a physics demo, not for a game.
My point with batman is that that is hailed as a good use of physx. However there is nothing I have seen that needs it. Clearly we have the graphical ability to do everything that physx or whatever currently does.
So why would it need a lot of computational power to do it. This is what we are disagreeing on. You seem to think the solution is to throw power at it. To do things which are not necessary for a good gaming experience.
What I am saying is that if something looks good but takes a few shortcuts then there is no drawback compared to the same looks but with unnecessary precision.
My point with batman is that that is hailed as a good use of physx. However there is nothing I have seen that needs it. Clearly we have the graphical ability to do everything that physx or whatever currently does.
So why would it need a lot of computational power to do it. This is what we are disagreeing on. You seem to think the solution is to throw power at it. To do things which are not necessary for a good gaming experience.
What I am saying is that if something looks good but takes a few shortcuts then there is no drawback compared to the same looks but with unnecessary precision.
necrophile
[H]ard|Gawd
- Joined
- Sep 6, 2008
- Messages
- 1,243
Fine.
Although until we get to a point where we can replace the fakes I will stick with them.
I do not think we are there, nor do I think we will get there in the next couple of years. Right now I think we need a scaling ability.
If you want the 50 shards of glass all calclated precisely then fine. However, I will settle for 5 with the rest animated. I do not want it the glass just removed or made solid.
Although until we get to a point where we can replace the fakes I will stick with them.
I do not think we are there, nor do I think we will get there in the next couple of years. Right now I think we need a scaling ability.
If you want the 50 shards of glass all calclated precisely then fine. However, I will settle for 5 with the rest animated. I do not want it the glass just removed or made solid.
Things will advance far faster than you expect. Just compare GF100's SPH performance compared to GT200. 30k particles @ 30fps vs 128k particles @ 120fps. Add in faster context switching and support for running multiple kernels in parallel and you just got yourself far more compute power than we had last year. But as always, we'll need to see what devs can or choose to do with it.
necrophile
[H]ard|Gawd
- Joined
- Sep 6, 2008
- Messages
- 1,243
Hmm, taking into account lag time between the tech being released, actually seeing if it can do what they claim in the wild and consumers adopting it I ain't so sure.
Nvidia will need a company like crytek to drag that sort of thing kicking and screaming into existence and even if they use the power, whether other companies will follow suit is hard to say. They haven't with crysis so who is to say they will with physics.
Nvidia will need a company like crytek to drag that sort of thing kicking and screaming into existence and even if they use the power, whether other companies will follow suit is hard to say. They haven't with crysis so who is to say they will with physics.
Why do you say that? A lot of things pioneered in Crysis have made their way into more titles. SSAO for one. Parallax Occlusion Mapping is another. DOF, god rays etc are all things first introduced in Crysis but people cried about the performance hit. Yet here we are today and those things are becoming standard. It's the same cycle for all new techniques. There will be naysayers at first and then we won't be able to live without them.