Nfs shift uses hardware acceleration !!!

I did some testing with the PhysX performance monitor http://developer.nvidia.com/object/agperfmon_home.html

http://img96.imageshack.us/img96/463...fmonui1250.jpg



I'm using PhysX driver 9.09.0814


http://downloads.guru3d.com/nVIDIA-P...load-2354.html



NOTE: NFS SHIFT INSTALLS 9.09.072 driver that doesn't work w/ AGEIA CARDS PERIOD.



AGEIA PPU is only in use when @ the TRACK

Blue light indicate cards is in use



This pic shows HARDWARE DIRECT MEMORY ACCESS

http://img17.imageshack.us/img17/841...wareaccele.png



This pic shows HARDWARE ULTILIZATION

http://img43.imageshack.us/img43/838...dwareaccel.png


This pic shows EVENT COUNT(Calls to ppu)

http://img707.imageshack.us/img707/5...ntcountave.gif


FPS didn't change but the driving is totally different.Tire's and damage are accelerated w/ the AGEIA ppu.:clap:

Can you define "driving is totally different"? Like the driving physics is different or just when you crash, it's different?

I feel the tire grip as if they were connected to me,when cornering this is crucial.:yepp:

Quote:

---

Originally Posted by Hell Hound

I feel the tire grip as if they were connected to me,when cornering this is crucial.:yepp:

---

whoa, that is awesome! i don't usually like driving games on the pc, but i might have to give this one a try now...

So, they cant do this on multicore CPUs?:rolleyes:

Quote:

---

Originally Posted by Rock&Roll

So, they cant do this on multicore CPUs?:rolleyes:

---

I don't think that this has to do with the cores but rather the calculations needed for this effect to occur.

I believe Cpu is not fast enough for dynamic physx,static yes but predetermined effects are not realistic.Medal of honor airborne shows this to be true,look @ how the rigid body effects look horrible in software mode.Bodies flying 30ft in the air and such.With hardware you can see the character arm and weapon bob in a more realistic manner and the bodies fall more naturally.:up:

Hmm. Might just need to steal back that 8800GT-256 from my sister, to use as a PPU when I upgrade my graphics setup. That sounds really cool to have non-eye-candy physics.

Quote:

---

Originally Posted by Hell Hound

I feel the tire grip as if they were connected to me,when cornering this is crucial.:yepp:

---

How does that actually work though? Physics, whether it's GPU or AGEIA, I thought they focus on effect physics. How would it actually affect the driving physics? I've played that game for about 20 hours with a G25, it felt pretty good. And if this was true, wouldn't having a PPU give you a big advantage racing online?

Quote:

---

Originally Posted by Hell Hound

I feel the tire grip as if they were connected to me,when cornering this is crucial.:yepp:

---

I don't understand.

Software only for multiplayer is your answer.If you can try it out and tell me what do you feel,that would be nice.One run w/ dedicated ppu,one run w/out it shouldn't take long.:up:

@munim I can tell how much pressure is needed for GAS/BRAKES much better.

Quote:

---

Originally Posted by Hell Hound

I feel the tire grip as if they were connected to me,when cornering this is crucial.:yepp:

---

My friend, I'm sorry to burst your bubble, but no PhysX gimmicks will transform the game physics (I mean how the car handles, to put it shortly) into something more "realistic" or something. It all boils down to engine coding, tyre behaviour moddeling, etc.

If you want acceptable, fun physics in Shift all you'll have to do is go to www.nogripracing.com and download a simple mod since this game is heavely based on ISI's gMotor2 engine (rFactor, GTR2, Race, GTL) and made by the guys who left SimBin. :up:

The PhysX api has Soft Bodies, Self-collision and damping support :shocked:;)

http://developer.download.nvidia.com...x/dev_soft.jpg

Quote:

---

Originally Posted by Hell Hound

The PhysX api has Soft Bodies, Self-collision and damping support :shocked:;)

http://developer.download.nvidia.com...x/dev_soft.jpg

---

It may have all that but it doesn't work like this.

An engine may be very powerful and allow a lot of things but you have to input a lot of data to it for the cars and tracks to behave accordingly. Depending on the data you feed it the game or mod can be very accurate or very arcadey.

Take gMotor2 for example. It's a very powerful, flexible and moddable engine despite it's age. There are, literally, hundreds if not thousands of mods for all the games based on it. Some are astounding, very accurate and better than the game itself. A lot are just good and even more are just plainly bad. There are even a few mods that aim deliberately for an arcade feel like the one that transforms rFactor in Looney Toons Racing! Again, it all depends on the data you feed the engine.

Try the Sharp Mod made for Shift available on nogripracing and you'll see what I mean. :)

Those are still static software accelerated vs dynamic hardware accelerated = in software every possible situation is not calculated,only predetermined.;)

Video of soft body effects @ 0:20 sec is what were talking about here. http://www.youtube.com/watch?v=gMQDP...ext=1&index=19 :up:

Quote:

---

Originally Posted by Hell Hound

Video of soft body effects @ 0:20 sec is what were talking about here. http://www.youtube.com/watch?v=gMQDP...ext=1&index=19 :up:

---

Carparroz is saying, PhysX is more for eye candies (ie dust, more movement on the trees, etc), and not for actual driving physics.

And from your replies, it's very vague to really conclude that PhysX modifies the driving physics (ie the game engine itself), sounds more like a placebo effect more than anything. I just don't see how PhysX can modify game engine itself, all the demos I see with PhysX or even the ATI physics demo, they are mainly collision detection, fluid, soft body effects, etc. From my understanding of it, this will change how things LOOK rather than how they actually behave to put in simplified terms.

The calculations of tire wear or suspension damage and engine damage are not seen but felt.:ROTF:

This is Damping

http://upload.wikimedia.org/wikipedi...ped_spring.gif

Quote from ShadowMyth on Guru3D.com nfs shift thread

Quote:

---

Maybe I'm wrong, the increase in fps could be from Hybrid SLI. My second gpu that's dedicated to Physx is defenitly being used. When playing the game its temperature goes up 12c & cools off once the game has been closed.

---

if the game engine features a modular design then it is possible that the physics library gets swapped out depending on whether there is any hardware support available. Just because physx has mainly been used for effects so far doesnt mean it cant be used for more, its a full physics API with all the tools necessary to do what the OP claims, whether it does so is another question.

Having played the game, i'm actually wondering what effects physx is being used for, there arent any real destructibles, explosions, cloth effects, etc... Maybe physx really kicks in once you have dedicated hardware for it? Its all conjecture on my part.

Soft bodies in racing/driving sims is nothing new, CPU's have been doing it since the guys who made 1nsane made their first game, which I forget the name of, at least that long anyways.

Only objects around the track are handled by PhysX. Under all the sugar coating and modifications the game it's still gMotor based, so driving physics are computed on your CPU.

If any one wants more sim like handling get this - NFS: Shift - Sharp Tires Mod 5.0

Quote:

---

Originally Posted by STEvil

Soft bodies in racing/driving sims is nothing new, CPU's have been doing it since the guys who made 1nsane made their first game, which I forget the name of, at least that long anyways.

---

Hungarian developers, Invictus. :)

Quote:

---

Originally Posted by Coldon

Just because physx has mainly been used for effects so far doesnt mean it cant be used for more, its a full physics API with all the tools necessary to do what the OP claims,

---

:yepp: Scalable terrain fedelity

Quote:

---

Delivers physical landscape accuracy and genuine surface reaction for true environmental experiences

---

http://www.bfgtech.com/physx/bfgphysx.html (under features & benefits tab) :up:

If your tires are not warmed up you may loose a second or two or If a piece of rubber is on the track and you drive over it you will feel it in your tire's.PhysX has potential.

Quote:

---

Originally Posted by Hell Hound

If your tires are not warmed up you may loose a second or two or If a piece of rubber is on the track and you drive over it you will feel it in your tire's.PhysX has potential.

---

I don't understand how you can feel things like these with your eyes through a PhysX enabled graphics card.

If you want to feel the tires you need to be in a car or on a seat with steering wheel with force feedback vibration system and you need of course the game to have support for all these things in the game engine.

Who said PhysX was limited to looks.Acceleration of the calculation of my suspension has nothing to do with looks or are you being funny. :shrug: Force feedback relies on data sent from the game to simulate the feel of the road,the data will be calculated by the ppu and sent to the steering wheel to simulate data given.

If you have more fun thinking PhysX works to alter game play then why care what other thinks? Why does one side works hard to put it down whilst the other tries to justify its existence? This is like Killer NIC all over again...

They have no proof of argument just talk,I'm stating what has been shown to be possible with the Api can change the way the game feels.We will see this in one form or another might not be with this API.They are acting like its not possible to change game play with this api but claim a mod can change game play.:shocked: THEY ALSO SAID THE WORLD WAS FLAT:rofl:The owner of the api doesn't matter its what it can do that matters.You want to kill it before it really gets utilized.

Quote:

---

Originally Posted by Hell Hound

If your tires are not warmed up you may loose a second or two or If a piece of rubber is on the track and you drive over it you will feel it in your tire's.PhysX has potential.

---

Tyre/brakes temperature altering the car behaviour has been around since the first Grand Prix game from the mid 90's, iirc. It's nothing new, very far from it. If you consider some other calculations like non-suspended mass and tyre modelling (the holy grail of race sim; there are books written about this subject!), tyre, brake and track temperatures are one of "simplest" things going on.

I'm a veteran race simmer and I know gMotor 1 and 2 engines from inside out. I can asure you that PhysX does nothing to change the game physics in Shift because it's impossible, really.

You don't have to believe in me but I ask you to go to some race-sim forums and ask some gMotor gurus/modders/developers and they'll tell you exactly what I'm trying to explain. :up:

Sure, PhysX support for actual car/suspension/tyre behaviour and track dynamics is very much possible, all I'm saying it's not possible for that in gMotor2, the engine powering Shift.

Edit/PS: I have nothing against PhysX and if it's possible for it to improve race sims, I'm all for it. After all, the only thing that comes to a real simmer when drinving on a virtual track is physics, physics, physics and to hell with everything else. :p:

Quote:

---

Originally Posted by Caparroz

I'm a veteran race simmer and I know gMotor 1 and 2 engines from inside out. I can asure you that PhysX does nothing to change the game physics in Shift because it's impossible, really.

Sure, PhysX support for actual car/suspension/tyre behaviour and track dynamics is very much possible, all I'm saying it's not possible for that in gMotor2, the engine powering Shift.

:p:

---

Physics @ 100% accuracy is what I want when racing.No need for a mod if its accurate,If the engine has been modded to use PhysX then there is more potential in shift than gtr2 @ least to me.

Quote:

---

I have nothing against PhysX and if it's possible for it to improve race sims, I'm all for it. After all, the only thing that comes to a real simmer when drinving on a virtual track is physics, physics, physics and to hell with everything else.

---

I would like to race you in gtr2 or something else if you like. :up:

Quote:

---

Originally Posted by Hell Hound

Physics @ 100% accuracy is what I want when racing.No need for a mod if its accurate,If the engine has been modded to use PhysX then there is more potential in shift than gtr2 @ least to me.




I would like to race you in gtr2 or something else if you like. :up:

---

Physics in Shift is 100% gMotor2. That's why all the physics mods available for Shift comes directly from parameters found in GTR2 and rFactor, iirc.

I wish that would be different, but a 100% accurate game or mod is impossible. :(

I don't race in GTR2 anymore. iRacing is all I have time for now. You should try it. It's miles ahead anything else on the market today.

Car list is limited.

This might be nice though

http://img299.imageshack.us/img299/2950/mustangq.png

Same guy who felt the difference with a 'gaming' NIC feels the difference with PhysX driving.
Just saying.

If you never used it how can you say anything about the product.:shrug:Reviews of gpus that they never used would be just lies.:shakes:Can we just talk about what physx can offer a sim racer and not this eye candy talk.

Sure we can.
If properly designed for it, a sim racer could see huge benefits with a PPU.
Unfortunately, this is not such a game.

Quote from YodaStar @ nogripracing.com

Quote:

---

"Live HDVehicle trace enable" traces car handling or such and outputs it to a log file in the game folder.

The "<prop name="Disable PhysX switching" type="Bool" />" is just the ID for the engine to understand the command. It just defines the type of value boolean, float etc.

For the "<prop name="Disable PhysX switching" data="true" />" you could check for perfomance impact when colliding with cars or track objects. Disable vsync, set graphic settings to low aswell as resolution and check framerate with for example FRAPs.

---

quote from East @ nogripracing.com @ same thread

Quote:

---

Looks like I spoke to soon. Changing it to false I do notice a slight increase in overall frame rates. I went back and forth a few times and it's noticeable to me. I don't see any difference in IQ. Can someone else try it?

---

Can anyone else say "placebo effect"?

Quote:

---

Originally Posted by Hell Hound

Quote from YodaStar @ nogripracing.com



quote from East @ nogripracing.com @ same thread

---

Quote:

---

Originally Posted by Hell Hound

Quote from Jonpo1 @ nogripracing.com



Testing is the only way to see and testing is what we are doing not just talking about it,this is not aimed @ anyone.

---

Gee, sounds like PERFORMANCE enhancements to me, which no one here is doubting.

I can't believe I'm even "arguing" about this.

Quote:

---

Originally Posted by Hell Hound

FPS didn't change but the driving is totally different.Tire's and damage are accelerated w/ the AGEIA ppu.:clap:

---

Quote:

---

Originally Posted by Hell Hound

I feel the tire grip as if they were connected to me,when cornering this is crucial.:yepp:

---

EDIT a bit off topic,
but this is all saying I can pop in my 8800gt to use as a PPU, even though my primary adapter is ATI?
Does it heat up the GPU much? Everything else is liquid cooled, I'd hate to buy another block, but stocker on low fan would be fine if it could keep up...

Quote from Jonpo1 @ nogripracing.com

Quote:

---

I decided to create a hybrid physX setup on my machine using my old 9800GT card as a PhysX hardware accelerator. Yes, I know that NFS Shift is reported as using the CPU for PhysX, but my testing results say otherwise, at least if you are running a Radeon card with an Nvidia card just for PhysX.

Anyway, after noticing a subjective improvement in the frame rate tanking whenever any of the visual physics effects came into play, I decided to check with FRAPS, just in case this was a placebo effect.

Well, according to FRAPS, it wasn't - the game was actually running better.

In a test I did at London, for example, I took the frame rate with vsync off when I was in the :banana::banana::banana::banana:pit at the back of the grid awaiting the race start. All the game settings were at medium, except for textures which were high. Antialiasing and anisotropic filtering were both set to 8x (box mode for antialiasing). The race was setup with 5 opponents and I of course was at the back of the grid. I also measured the lowest frame rate during the one lap race.

So here are the results:

Radeon 5870 only:

FPS at back of grid awaiting start: 91
Lowest FPS recorded: 28


Radeon 5870 plus 9800GT for PhysX processing:

FPS at back of grid awaiting start: 104
Lowest FPS recorded: 38


I repeated these tests about 10 times and the results were consistent - the above are averages, though there was little variation between averages.

---

Testing is the only way to see and testing is what we are doing not just talking about it,this is not aimed @ anyone.

@ first everyone was saying its cpu only now we can say that's not true and also find out what else is happening,I'm not arguing I'm talking about hardware accelerated physics and the effect it can have on sim racing.:up:

Quote:

---

Originally Posted by Hell Hound

I'm talking about hardware accelerated physics and the effect it can have on sim racing.:up:

---

Exactly: it can affect the graphics, FPS, and little stuff, which no one here doubts (yes, thank you for the news, it doesn't sound well known). But no, adding a PPU is not going to change the "feel" of the tires, make the [non-visual] engine damage more realistic, etc... all that is just absurd.

But it can we haven't tried yet.It seems to me you would not use gpu @ 100% all of the race so heat should not be a problem but that is single slot so I'm not sure sorry.

Quote:

---

Originally Posted by Hell Hound

But it can we haven't tried yet.It seems to me you would not use gpu @ 100% all of the race so heat should not be a problem but that is single slot so I'm not sure sorry.

---

I think we all agree that PhysX CAN change the driving physics, but only IF the game engine is designed to and allows it. Caparroz seems to be a veteran when it comes to the gMotor2 game engine, and he noted that PhysX cannot modified the game engine.

Yes, collision dmg, debris, etc, can/may be accelerated with PhysX in the game, but improving/making the game engine (ie driving physics) is impossible in Shift.

Some quick searching also turn up this thread:
http://forums.nvidia.com/index.php?s...dpost&p=938354

Quote:

---

Need for Speed Shift is also using PhysX. But it only uses it on the CPU

---

Quote:

---

Originally Posted by 570091D

whoa, that is awesome! i don't usually like driving games on the pc, but i might have to give this one a try now...

---

Have you played GRID or Dirt 2? I used to share your sentiment not finding any good PC racing games since NFS Porsche until I discovered the Codemasters games. :)

@ thephenom I quoted some people @ nogripracing.com that say they have activated hardware acceleration.;)

Wizzard can you include this in your PhysX benchmarks.:)

*reads thread*
:rofl: :rofl: :rofl: :rofl: :rofl:





PS: Seriously though. If PhysX were actually DOING something here that even comes close to what you are saying.

NVIDIA WOULD BE ALL OVER IT.

Nvidia does not control what the devs do,The api is there but the instruction calls for what we want might not be.Writing instructions to call for what we want will not be like making the game.:up:

Quote:

---

Originally Posted by BieBiep

*reads thread*
:rofl: :rofl: :rofl: :rofl: :rofl:





PS: Seriously though. If PhysX were actually DOING something here that even comes close to what you are saying.

NVIDIA WOULD BE ALL OVER IT.

---

I'm not sure why this is in the news section. From the previous examination of this before and after the update clearly show no GPU physx "enhancments" for the game. :yepp:

Did you see the first post buddy there's no denying what that chart says hardware DMA is being used and the ppu lights up.Stop talking a show real proof as I did.:up:I will tell wizzard every step I took to enable it if he wants to try and debunk this,do you have access to a ageia ppu or a ati gpu + nvidia gpu,if not you cant contribute to this conversation,we need more testers not haters.

Quote:

---

Originally Posted by Hell Hound

Did you see the first post buddy there's no denying what that chart says hardware DMA is being used and the ppu lights up.Stop talking a show real proof as I did.:up:

---

I think it maybe the case less lag in the physics calculations that could give a different feeling but not a change to the physics itself.

Quite possible.:up:

Quote:

---

Originally Posted by Hell Hound

Did you see the first post buddy there's no denying what that chart says hardware DMA is being used and the ppu lights up.Stop talking a show real proof as I did.:up:I will tell wizzard every step I took to enable it if he wants to try and debunk this,do you have access to a ageia ppu or a ati gpu + nvidia gpu,if not you cant contribute to this conversation,we need more testers not haters.

---

It's obvious that you cannot confirm the information on the chart. Others have tested this theory some time ago and have determine that there is no in game effect of GPU physx. Since you are still asking for more testers you only contradict the very results you want me to deem creditable. If that could stand on it's own you wouldn't need further evaluation. And unfortunately, saying that the car corners different is way to subjective to be taken as concrete proof at this time.

Which for the most part is why I ask this is in the news section for? You, the OP are asking for testers in the New Section. This should be moved.

I believe test were not done in a configuration like this.So other users testing is needed to see what other types of setup can enable it,as well see if performance differences occur.Which has been shown to be true,other types of setups (hybrid setups) can enable hardware acceleration and a performance difference has been seen as well.:up:The thread is about Hardware acceleration,which was/is thought to be Cpu accelerated only.Thats news.

Quote:

---

Originally Posted by Hell Hound

I believe test were not done in a configuration like this.So other users testing is needed to see what other types of setup can enable it,as well see if performance differences occur.Which has been shown to be true,other types of setups (hybrid setups) can enable hardware acceleration and a performance difference has seen as well.:up:The thread is about Hardware acceleration,which was/is thought to be Cpu accelerated only.Thats news.

---

It is my understanding that hardware accelerated is GPU. While software accelerated is CPU.

AGEIA ppu is what I'm using so that is why I wanted to know if people could use there gpu's for acceleration too.I found 2 user who have activated gpu acceleration.Remeber some physx titles are only accelerated with the ageia ppu.:up:

Quote:

---

Originally Posted by Hell Hound

AGEIA ppu is what I'm using so that is why I wanted to know if people could use there gpu's for acceleration too.I found 2 user who have activated gpu acceleration.Remeber some physx titles are only accelerated with the ageia ppu.:up:

---

You have to separate the distinction between the PPU & the GPU acceleration in your threads as option of use is vastly different.

If a game does not have menu option to turn on hardware physx then GPU physx can not be used.

Quote:

---

Originally Posted by Hell Hound

Quote from Jonpo1 @ nogripracing.com

I decided to create a hybrid physX setup on my machine using my old 9800GT card as a PhysX hardware accelerator. Yes, I know that NFS Shift is reported as using the CPU for PhysX, but my testing results say otherwise, at least if you are running a Radeon card with an Nvidia card just for PhysX.

Anyway, after noticing a subjective improvement in the frame rate tanking whenever any of the visual physics effects came into play, I decided to check with FRAPS, just in case this was a placebo effect.

Well, according to FRAPS, it wasn't - the game was actually running better.

In a test I did at London, for example, I took the frame rate with vsync off when I was in the pit at the back of the grid awaiting the race start. All the game settings were at medium, except for textures which were high. Antialiasing and anisotropic filtering were both set to 8x (box mode for antialiasing). The race was setup with 5 opponents and I of course was at the back of the grid. I also measured the lowest frame rate during the one lap race.

So here are the results:

Radeon 5870 only:

FPS at back of grid awaiting start: 91
Lowest FPS recorded: 28


Radeon 5870 plus 9800GT for PhysX processing:

FPS at back of grid awaiting start: 104
Lowest FPS recorded: 38


I repeated these tests about 10 times and the results were consistent - the above are averages, though there was little variation between averages



Testing is the only way to see and testing is what we are doing not just talking about it,this is not aimed @ anyone.

---

Gpu physx acceleration for you right here.:up:

What if the PhysX processing is being done on the CPU? Did you monitor task manager?

PhysX runtimes come with demo installer.... So what this is? nVidia fooled again someones that NFS shift doenst use PhysX? Hahaha!!

Nice PhysX program... havent notice that is exist. Very nice. Im used long time Ageia PPU With my ATi´s (X3)

Quote:

---

Originally Posted by STEvil

What if the PhysX processing is being done on the CPU? Did you monitor task manager?

---

You mean simultaneously or do you want to know how much cpu is used in software mode.My understanding is when ppu is in use the program is being hardware accelerated only.

Quote:

---

Originally Posted by Hell Hound

You mean simultaneously or do you want to know how much cpu is used in software mode.My understanding is when ppu is in use the program is being hardware accelerated only.

---

Both.

Software

http://img686.imageshack.us/img686/6...twareaccel.gif


Hardware

http://img405.imageshack.us/img405/1...dwareaccel.gif

http://www.iq.harvard.edu/blog/sss/ci700332kn00001.gif

Here we go again.Get yourself a dedicated ppu and stop hating on the rest of us.:rofl:

Why?

Why do you even try to talk to him?


If you read ANY of his threads you'll see how deluded he is. Read ANY of his comments, and you'll see how fanatically he defends the most absurd claims. If you read any reviews on the aforementioned products (Killer NIC 2100 and this PPU thing), you'll either get devoted fans with the same taste in 'performance products,' which really bring new features to this innovative and creative period of gaming, or, people saying the product is utter :banana::banana::banana::banana:.


Don't talk to him. It furthers his delusion and angers others when they realize, but can't change, his views.

Why call people deluded as if your views are any better than the next person.And a claim is when you have no evidence,which is what most of you do claim this claim that.:rofl:

I dont see any task manager or cpu load monitoring progams happening.

Are you saying there's no software physics happening like whats point the load up the demo dude.It will ask you to install physx driver bla bla bla.I don't care about cpu physx only hardware that's what the thread is about.

No, that is not my statement at all.

My statement is there are no CPU load monitoring graphs and because of such we cannot be absolutely sure that the physics/physx code is not being run on the main system CPU. Its obvious some hardware access to the additional PPU (be that an ageia processor or an nV one) is happening, but that does not tell the full story.

By your replies you are either avoiding my statement or just dont have a grasp of how PhysX works.

I don't know what Hell_Hound is talking about in any of his posts in this thread. I've seen euros at least 4 times as coherent as this.

I'm even beginning to suspect he doesn't know what he's talking about either. It's the only reasonable explanation.

I think it's just that game is utilizing multicore CPU better and thus, getting better results. I don't know why would a game use hardware PhysX and not make a big deal out of it. It just doesn't make sense considering how they're always boasting about PhysX in every game they release.

Did you read that THE AGEIA PPU LIGHTS UP WHEN IN USE

This doesn't make any sense show me that it isn't which you cant,get you some hardware and prove me wrong.:ROTF:

Fact is my ppu is in use,you dont have one so stop complaining about it.:p:

@ iddqd prove I don't know what I'm talking about.

This is not a TWIMTBP title nvidia can't tell them how to use the api it doesn't even say physx on the box,but it installs physx drivers.

No advertising physx from EA

Quote:

---

Originally Posted by Hell Hound

Did you read that THE AGEIA PPU LIGHTS UP WHEN IN USE

This doesn't make any sense show me that it isn't which you cant,get you some hardware and prove me wrong.:ROTF:

Fact is my ppu is in use,you dont have one so stop complaining about it.:p:

@ iddqd prove I don't know what I'm talking about.

This is not a TWIMTBP title nvidia can't tell them how to use the api it doesn't even say physx on the box,but it installs physx drivers.

No advertising physx from EA

---

http://www.nzone.com/object/nzone_tw...gameslist.html
http://www.nzone.com/object/nzone_shift_home.html

Looks like Shift IS a TWIMTBP title after all.

I'm not sure what you don't get. STEvil wants to know if the original CPU physics is offloaded to the PPU. By having a CPU utilization log with PPU vs without, that would show whether it's offloaded (ie lower CPU utilization with PPU would show hardware physics is used).

People are VERY skeptical about this since this is a TWIMTBP title, and if PPU does accelerate the game or "change' the driving physics, NV would be ALL over it, marketing the crap out of it boosting the usefulness of a PPU.

EDIT: Also, a simple test to see if a game uses CPU or GPU, you can go into NV control panel and in the 3D settings, you check the Show PhysX Visual Indicator. It'll show in-game whether it's using CPU or GPU.

Its not on the box that's for sure,where on that page does it says physx support for this title nowhere.I showed software utilization and Hardware,and you still dont acknowledge that ppu only lights up in use do you.:shrug:

edit:That's not the way I enabled it,there is quite a few steps it takes to use the newer driver with an ATI gpu.I mixed dll's from diff drivers for better performance also.

So Shift doesn't use PhysX hardware acceleration with the supplied Nvidia drivers but it does with the hacked ATI+PhysX drivers? Lol, that's rich.

That's the only way I can enabled it with my gpu,also notice the poster from nogripracing.com has a ATI+Nvidia for physx setup.

Well if the hacked drivers are unlocking hardware acceleration for Shift, I wonder how many other games it would work for.

Quote:

---

Originally Posted by Hell Hound

:yepp: Scalable terrain fedelity http://www.bfgtech.com/physx/bfgphysx.html (under features & benefits tab) :up:

If your tires are not warmed up you may loose a second or two or If a piece of rubber is on the track and you drive over it you will feel it in your tire's.PhysX has potential.

---

Hell Hound

Bro, I really don't think you understand what you are talking about. I have read all your posts and you seem to think the whole game is physics.

Pavement & friction? Do you understand what a Physics API is, does? Or, the difference between physics & an algorithm? Do you even understand why the compression of spring rate or dampening, would NEVER need the use of PhysX to resolve... etc. It appears that you seem to think that a physic's library (ie: PhysX) is anything calculated in a game. It is not.


Additionally, I though it was made quite clear to you (in another thread), that nobody here is discussing Scientific Computing.

Using hardware (PhysX or Physics), where people render a physical event, to illustrate an effect, scenario, or law of physics, etc. Where someone in a lab, used a bevy of Nvidia cards in SLI, utilizing PhysX to help them render something (non game related). Like your Frog demo... that is nothing more than illustrating how powerful, or robust the API is @ scientific computing, not gaming!


Ironically, the PhysX doesn't "accelerate" anything. "PhysX" can be accelerated... but real time, is real time... , so either it works, or it doesn't!


Coincidentally, whether or not Need For Speed utilized the PPU, CPU or GPU is irrelevant to this discussion, because it is safe to say that you DON'T need an Nvidia card to play the game, nor calculate the PhysX being used. Your own links demonstrate this.

It could easily just use a few idle cores of a modern CPU, with the same result.


Lastly, almost all of us understand what "physics", in a game are, what it does... we are saying that PhysX Only on the GPU is to protect Nvidia (so they can sell more video cards) and is an illusion created by Nvidia.

lol reading comprehension is bad here,I was talking about the possibilities of using terrain fidelity.Second physics is physics the way you use it is your choice,and yes the whole game is about physics its a driving game.Now what part of that is being accelerated by physx,not engine but collision and objects around the track and the handling.

Quote from YodaStar @ nogripracing.com

Quote:

---

"Live HDVehicle trace enable" traces car handling or such and outputs it to a log file in the game folder.

The "<prop name="Disable PhysX switching" type="Bool" />" is just the ID for the engine to understand the command. It just defines the type of value boolean, float etc.

For the "<prop name="Disable PhysX switching" data="true" />" you could check for perfomance impact when colliding with cars or track objects. Disable vsync, set graphic settings to low aswell as resolution and check framerate with for example FRAPs.

---

Quote:

---

whether or not Need For Speed utilized the PPU, CPU or GPU is irrelevant to this discussion

---

The thread is about hardware accelerated physx but people keep saying cpu cpu cpu :shakes:

It seems you don't know that its take computational power to game,many many calculation's,how is calculating not scientific.The more real or accurate the program the more computational power is needed.Dedicated hardware is made to do specific calculation to simulate a more real or accurate program.Who said damping calculations was depended upon physx that's stupid to even say that.I said PhysX supports damping calculation acceleration.I believe your understanding of what is being talked about is limited not trying to say anything wrong just saying that's all.;)

http://img101.imageshack.us/img101/4225/damping.png

The PhysX api can use hardware to more accurately accelerate the calculating of damping,when the steering wheel simulates data given it will be more accurate or real.




Someone's Grand Prix Legends Core.ini file settings

allow_force_feedback = 1 ; Use FF if device has it

force_feedback_damping = 200.0 ; force feedback damping coefficient

force_feedback_latency = 00.001 ; force feedback latency (secs)

max_steering_torque = 250.0 ; steering torque in N*in giving max device force

Hell Hound, I think you're running in circles here, mate.

The thing is, all the "effects" that you think PhysX or whatever API could bring to a racing or driving simulation is already being done for DECADES. Nothing that PhysX brings to the table is new per se.

Modern race-sims engines can be pretty CPU intensive and, theoretically, all the PhysX API could do is offload some calculations from the CPU.

Edit:

Sorry, mate, but you just proved that you don't know what you're talking about.

Spring and damper are just FFB effects that can be simulated by your wheel. And every simmer knows that both allways should be set to 0% and centering should be turned off aswell. ;)

That's what the whole thread is about hardware acceleration ,not bringing something new to racing sim's.

Quote:

---

Originally Posted by Hell Hound

That's what the whole thread is about hardware acceleration ,not bringing something new to racing sim's.

---

Which still remains to be proved.

Read my edit. It should improve your lap times. :up:

Quote:

---

Originally Posted by Caparroz

Which still remains to be proved.

Spring and damper are just FFB effects that can be simulated by your wheel. :up:

---

Because sim's didn't use it,where do you think the data comes from.I used that to point out damping being used on a game.:up:

Also why are people not acknowledging the ppu only lights up in use.

Get a ppu and you will see the light.:ROTF:

Quote:

---

Originally Posted by Hell Hound

Because sim's didn't use it,I used that to point out damping being used on a game.:up:

---

LOLWUT?!?

Do you even know what spring and damper Force Feedback effects are?

You do know that Grand Prix Legends is a sim made in 1997, right?

http://www.teamslm.com/downloads_momo.html

Damper Effects Strength - 0% very VERY important to set this to 0%, this is probably THE most important setting, as it will greatly increase notchiness and wheel twitches, even though the Papyrus sims do not exactly enable it

Quote:

---

Do you even know what spring and damper Force Feedback effects are?

---

The mechanical production of information sensed by the human kinaesthetic systems.

FORCE FEEDBACK EFFECTS are EFFECTS SIMULATED FROM GAME DATA TO GIVE A HUMAN THE SENSE OF REALISM.

Now Let me ask you do you know what force feedback is ?

hmm, perhaps you are correct, ill be picking up my new PPU today and installing it to get the full feeling in a game. team fortress 2 is feeling pretty bland.

Quote:

---

Originally Posted by AndrewZorn

hmm, perhaps you are correct, ill be picking up my new PPU today and installing it to get the full feeling in a game. team fortress 2 is feeling pretty bland.

---

Think if a pedals had force feedback,each car would drive in a more believable manner.

look @ this

http://www.frex.com/gp/hydro/hydroR1w448.gif

http://www.frex.com/gp/hydro/SCcup320.jpg

These are based off of oil pressure.

Quote:

---

Originally Posted by Hell Hound

http://www.teamslm.com/downloads_momo.html

Damper Effects Strength - 0% very VERY important to set this to 0%, this is probably THE most important setting, as it will greatly increase notchiness and wheel twitches, even though the Papyrus sims do not exactly enable it



The mechanical production of information sensed by the human kinaesthetic systems.

FORCE FEEDBACK EFFECTS are EFFECTS SIMULATED FROM GAME DATA TO GIVE A HUMAN THE SENSE OF REALISM.

---

Man, you're making such a mess and confusing two complete and absolute distintic things that I don't even know how I can awnser you. Sorry if I'm being harsh, I really do not intend to be, but it's like writinig a thesis called "French Fries and Gamma Rays Burts: a correlation between the two".

I'll try anyway...

GPL in it's glory days had a very freaking advanced FFB code. It blew everyone away. It's one of the best around even after 13 years of it's launch. But like every piece of software ever released it came with a "few" bugs.

The FFB tweak in the core.ini for the FFB to work like it should is widely known among GPL'ers. In the first few days after the lauch Dave Kaemmer (race-sim supremo, developer of GPL) himself acknowledge the bug in FFB and worked with the community and this very simple fix was born instead of waiting for a patch.

Why you think FFB an PhysX are correlated, or "FORCE FEEDBACK EFFECTS are EFFECTS SIMULATED FROM GAME DATA TO GIVE A HUMAN THE SENSE OF REALISM" as you put it, is really beyond me. :shrug:

I'm under the impression that you're thinking PhysX works like this:

A developer types "Air" in the PhysX developer tools and, BAM, there you have it, full drag/air resistance and flow accurate simulation inside the game. Type in "tyre and road" and you can thrown away Pacejka Formula for tyre/suface interaction between lateral and longitudinal forces.

Edit: My God, these set of Frex pedals is around for YEARS! What the heck does it have to do with PhysX?!? I had one of those when Athlon64 and X800 were braking world records, ffs...

I'm sorry for being harsh upon you but you keep thinking eye candy.This is not batman lol I don't want flying paper.Track data is simulated with force feedback,okay now if the physx api handles the track data where is the force feedback device getting its data from.?

Quote:

---

Originally Posted by Hell Hound

I'm sorry for being harsh upon you but you keep thinking eye candy.This is not batman lol I don't want flying paper.Track data is simulated with force feedback,okay now if the physx api handles the track data where is the force feedback device getting its data from.?

---

No, I didn't think you're harsh at all. :up:

Force Feedback is a very complex subject and is not by any means applied only in computer games. I'm no expert on that (you got to have an engineering/math degree to be an expert on that) and "interaction" with the virtual enviroment is only part of it. That's why only a handful of games deals properly with it. I suggest a google search with you're interested in that.

Back to PhysX. Virtual tyre/surface interaction isn't and it will never be handled by PhysX. Even if it would, FFB code would work completely independant of it.

what a mess. i can't get rid of the impression that the OP has no clue what he's talking about and how things work at all :p:

Quote:

---

Originally Posted by Caparroz

No, I didn't think you're harsh at all. :up:

Force Feedback is a very complex subject and is not by any means applied only in computer games. I'm no expert on that (you got to have an engineering/math degree to be an expert on that) and "interaction" with the virtual enviroment is only part of it. That's why only a handful of games deals properly with it. I suggest a google search with you're interested in that.

Back to PhysX. Virtual tyre/surface interaction isn't and it will never be handled by PhysX. Even if it would, FFB code would work completely independant of it.

---

Quote:

---

Originally Posted by Hell Hound

:yepp: Scalable terrain fedelity http://www.bfgtech.com/physx/bfgphysx.html (under features & benefits tab) :up:

---

surface reaction means what to you ?

Quote:

---

Originally Posted by RaZz!

what a mess. i can't get rid of the impression that the OP has no clue what he's talking about and how things work at all :p:

---

:p: :rolleyes: :hehe:

Quote:

---

Originally Posted by Hell Hound

surface reaction means what to you ?

---

PR talk.

Quote:

---

Originally Posted by Hell Hound

lol reading comprehension is bad here,I was talking about the possibilities of using terrain fidelity.Second physics is physics the way you use it is your choice,and yes the whole game is about physics its a driving game.Now what part of that is being accelerated by physx,not engine but collision and objects around the track and the handling.

Quote from YodaStar @ nogripracing.com


The thread is about hardware accelerated physx but people keep saying cpu cpu cpu :shakes:

It seems you don't know that its take computational power to game,many many calculation's,how is calculating not scientific.The more real or accurate the program the more computational power is needed.Dedicated hardware is made to do specific calculation to simulate a more real or accurate program.Who said damping calculations was depended upon physx that's stupid to even say that.I said PhysX supports damping calculation acceleration.I believe your understanding of what is being talked about is limited not trying to say anything wrong just saying that's all.;)

---

Dude,

In your latest posts, it is clearly demonstrated, that you have no idea what a physic's library is (ie: PhysX/Physics) when related to gaming..!


Games don't have real physical objects, so when making 3d games, how objects react to one another has always been left up to the developer. Sometimes they get their calculation right, sometimes, things seems "off", or not simulated well. (ie: characters dieing animation in a game)

That is where a set of software libraries, that run a separate sub-routine (if you will), that uses KNOWN physical attributes. So that when objects, applied using such a lib, behave correctly without intervention from the game developer..

Understand?


Now.. that^ is not the same as having a physical object (spring on a car) that has known dynamics. Which would never need to use a "specialized library" (physx) to compute the spring weight & rate while it compresses, because the environmental dynamics are already known. The Developer would just use mathematics & arrive with the same answer, without all the complexity.

Additionally, it is only when the desires effect is needed on a myriad of different objects variables... that you may need to make use of such "specialized library". That is when physx/physics is a better choice, because it can then be applied to every instance of that object, without the need to program it, it will function on it own, within the confines of whatever parameters it is given.

Hooking stable objects into a real time physical emulator would be a waste of processing power & utterly pointless. (ie: idiotic)


Ironically, many forum members here have reminded you numerous times that you have no clue, matter -of-fact I don't think I can find anyone that does agree with you on any of this.....




BTW, there is no such thing as software physics, everything is done on hardware. Whether it CPU, GPU, or dedicated PPU.. get your terminology right!

What I'm telling you is stop talking about every thing but hardware acceleration in this thread.

Also its you who does not know what physics is these calculation's are set in stone already,so to have a library of them is to have storage full of calculation's (data).anyway as I have showed the whole thread was about hardware acceleration in nfs shift anyway.

Funny thing is every one calls cpu acceleration software acceleration,and I bet I will see you say it as well.:clap:

Quote:

---

Originally Posted by Hell Hound

surface reaction means what to you ?

---

OK.. wait. That^ linked quote, is where you derived at Need for Speed using PhysX to simulate tire/tarmac cohesion...?

Then, you extrapolate that to also mean force feedback and spring rate/dampening, because a game can output spring rate/dampening to a Force Feedback device..?

:rehab:




?
er, how about..
The surface of the quarter panels as they hit the wall? The surface of the windshield as it cracks? The surface of the water as a pebble hits it?



There are thousands of surfaces in a racing game, but you took that simple quote and argued a whole point, because you thought they meant the track surface?

I would think that one of the most stable environment in a racing game would be the tarmac. Slip angles and lateral G's and all types of physical forces are already known. Even if applying a variable surface co-efficient, such as being hotter on one part of the track, over the next, even oil spill..

Because the tarmac & tires have known physical attributes, that given the depth of the game, can have many different surfaces... but no need to involve a physic.lib

Absolutely one of the funniest things I have ever read... along with my new sig, thnx!




{edit] BTW... the topic of this thread is "Need for Speed; Shift.. uses hardware acceleration !!!" and you have done everything BUT discuss the use of PhysX in the game. Also, libraries don't have calculatins set in stone, the FORMULAS are.. do you even know how to do geomatrey, let alone diff eq & such...?

You can not say that track surface isn't a surface that objects react to.

The PhysX api has Soft Bodies, Self-collision and damping support

http://developer.download.nvidia.com...x/dev_soft.jpg

This is how we got to damping acceleration.:up:

Quote:

---

Originally Posted by Xoulz

Dude,

In your latest posts, it is clearly demonstrated, that you have no idea what a physic's library is (ie: PhysX/Physics) when related to gaming..!


Games don't have real physical objects, so when making 3d games, how objects react to one another has always been left up to the developer. Sometimes they get their calculation right, sometimes, things seems "off", or not simulated well. (ie: characters dieing animation in a game)

That is where a set of software libraries, that run a separate sub-routine (if you will), that uses KNOWN physical attributes. So that when objects, applied using such a lib, behave correctly without intervention from the game developer..

Understand?


Now.. that^ is not the same as having a physical object (spring on a car) that has known dynamics. Which would never need to use a "specialized library" (physx) to compute the spring weight & rate while it compresses, because the environmental dynamics are already known. The Developer would just use mathematics & arrive with the same answer, without all the complexity.

Additionally, it is only when the desires effect is needed on a myriad of different objects variables... that you may need to make use of such "specialized library". That is when physx/physics is a better choice, because it can then be applied to every instance of that object, without the need to program it, it will function on it own, within the confines of whatever parameters it is given.

Hooking stable objects into a real time physical emulator would be a waste of processing power & utterly pointless. (ie: idiotic)


Ironically, many forum members here have reminded you numerous times that you have no clue, matter -of-fact I don't think I can find anyone that does agree with you on any of this.....

---

you dont know what you're talking about do you?

Quote:

---

BTW, there is no such thing as software physics, everything is done on hardware. Whether it CPU, GPU, or dedicated PPU.. get your terminology right!

---

HW physics refers to physics simulations that are run on an ASIC to boost performance.

SW physics refers to physics simulations that run on a cpu.

DID ANYONE READE POST NUMBER 66?

I can't believe you guys are actually talking to him. It's not a matter of right or wrong opinions, it's a matter of being open minded, and this guy is totally open minded in only one thing, the endorsement of fail products.