[Nvidia.com] Nvidia GameWorks unleashed at GDC.

[VulgarDisplay]

PhysX is GPU-accelerated on systems with NVIDIA GPU's (which represents a huge install base), and it is CPU-accelerated on countless other systems. NVIDIA claims that PhysX is the most popular physics SDK on the market (used in more than 500 games), while also being included in major game engines such as UE3 and UE4 (among others): https://developer.nvidia.com/gameworks-physx-overview

They are referring to physx that is used for collision detections and fundamental engine physics and not GPU physx which separate.

 

[Keysplayr]

were it not proprietary it would gain more adoption...as most open things do. Nvidia could then have more games that use this technology but they would keep the prestige of introducing/developing it...Its all about perception, IMO that is.

Perception of what? The adoption rate of a feature? Or whether you can use it or not? IMO, the latter is where it matters.

As for adoption rate? PhysX, either GPU or CPU based, is currently in over 400 games and there are many currently under development along with GameWorks.
I think the adoption argument is non-existent to tell you the truth.

 

[wand3r3r]

Yeah, it really doesn't. I think we've come to the end of the road for the "proprietary arguments".

Why would you say that? That's a baseless conclusion simply because both now have something proprietary.

Simply because both have something currently unused by others doesn't mean that the average person now suddenly supports proprietary features. If the manufacturer's have chosen to create proprietary features it doesn't mean that makes the consumer happy, it may be to the contrary since now you are limited in what you get with the card.

 

[Keysplayr]

Why would you say that? That's a baseless conclusion simply because both now have something proprietary.

Simply because both have something currently unused by others doesn't mean that the average person now suddenly supports proprietary features. If the manufacturer's have chosen to create proprietary features it doesn't mean that makes the consumer happy, it may be to the contrary since now you are limited in what you get with the card.

It doesn't matter if it's supported or not. It is what it is. It's here. It's the way these companies tout value over the other (well, one of the ways).

Let me run a scenario by you all for a moment, then you tell me how you think it's going to turn out.

Nvidia creates a ton of features that will only work on their GPU's.
AMD creates a ton of features that will work on any GPU.(Intel, Nvidia, AMD)

Down the road, which will you choose?

 

[blackened23]

Its getting more press because it has broken years of stagnation on the api front, and is genuinely awesome for its users.

I fail to see how blowing newspapers and jello are earth shattering implementations of physx in bl2 and batman. As another poster said physics is held back by its proprietary nature. It can't change gameplay so its always going to relegated to fluff that you would not notice if it was missing. We are still waiting for the dust to settle from mantle. No one knows if it will always be proprietary or not, buy physx is staying that way.

Just so you know, since you brought this argument up, GPU accelerated physx is only a small subset of what physx does; it is actually a full blown physics engine that isn't proprietary - it works on all major gaming platforms. The actual physics engine portion is widely used and works on all hardware. That includes the xbox 360, ps3, pc, windows, android, you name it, physx works for physics engines, and it is widespread in use with nearly 500 titles. I could name an endless list of console 360, ps3, PS4 and XB1 titles that are using the physics engine portion of physx. Now that doesn't sound proprietary to me. You're focusing only on the gpu accelerated particle features, which does require cuda hardware. Again. Small subset of physx.

Conversely the open standard which AMD proposed - openCL physics - hasn't gained steam because AMD just doesn't have the resources to create developer tools that are compelling or desirable. Therefore it hasn't gained steam, despite having been around forever. This is what nvidia excels at IMO, software - they create developer tools that are easy to work with based on what i've read and because of this, they gain adoption. Physx has done very well in this respect, it is used in a ton of titles. Even looking at Intel's Havoc - it is all about developer tools. Just throwing something at a developer and stating "hey this is an open standard" means nothing. How does it make their job easier? How does the open standard make their life or job easier as a developer just by merely being an "open" standard? That does nothing. It seems in this case that the open standard failed. I could name more open standards that failed because of poor developer tools, but I won't get into that. This is not meant as any sort of slight to AMD, but open standards in and of itself does not make something compelling. Yet, since you brought it up, physx isn't fully "proprietary" either since the physics engine portion works on well....everything.

And on that note, as mentioned earlier - some implementations of this GPU particle physx (which requires CUDA hardware) are excellent while not all are. It isn't meant to be a pre-requisite and it never has. It's merely a value add. Oddly enough, Mantle is also a value add for AMD users. I think both respective technologies in their proprietary aspects (again, i'm only talking of the gpu accelerated particle features of physx, not physx in the general sense) are value adds that are good for AMD or NV users respectively. It's all in the eye of the beholder. They're both good in different ways. If you like one or the other, by all means, buy based on what you like. Now I'm not a stickler for physx, but I buy NV based on other reasons which I won't get into. But the value adds do mean something to a lot of people.

 

[Zanovar]

It doesn't matter if it's supported or not. It is what it is. It's here. It's the way these companies tout value over the other (well, one of the ways).

Let me run a scenario by you all for a moment, then you tell me how you think it's going to turn out.

Nvidia creates a ton of features that will only work on their GPU's.
AMD creates a ton of features that will work on any GPU.(Intel, Nvidia, AMD)

Down the road, which will you choose?

He shouldnt have to,but hey thats just me,Im loving how gsync works so the next great monitor(not that tn crap peeps seem to be wetting themselves over)that has it has me tied to nvidia.not happy about that but it is what it is.

 

[Keysplayr]

He shouldnt have to,but hey thats just me,Im f****** loving how gsync works so the next great monitor(not that tn shit peeps seem to be wetting themselves over)that has it has me tied to nvidia.not happy about that but it is what it is.

Could you possibly re-type this please? Having a rough time reading posts like this is all.

And please explain why "He shouldn't have to". How does that work?

 

[Zanovar]

Could you possibly re-type this in English please?

And please explain why "He shouldn't have to". How does that work?

I think you undertood my post quite perfectly,stop being a daffodil..the gist of it anyways.

 

[Keysplayr]

I think you undertood my post quite perfectly,stop being a daffodil..the gist of it anyways.

Sorry, I just hate when people can't take the time.

 

[NTMBK]

It doesn't matter if it's supported or not. It is what it is. It's here. It's the way these companies tout value over the other (well, one of the ways).

Let me run a scenario by you all for a moment, then you tell me how you think it's going to turn out.

Nvidia creates a ton of features that will only work on their GPU's.
AMD creates a ton of features that will work on any GPU.(Intel, Nvidia, AMD)

Down the road, which will you choose?

I would buy whichever GPU offers the best performance/$, because I'm not really that interested in gimmicks. Especially ones that only work in a handful of games which I'm not interested in.

 

[Keysplayr]

I would buy whichever GPU offers the best performance/$, because I'm not really that interested in gimmicks. Especially ones that only work in a handful of games which I'm not interested in.

Excellent answer. Since you don't care for gimmicks, you probably don't care if a feature is proprietary since all proprietary features are gimmicks. Right?

 

[NTMBK]

Excellent answer. Since you don't care for gimmicks, you probably don't care if a feature is proprietary.

The sad thing is that if these features weren't proprietary, they could be promoted to more than gimmicks. Since GPU PhysX is unsupported on over half the market, it can't be made into an integral part of the game engine- if the actual game simulation depended on an NVidia only feature, the game would be utterly broken on rival GPUs. That dooms GPU PhysX to completely extraneous, irrelevant swirling nonsense. If it was a cross-platform GPU physics solution, supported on all GPUs, then developers could rely on it always being present and hence make it more integral to their engine.

 

[Keysplayr]

The sad thing is that if these features weren't proprietary, they could be promoted to more than gimmicks. Since GPU PhysX is unsupported on over half the market, it can't be made into an integral part of the game engine- if the actual game simulation depended on an NVidia only feature, the game would be utterly broken on rival GPUs. That dooms GPU PhysX to completely extraneous, irrelevant swirling nonsense. If it was a cross-platform GPU physics solution, supported on all GPUs, then developers could rely on it always being present and hence make it more integral to their engine.

So if something is proprietary, it can never be anything more than a gimmick. What logic was used to lead to this conclusion?

PhysX doesn't have to be run on a GPU. There are over 450 titles that utilize PhysX whether on GPU or CPU. So, PhysX isn't just about GPUs. And it's a lot more than swirling. And it absolutely does not have to be a part of a game engine to be effective.

 

[AnandThenMan]

So if something is proprietary, it can never be anything more than a gimmick. What logic was used to lead to this conclusion?

What makes people come to that conclusion in respect to Nvidia, is the high profile GPU accelerated PhysX stuff has always been limited to superficials not something that is integral to gameplay. So it comes across as a gimmick, and in some cases looking very unrealistic, designed to be a "hey look at what PhysX can do" rather than a seamless integration.

 

[ams23]

At the end of the day, PhysX is NVIDIA's intellectual property, and they have paid well north of one hundred million dollars to secure and advance that technology. To suggest that this technology should be given away for free and without license fees to other IHV's is insane.

 

[Zanovar]

What makes people come to that conclusion in respect to Nvidia, is the high profile GPU accelerated PhysX stuff has always been limited to superficials not something that is integral to gameplay. So it comes across as a gimmick, and in some cases looking very unrealistic, designed to be a "hey look at what PhysX can do" rather than a seamless integration.

Hey you dont like the physx,you sir are obviosly bad.rotten to the core

 

[AnandThenMan]

To suggest that this technology should be given away for free and without license fees to other IHV's is insane.

Can you point to people that have suggested this? I've said for years that Nvidia should give away the dev tools for PhysX and make it work with all GPUs. Then charge a fee for games that get released. It's a great business model.

 

[ams23]

It's not that simple. AAA games get churned out at an unpredictable rate each year. Also, NVIDIA can only be responsible for quality, performance, and assurance testing on their own hardware using their own software drivers. And if another IHV refuses to pay for the technology or rejects the licensing terms, then they are back to square one. In other words, monetizing this investment based on support for non-NVIDIA GPU's is easier said than done.

 

[NTMBK]

So if something is proprietary, it can never be anything more than a gimmick. What logic was used to lead to this conclusion?

PhysX doesn't have to be run on a GPU. There are over 450 titles that utilize PhysX whether on GPU or CPU. So, PhysX isn't just about GPUs. And it's a lot more than swirling. And it absolutely does not have to be a part of a game engine to be effective.

And that is why I specified GPU PhysX, not CPU PhysX. The CPU PhysX is a great example of how a library should work- it runs fine on any CPU, and can be relied on working in any situation. But the GPU stuff just tends to be limited to "add some swirls on top". (See Borderlands 2 for example.) And it is such a waste of what GPU physics could be, if developers could actually depend on it being present on all brands of GPU. Can you imagine how awesome it would be if GPU physics was used to actually update the game world dynamically, causing dynamic destruction? If PhysX was running on your Intel integrated graphics, while your NVidia discrete GPU handled the rendering?

 

[stahlhart]

Thread cleaned. A warning to everyone here to stay focused on the topic of discussion when participating in threads here, and not to use forum threads to engage in personal attacks against other members.
-- stahlhart

 

[SirPauly]

How can you think any company that creates something that only half of consumers can use good? If things continue on this path, everybody is going to need to have one PC with an AMD card, and one with an nVidia just so they can play all the games. Because if things continue as they are, games are going to start becoming vendor specific.

I actually read this type view a lot and simply don't understand. The core game-play is the same.

Let's take Just Cause 2 utilizing Cuda; AMD owners could enjoy the core game-play and the title was great with Cuda or without! Where is the harm?

 

[monstercameron]

I actually read this type view a lot and simply don't understand. The core game-play is the same.

Let's take Just Cause 2 utilizing Cuda; AMD owners could enjoy the core game-play and the title was great with Cuda or without! Where is the harm?

there is no harm but if one user pays the same money as another with different hardware but are locked out of features, that is ethically wrong.

 

[SirPauly]

there is no harm but if one user pays the same money as another with different hardware but are locked out of features, that is ethically wrong.

It's a competitive sector in an unideal world, imho! I would like to see ideal harmony but not very realistic!

 

[Keysplayr]

there is no harm but if one user pays the same money as another with different hardware but are locked out of features, that is ethically wrong.

I think you might have it a bit wrong there.

Two GPUs. One game. Both can play the game, but one GPU company adds extra features to the game that work only on their GPUs.

The GPU isn't taking anything away from the other GPU company, but giving more to it's own GPUs and the users of those GPUs.

Nothing unethical about it at all. You can't block what was never there in the first place.

 

[3DVagabond]

I don't know how many times we are going to have this discussion? The nVidia proponents think it's a perfectly fine business practice and the AMD proponents think it's dodgy. That's part of the reasons some people have a brand preference. Nobody is ever going to change the others mind. Have your say and move on.