DX11 = PhysX killer?

[jeremyshaw]

Real world numbers, eh? post them - if you can find 'FUD' that can support your view .

Also. DoF was just lazy. The only things I have seen worse, are the default Softimage particle 'explosion' settings.

Just look at, let's say, Halo 3. That put a lot more work into 'smoke' effects then DoF did. DoF did... what? Overuse the physx smoke effect to block your sight? Because they really made it 'look good'? It doesn't. HINT. Also, why not just add pWind to the pSmoke effects to clear out the smoke? It would of made the pGame seem so much more awesome. After all, pEffects is all this game can tout. Besides, smokeless powder since the late 1800s made this moot.

Just so you know, I used to be a Physx nvidiot, until crap games like this just flooded out the gates.

Also... Physx is better on the GPU then the CPU.

Everything else doesn't care, as we all knw nVidia just wants GPU sales, not CPU sales .

FUD is what you love to spell. After all, you are afraid of switching out after Physx is proven to be nothing more then a marketing point for nVidia.


Real world numbers... hmm...

Halo 2 = 6.3 million copies, according to NPD. [EDIT: whoops, that was an old number, it''s now 8.46 million]

Batman AA + DoF + Cryostasis = not even close.


Yeah... Physx has a 'lot' of games under it's belt... I mean, a lot. Very much lot. No kidding.

FUD... what a joke. You love spilling that at every turn when people present supurior arguements.

You're gonna ignore this, because you know you cannot defeat it. Even with your pZealot mode enabled. [EDIT:]

 

[Silus]

PhysX is a proproetary standard that is not open to the Industry.

OPENCL is a standard open to the industry.

If you honestly think for a single second a middleware wont be built upon that foundation your smoking something heavy.

When we say "OpenCL will kill PhysX" we more than understand a middleware needs to exist. We also fully understand it will.

So many clueless people, it hurts...

Do you understand that NVIDIA supports ALL standards and will eventually port PhysX to use them ? Right now, it's just CUDA; because OpenCL is still in its infancy. But when it is good to go, PhysX will be ported to OpenCL, just like it was ported to CUDA.

 

[Silus]

If PhysX is built on top of OpenCL how will it kill PhysX?

Precisely.

 

[Silus]

I am amazed at the sheer number of people not getting it...it's like saying OSX will kill EAX...most people really havn't got a clue about this and they mix (and confuse) DX, Physx, OpenCL, Havok ect in one big pot...and not getting any of them

But funny to see how this "unimportant useless PhysX" really needs to be beaten...by anything people can dig up...related or not

I can't agree more with you. I'm in awe with the amount of cluelessness in a tech-forum. It's one thing to be a supporter of one company. It's quite another to be completely ignorant about a technology, but still comment on it.

 

[Silus]

I think it's interesting how AMD appears to be moving away from Havok with their recent support for Bullet. Apparently Intel wasn't too charmed of OpenCL Havok after all, so AMD went with the next best thing: Bullet on OpenCL

Methinks Havok may very well join PhysX in that future footnote

But that was known since the very start of AMD's partnership with Intel and Havok. Only the ferverous AMD fanboys that constantly hammered ALL the PhysX threads in the Physics sub-forum, thought that AMD was in a "better position" to succeed in GPU physics, which was hilarious since they have absolutely NOTHING to show right now.

Until Larrabee is out, Intel has no interest in "helping" AMD. Actually, Intel has no interest in helping AMD, period.

 

[Cali3350]

So many clueless people, it hurts...

Do you understand that NVIDIA supports ALL standards and will eventually port PhysX to use them ? Right now, it's just CUDA; because OpenCL is still in its infancy. But when it is good to go, PhysX will be ported to OpenCL, just like it was ported to CUDA.

It will eventually port PhysX to work on all OpenCL supported cards and prove a industry wide standard? Fine.

Nvidia has just recently decided running a Nvidia PhysX card with a AMD graphics chip is not allowed in their drivers? Ok.

I totally see the convergence. They obviously care about PhysX being a open standard.

 

[BlueStorm]

^This.


Because PhysX is owned by Nvidia that means you have to pay to use it. Why would anyone pay to use something when they could use something similar for free?

Basically you're wrong all around. By your logic, everyone would be running linux today. But we're not, because people are willing to pay for a better experience.

The money goes to the company who makes it, who in turn invests it into more of that technology to make it better and better. By the way, Havok charges outrageous license fees on a per SKU basis (1 fee for PS3, 1 fee for XBOX, 1 fee for PC, all for the same game). Works out pretty well for them eh?

You guys are confusing open standards with open source.. not the same thing. Just because something is an "open standard" (i.e. runs on all hardware) doesn't mean a company will give it away. People need this funny thing called food, and a place to sleep, so they tend to like to be compensated for their hard work. And just because something is an "open standard" doesn't mean it will run well on all hardware. So the experience you end up with could be a big cluster, despite the fact that it "works".

Basically at the end of the day, you'll never get something with the breadth of tools and quality unless it's funded by something. In NVIDIAs case, that's GPU sales.

NVIDIA is also free to improve PhysX to take advantage of new CUDA features.. hand in hand development with no latency in the process, with no issues of IP or hardware details when talking to third party physics providers. Meanwhile, getting 1 feature to work and be QA'd across multiple platforms/hardware configs with an "open" standard will be a nightmare for a developer.

Developers are willing to pay money for something that will work the way they want, be well supported by the creators, and save them time in the long run. Upfront cost is nothing if it save you hundreds or thousands of man-hours of work. That's called an "investment".

People who get paid tend to produce better quality work than those who don't. Basically all of these are distinct advantages that people have a tendency of overlooking, and instead just spout the "OPEN STANDZRASD!!!1!" line without understanding it's implications, or what it means.

NVIDIAs advantages with their "closed standard" :
-Faster physics feature deployment
-No need to dumb-down physics features to the greatest common denominator
-Very good understanding of and tight optimization for 1 type of hardware (NVIDIA GPUs)
-Well funded and supported physics solution
-Potentially better tools than some random smaller company could produce
-Better control over the end-user experience

Lastly, people are overreacting to PhysX, saying it will die because of lack of adoption. The funny thing is it will piggyback on GeForce's success as well as it's own, and for those keeping track, the Steam survey says 65% or so of PC gamers (of course we use this rather than some market perspective because most of the large number of intel integrated users dont game at all) are on NVIDIA hardware. So for a developer, serving over half of your users with enhanced features is pretty well worth it.

 

[Silus]

Yeah except if you look at the PhysX title list and compare it to upcoming games using Havok... that link doesn't make a convincing argument at all. PhysX still needs that AAA title that will impress gamers and make them run out to get an nvidia card, because at the moment it has none. Red Faction Guerilla could have been that game.

And with upcoming heavyweights such as Bioshock 2, Diablo 3/Starcraft 2, Crysis 2, Witcher 2 all using non-PhysX engines, you have to at least understand why people are skeptical about the whole issue.

Red Faction ? Are you joking ? Red Faction was never a AAA title. Why would it constitute a good thing for PhysX ?

Also, Mafia 2 and Mass Effect 2, two AAA titles are coming with PhysX support. And there's no issue here. Just some people that don't "like" PhysX, because it is owned by NVIDIA, going as far as say "as long as it's not PhysX". No game is going to be good, because of the physics API used. But PhysX does provide the ability to create more realistic and complex physics simulations, because it can offload the physics processing from the CPU, to a GPU or PPU, unlike other physics APIs in the market right now.

The other aspect of how ridiculous these threads are is the fact that the naysayers expect PhysX to either be awesome, with no drawbacks or fail completely. Do you have the slighest idea what it was to run Havok's first uses, in games like HL2, in a medium range PC ? The framerate would tank with just a few objects that got physics calculations, on the screen. But every sane person knew that it was normal. New tech always kills current hardware and will take some time to improve. PhysX seems to be in a realm of its own, in the head of these people. It's still new, but it must do everything without penalty. It must appear in AAA titles or it sucks...See how ridiculous that is ?

 

[Silus]

It will eventually port PhysX to work on all OpenCL supported cards and prove a industry wide standard? Fine.

So at last, do you understand what OpenCL, CUDA and PhysX is ?

Will it confuse you, if I say OpenCL == CUDA ?

Nvidia has just recently decided running a Nvidia PhysX card with a AMD graphics chip is not allowed in their drivers? Ok.

They do it due to the fact that they cannot provide support to correct any fault, when a non CUDA ready GPU is being used for graphics rendering. And obviously, also some business reasons..
But I would actually love to see them remove that limitation and then see everyone trying to get support from them and they would deny it, because they cannot support someone else's hardware. I would bet that when that happens, we would get another couple hundred thousand threads about how "NVIDIA sucks" because they provide no support, just like we already have a couple hundred thousand threads about how "NVIDIA sucks", because they imposed that limitation.

I totally see the convergence. They obviously care about PhysX being a open standard.

They do. It's free to license, but it's not free to use as a whole. You still need to pay for the source code, if you really need it. And developers usually do. Havok on the other hand, has huge licensing fees, which developers will need to keep paying, if they use the same engine in their new games, as they used in their past games. Changing the physics engine is a major undertaking and in a time of cross-platform development, many developers simply won't do it.

 

[Deleted member 174368]

Red Faction ? Are you joking ? Red Faction was never a AAA title. Why would it constitute a good thing for PhysX ?

I'm not going to discuss the quality of the title even though it's a much better than the objectively terrible Darkest of Days (which Atech keeps mentioning for some reason), but Red Faction is a great example of a game that has been highly praised for its physics since it forms a major, non-gimmicky part of the gameplay. That's what PhysX needs.

Also, Mafia 2 and Mass Effect 2, two AAA titles are coming with PhysX support. And there's no issue here. Just some people that don't "like" PhysX, because it is owned by NVIDIA, going as far as say "as long as it's not PhysX". No game is going to be good, because of the physics API used. But PhysX does provide the ability to create more realistic and complex physics simulations, because it can offload the physics processing from the CPU, to a GPU or PPU, unlike other physics APIs in the market right now.

We'll see to what extent it actually improves the Mafia 2 experience. I'm honestly just hoping for more than a few extra particle effects.

 

[Silus]

I'm not going to discuss the quality of the title even though it's a much better than the objectively terrible Darkest of Days (which Atech keeps mentioning for some reason), but Red Faction is a great example of a game that has been highly praised for its physics since it forms a major, non-gimmicky part of the gameplay. That's what PhysX needs.

Why must it be immediately part of the gameplay ?

You should read my post above, about how some people seem to want PhysX to do everything from the start, or fail completely. PhysX carried by NVIDIA is relatively new and it has established itself as a good alternative to what was the biggest physics engine: Havok.

Havok's SDK was released in 2000 and only in 2004, with the release of HL2, did we actually see any influence of physics in gameplay and after that, nothing really impressive showed up...

Why must PhysX accomplish this right away ?

We'll see to what extent it actually improves the Mafia 2 experience. I'm honestly just hoping for more than a few extra particle effects.

There you go again. Assuming something about a game you haven't seen yet. That's precisely the point I made above and in a previous post.

 

[Atech]

Real world numbers, eh? post them - if you can find 'FUD' that can support your view .

Also. DoF was just lazy. The only things I have seen worse, are the default Softimage particle 'explosion' settings.

Just look at, let's say, Halo 3. That put a lot more work into 'smoke' effects then DoF did. DoF did... what? Overuse the physx smoke effect to block your sight? Because they really made it 'look good'? It doesn't. HINT. Also, why not just add pWind to the pSmoke effects to clear out the smoke? It would of made the pGame seem so much more awesome. After all, pEffects is all this game can tout. Besides, smokeless powder since the late 1800s made this moot.

Just so you know, I used to be a Physx nvidiot, until crap games like this just flooded out the gates.

Also... Physx is better on the GPU then the CPU.

Everything else doesn't care, as we all knw nVidia just wants GPU sales, not CPU sales .

FUD is what you love to spell. After all, you are afraid of switching out after Physx is proven to be nothing more then a marketing point for nVidia.


Real world numbers... hmm...

Halo 2 = 6.3 million copies, according to NPD. [EDIT: whoops, that was an old number, it''s now 8.46 million]

Batman AA + DoF + Cryostasis = not even close.


Yeah... Physx has a 'lot' of games under it's belt... I mean, a lot. Very much lot. No kidding.

FUD... what a joke. You love spilling that at every turn when people present supurior arguements.

You're gonna ignore this, because you know you cannot defeat it. Even with your pZealot mode enabled. [EDIT:]

PhysX is the most used physcis API by developers now, no matter the FUD you try and spin

 

[kadozer]

So much marketing fluff here and physx is so great it speaks for itself..oh wait.
Cliff notes: If you like overly done flying papers, broken glass, and other ancillary effects go with Physx.

 

[kumquatsrus]

lol, trying to sell us access to the 'top 10' list?

What's there to use it: a whole bunch of nVidia puppets?
At least Havok has got some games under it's belt.
Physx just came out with one.

Dunno. If you want to just count the 2 Halo games that use it: 200% lead.

But if you want to count: "cloth tearing for points" then... (I guess the Warmonger demo counts, too...)

besides... what does this even have to do with the sheer number of games sold? Compare that!

since when does the amount of games sold correlate with the particular physics engine being used under the hood?

yes, physx has one game only...indeed. comparing a niche market like games taking advantage of gpu physx with a mainstream one like cpu havok seems rather disingenuous. at least compare software physx based games with software havok based games for a fair comparison. and truly the havok list (regardless of "quality" of games) should be longer seeing as how it has been around 3 times as long as physx has, so no qualms there.

Yeah except if you look at the PhysX title list and compare it to upcoming games using Havok... that link doesn't make a convincing argument at all. PhysX still needs that AAA title that will impress gamers and make them run out to get an nvidia card, because at the moment it has none. Red Faction Guerilla could have been that game.

And with upcoming heavyweights such as Bioshock 2, Diablo 3/Starcraft 2, Crysis 2, Witcher 2 all using non-PhysX engines, you have to at least understand why people are skeptical about the whole issue.

why does physx need a "AAA" title to impress gamers to buy a nvidia card when it is merely physics middleware that runs capably on the cpu. if you are only talking about the very small niche of gpu physx, this is not a plausible demand because if it was and there was such a game, i don't think you would want people only running out to buy only nvidia cards for a game. maybe if nvidia buys a developer to make quality exclusive games with gpu physx could it actually work.

hypothetically, if red faction guerilla was such a game, and millions went out to buy a nvidia card in order play this killer app, that wouldn't leave ati in a good position now would it. so even in spite of the sarcasm, you should be leery of this happening in any possible scenario. physx just needs to be accepted by developers/ publishers to be successful. that's it. the more it is used, the more successful it is. just like the more havok is used, the more successful it is. the more options, the better. just like having more than one cpu maker or one gpu maker is better due to more options.

Ach... I can count Batman AA in, but nothing else as a 'AAA' title. It's not gonna be selling on the same level as Halo 3 or CoD4, but it will be the best Physx has to offer, for now.

But you are right. Havok and others (even Battlefront 2!) have already established Physx is not needed.
Any simple view of the before and after shots of 'Darkest of Days' makes this pretty obvious.

labeling a game as a "AAA" title is purely subjective to every gamer, and therefore lies in the realm of opinion. but just to play along and humor you, there is a little old game called gears of war. dunno, you might have heard of it. pretty sure gears of war 1+2 which use physx could be in the same category as halo3 or cod4 since it sold millions and was a system seller for the xbox360.

mafia 2, dark void, mass effect 2, borderlands, aliens colonial marines, dragon age origins, etc. are all games coming out in the next couple of months using physx. and yes, i'm sure you expect all of them to be either mediocre or suck completely. not sure what quality games coming out in the next six months will use havok, though i'm sure there are some. perhaps you can enlighten us.

moreover, physx being used in two of the biggest and most heavily licensed game engines in unreal engine 3 and gamebryo, and being backed by some of the biggest publishers such as ea, sega, capcom, eidos, thq, etc. might lend credence to the notion that there will actually continue to be a surge, and not a waning in physx based games for the next couple of years. being available to use across the widest array of platforms (pc, mac, linux, ps3, xbox360, wii, iphone) certainly helps as well, i'm sure.

crysis proves havok isn't needed either if that is the argument. which just goes to show that a quality game isn't defined solely by it's physics engine. it is the complete package that matters. havok and physx merely make it much easier for devs to integrate more realistic physics into their games. so again, what is with the physx hate? it seems unwarranted. hating on nvidia is a different story. just because you hate the "parent", doesn't mean the "child" will be the exact same way (poor metaphor, i know, lol).

Real world numbers, eh? post them - if you can find 'FUD' that can support your view .

Also. DoF was just lazy. The only things I have seen worse, are the default Softimage particle 'explosion' settings.

Just look at, let's say, Halo 3. That put a lot more work into 'smoke' effects then DoF did. DoF did... what? Overuse the physx smoke effect to block your sight? Because they really made it 'look good'? It doesn't. HINT. Also, why not just add pWind to the pSmoke effects to clear out the smoke? It would of made the pGame seem so much more awesome. After all, pEffects is all this game can tout. Besides, smokeless powder since the late 1800s made this moot.

Just so you know, I used to be a Physx nvidiot, until crap games like this just flooded out the gates.

Also... Physx is better on the GPU then the CPU.

Everything else doesn't care, as we all knw nVidia just wants GPU sales, not CPU sales .

FUD is what you love to spell. After all, you are afraid of switching out after Physx is proven to be nothing more then a marketing point for nVidia.


Real world numbers... hmm...

Halo 2 = 6.3 million copies, according to NPD. [EDIT: whoops, that was an old number, it''s now 8.46 million]

Batman AA + DoF + Cryostasis = not even close.


Yeah... Physx has a 'lot' of games under it's belt... I mean, a lot. Very much lot. No kidding.

FUD... what a joke. You love spilling that at every turn when people present supurior arguements.

You're gonna ignore this, because you know you cannot defeat it. Even with your pZealot mode enabled. [EDIT:]

i can't really comment on the quality of darkest of days since i only played the demo, but apparently that's how smoke and fog work in real life - they can obscure your view of anything that they cover (and the view of others that are trying to see you). just like when dense weather can make visibility difficult in a car. on the other hand, a game that uses havok that could of made use of gpu physx is wolfenstein. just played the demo and the nazi banners look drab without cloth physics. is that important to you? probably not. but after seeing it in other games, that little touch makes it just a smidge more immersive for me. i'm sure there are other things that could have been improved upon as well.

really? comparing sales of a massively popular 5 year old console game with a month old critically acclaimed game and two small indie games is really stretching it, especially considering gpu physx support has only been available for less than a year and a half at this point.

and the better statement would be physics is handled better on the gpu than the cpu based on extant hardware (just like 3d acceleration). hence, amd's interest in getting physics support working on its gpus vs. just sticking with physics that would work fine on their cpus. and in the meantime, nvidia will continue to heavily market physx.

i wouldn't call a couple hundred games something to sneeze at. and if you want to discuss quality over quantity instead. there are tons of crap games that make use of havok (stormrise and hellgate london, for instance) as much as there are tons of crap games that make use of physx. (just look at the havok site or the nzone site and the evidence speaks for itself). the more alarming assertion would be that there are a ton of crap games in general, irregardless of the physics engine being used. i could say the same for any form of media. just the internet alone would be enough of an argument.

so if you want to hate physx, hate it because it offers a small niche implementation that currently accelerates it on only one vendor's gpus that enables completely optional effects, while providing the framework for a better solution to the physics "problem" in gaming, thereby causing a shift in game development so that games can eventually become even better with this technology, and also bringing the impetus necessary for competition to exist in this arena in the first place.

It will eventually port PhysX to work on all OpenCL supported cards and prove a industry wide standard? Fine.

Nvidia has just recently decided running a Nvidia PhysX card with a AMD graphics chip is not allowed in their drivers? Ok.

I totally see the convergence. They obviously care about PhysX being a open standard.

i see your point, but the only reason hardware-accelerated physx isn't an open standard is because there is no competition in the gpu accelerated physics arena. as long as there is none, nvidia will continue to see this as a competitive advantage. only when something arrives will the catalyst (pun intended) be provided for change to take place (possibly opencl bullet being pushed by amd; apparently moreso than havok now). either nvidia will make gpu physx open and continue to make a return on their investment, or they will let it die off to where a competitor might take its place. we will see how serious amd is about this iniatitive and if they can truly leverage some momentum in the same way nvidia has done for ages with the "twimtbp" program.

 

[Elledan]

So much marketing fluff here and Havok is so great it speaks for itself..oh wait.
Cliff notes: If you like overly done flying papers, broken glass, and other ancillary effects go with Havok.

It's all the same

*cough*Half-Life 2*cough*

 

[Silus]

moreover, physx being used in two of the biggest and most heavily licensed game engines in unreal engine 3 and gamebryo, and being backed by some of the biggest publishers such as ea, sega, capcom, eidos, thq, etc. might lend credence to the notion that there will actually continue to be a surge, and not a waning in physx based games for the next couple of years.

Precisely, It clearly shows how developers are seeing PhysX as a perfectly viable alternative to Havok, And given that PhysX is still quite new, it's definitely a good sign for them, to already have such a big list in supporters. Obviously, everything starts with small studios, but we are already seeing very big (the biggest actually: EA), using PhysX in some of their upcoming titles.

crysis proves havok isn't needed either if that is the argument. which just goes to show that a quality game isn't defined solely by it's physics engine. it is the complete package. havok and physx merely make it much easier for devs to integrate more realistic physics into their games. so again, what is with the physx hate? it seems unwarranted. hating on nvidia is a different story. just because you hate the "parent", doesn't mean the "child" will be the exact same way (poor metaphor, i know, lol).

Yes. No game is a good game, because of the use of a specific physics API. But the easy of use and development tools, make a specific physics API, the preferred one over another. And that's why PhysX is not a "gamer" thing, but actually a "developer" thing, that will then provide the effects for the "gamer" in their games. First and foremost PhysX appeals to developers and then to gamers.

so if you want to hate physx, hate it because it offers a small niche implementation that currently accelerates it on only one vendor's gpus that enables completely optional effects, while providing the framework for a better solution to the physics "problem" in gaming, thereby causing a shift in game development so that games can eventually become even better with this technology and bringing the impetus necessary for competition to exist in this arena in the first place.

I can't see that as a reason to hate it, but it is indeed why some people hate it. Yes, GPU Physics is a GeForce only thing for now, but PhysX itself defaults to the CPU anyway, so anyone can enjoy the "normal" effects that we've seen for years. What happens is that some people think that most of the new effects we are seeing today, accelerated by GPU Physics, is doable on a CPU, when in reality it isn't. People go as far as say that developers don't know what they're doing, when that's the most ludicrous argument I have ever heard. The people that create games, don't know what they're doing ?!? Really ?

Still, even though these people criticize PhysX on a daily basis, they do see the need for the effects it provides and thus are quite "mad" to not being able to see them done on a non-GeForce graphics card. If they didn't see the need for them or how good some of those effects are, they wouldn't "hate" PhysX and would just ignore it...

 

[kumquatsrus]

Basically you're wrong all around. By your logic, everyone would be running linux today. But we're not, because people are willing to pay for a better experience.

The money goes to the company who makes it, who in turn invests it into more of that technology to make it better and better. By the way, Havok charges outrageous license fees on a per SKU basis (1 fee for PS3, 1 fee for XBOX, 1 fee for PC, all for the same game). Works out pretty well for them eh?

You guys are confusing open standards with open source.. not the same thing. Just because something is an "open standard" (i.e. runs on all hardware) doesn't mean a company will give it away. People need this funny thing called food, and a place to sleep, so they tend to like to be compensated for their hard work. And just because something is an "open standard" doesn't mean it will run well on all hardware. So the experience you end up with could be a big cluster, despite the fact that it "works".

Basically at the end of the day, you'll never get something with the breadth of tools and quality unless it's funded by something. In NVIDIAs case, that's GPU sales.

NVIDIA is also free to improve PhysX to take advantage of new CUDA features.. hand in hand development with no latency in the process, with no issues of IP or hardware details when talking to third party physics providers. Meanwhile, getting 1 feature to work and be QA'd across multiple platforms/hardware configs with an "open" standard will be a nightmare for a developer.

Developers are willing to pay money for something that will work the way they want, be well supported by the creators, and save them time in the long run. Upfront cost is nothing if it save you hundreds or thousands of man-hours of work. That's called an "investment".

People who get paid tend to produce better quality work than those who don't. Basically all of these are distinct advantages that people have a tendency of overlooking, and instead just spout the "OPEN STANDZRASD!!!1!" line without understanding it's implications, or what it means.

NVIDIAs advantages with their "closed standard" :
-Faster physics feature deployment
-No need to dumb-down physics features to the greatest common denominator
-Very good understanding of and tight optimization for 1 type of hardware (NVIDIA GPUs)
-Well funded and supported physics solution
-Potentially better tools than some random smaller company could produce
-Better control over the end-user experience

Lastly, people are overreacting to PhysX, saying it will die because of lack of adoption. The funny thing is it will piggyback on GeForce's success as well as it's own, and for those keeping track, the Steam survey says 65% or so of PC gamers (of course we use this rather than some market perspective because most of the large number of intel integrated users dont game at all) are on NVIDIA hardware. So for a developer, serving over half of your users with enhanced features is pretty well worth it.

+1

Red Faction ? Are you joking ? Red Faction was never a AAA title. Why would it constitute a good thing for PhysX ?

Also, Mafia 2 and Mass Effect 2, two AAA titles are coming with PhysX support. And there's no issue here. Just some people that don't "like" PhysX, because it is owned by NVIDIA, going as far as say "as long as it's not PhysX". No game is going to be good, because of the physics API used. But PhysX does provide the ability to create more realistic and complex physics simulations, because it can offload the physics processing from the CPU, to a GPU or PPU, unlike other physics APIs in the market right now.

The other aspect of how ridiculous these threads are is the fact that the naysayers expect PhysX to either be awesome, with no drawbacks or fail completely. Do you have the slighest idea what it was to run Havok's first uses, in games like HL2, in a medium range PC ? The framerate would tank with just a few objects that got physics calculations, on the screen. But every sane person knew that it was normal. New tech always kills current hardware and will take some time to improve. PhysX seems to be in a realm of its own, in the head of these people. It's still new, but it must do everything without penalty. It must appear in AAA titles or it sucks...See how ridiculous that is ?

+1 also

So much marketing fluff here and physx is so great it speaks for itself..oh wait.
Cliff notes: If you like overly done flying papers, broken glass, and other ancillary effects go with Physx.

don't know what you are expecting here. you want something more than "fluff" looking physics effects in games, then prepare to face the consequences. let's say in an alternate reality, nvidia buys out and owns a game developer like crytek and causes them to make crysis 2 an nvidia exclusive that requires gpu physx to run the game and that will actually make it more than just "fluff" in order to turn the gameplay into something more meaningful and dynamic. is this something that would be acceptable to you? if not, i suggest you either enjoy the "fluff", or just ignore it. i'm sure ati only users would be very unhappy with this hypothetical scenario. but this has been par for course in the console world when it comes to exclusive games. be content that this isn't an issue when it comes to the gpu vendor being used. otherwise, wait (im)patiently like everyone else for accelerated physics with "widespread" support (for all capable gpus regardless of color - no more racism, lol) that will allow for meaningful gameplay improvements.

btw, i don't see the "fluff" of dx10/11 enabled graphic enhanced options being derided here. boggles the mind methinks.

 

[Atech]

Still, even though these people criticize PhysX on a daily basis, they do see the need for the effects it provides and thus are quite "mad" to not being able to see them done on a non-GeForce graphics card. If they didn't see the need for them or how good some of those effects are, they wouldn't "hate" PhysX and would just ignore it...

Like I said before, it's just a matter of time before the anti-GPU-PhysX (physics) crowd start hailing GPU-physics.
That is when AMD pull's it shite together.
The all of a sudden the red camp will hail GPU-physics as the second comming...even bigger than "Eyefinity" (that you alreday have had via nHancer on Geforce cards for years)...

The tune used to be:
"Havok will kill PhysX"...but they forgot to factor in that Havok is owned by Intel.
Now the tune is:
"Bullet will kill PhysX"...just a shame that currently only ~10% of the developers use Bullet..and to use the insanly badly matra of the clueless people:
"But bullet dosn't have any AAA titles"

Dosn't ring well now does it?

And I am ignoring the clueless minons that can't seperate a physics middleware from an API...(the "OpenCL will kill PhysX" minions"...

The grapes are indeed sour

It was the same thing before NVIDIA bought AGEIA.
There both NVIDIA fans and AMD fans really dislike "PhysX".
The NVIDIA crowd has learned now...only the red team is still stuck in the mud

 

[kadozer]

What secondary card should I get so I can run Havok? I gots mad cash yo

 

[Atech]

What secondary card should I get so I can run Havok? I gots mad cash yo

Read the fucking thread

Havok = Intel...you do the math.

 

[p3sty]

halo 2 sucked Die Nvidia DIE!!!!!

 

[kumquatsrus]

I can't see that as a reason to hate it, but it is indeed why some people hate it. Yes, GPU Physics is a GeForce only thing for now, but PhysX itself defaults to the CPU anyway, so anyone can enjoy the "normal" effects that we've seen for years. What happens is that some people think that most of the new effects we are seeing today, accelerated by GPU Physics, is doable on a CPU, when in reality it isn't. People go as far as say that developers don't know what they're doing, when that's the most ludicrous argument I have ever heard. The people that create games, don't know what they're doing ?!? Really ?

Still, even though these people criticize PhysX on a daily basis, they do see the need for the effects it provides and thus are quite "mad" to not being able to see them done on a non-GeForce graphics card. If they didn't see the need for them or how good some of those effects are, they wouldn't "hate" PhysX and would just ignore it...

lol, sorry. subtle sarcasm ftl. but yeah, the real complaint is ati needs to have the same thing or better so no one gets left out. well, the only thing standing in the way now is time, assuming amd is really interested in participating in propelling gpu accelerated physics forward.

 

[kumquatsrus]

What secondary card should I get so I can run Havok? I gots mad cash yo

this:

http://www.newegg.com/Product/Product.aspx?Item=N82E16814195055

Read the fucking thread

Havok = Intel...you do the math.

whoa down boy! i'm pretty sure he's just being sarcastic lol.

halo 2 sucked Die Nvidia DIE!!!!!

yeah suck it!!! oh wait, halo2 was made by bungie using havok. uhhh, yeah nvidia die!!!

 

[Silus]

lol, sorry. subtle sarcasm ftl. but yeah, the real complaint is ati needs to have the same thing or better so no one gets left out. well, the only thing standing in the way now is time, assuming amd is really interested in participating in propelling gpu accelerated physics forward.

And they had plenty of time to do so, but they chose to partner with Intel and its Havok and we can all see where that led them: Nowhere. That's why we are now seeing them doing some PR about Bullet physics and how that's the way to go. A while back it was Havok...

As I've said many times in the physics sub-forum, AMD chose a very bad partner to slingshot their own GPU physics efforts and they are paying for it. That's their own fault. No one else's.

And as been established already, the OP's thread title is ridiculous, because DX11 as a graphics API won't change anything. If OpenCL becomes a standard, PhysX will run on it. If DirectCompute becomes the standard, PhysX will run on it. If nothing happens, it will run on a CUDA ready GPU. NVIDIA has all these "hypothetical standards" covered, to keep PhysX being supported without issues.

 

[ravaii]

If Physx gets ported to OpenCL or DirectCompute then won't ATI gpus be able to run Physx too?

I am trying to educate myself on the subject. Physx is an application that runs on CUDA, which is an API that Nvidia GPUs support but not ATI. If Physx is ported to OpenCL or DirectCompute, then won't Physx be able to run on ATI GPUs too? From what I understand, software developers currently have free access to the Physx SDK, so all ATI has to do is make sure its GPUs can run the OpenCL or DirectCompute port of Physx. I don't see where the additional licensing fees for ATI supporting Physx come in.

Is it possible for Physx to have an OpenCL or DirectCompute port and still not be able to run on an ATI gpu? Can Nvidia do anything to prevent ATI from using physx if it does get ported?

 

[Cali3350]

If Physx gets ported to OpenCL or DirectCompute then won't ATI gpus be able to run Physx too?

I am trying to educate myself on the subject. Physx is an application that runs on CUDA, which is an API that Nvidia GPUs support but not ATI. If Physx is ported to OpenCL or DirectCompute, then won't Physx be able to run on ATI GPUs too? From what I understand, software developers currently have free access to the Physx SDK, so all ATI has to do is make sure its GPUs can run the OpenCL or DirectCompute port of Physx. I don't see where the additional licensing fees for ATI supporting Physx come in.

Is it possible for Physx to have an OpenCL or DirectCompute port and still not be able to run on an ATI gpu? Can Nvidia do anything to prevent ATI from using physx if it does get ported?

Nvidia could port PhysX over to OpenCL and the world would be a wonderful place. I dont see that happening though.

If Nvidia did port it over to OpenCL then ATI would be able to run it, as its a open language for all graphics hardware.

 

[Silus]

Nvidia could port PhysX over to OpenCL and the world would be a wonderful place. I dont see that happening though.

If Nvidia did port it over to OpenCL then ATI would be able to run it, as its a open language for all graphics hardware.

Incorrect, since ATI would still have to port PhysX instructions translated from OpenCL, to Brook+, in their drivers.

Translating PhysX to whatever a GPU understands, requires that the drivers are ready for it. If ATI doesn't form a deal with NVIDIA in order to get support for PhysX, then PhysX will never run on any ATI card.

 

[Silus]

If Physx gets ported to OpenCL or DirectCompute then won't ATI gpus be able to run Physx too?

Yes, but only if ATI translates PhysX instructions from OpenCL to what their GPUs understand: Brook+. And this requires that ATI licenses the tech from NVIDIA.

I am trying to educate myself on the subject. Physx is an application that runs on CUDA, which is an API that Nvidia GPUs support but not ATI. If Physx is ported to OpenCL or DirectCompute, then won't Physx be able to run on ATI GPUs too? From what I understand, software developers currently have free access to the Physx SDK, so all ATI has to do is make sure its GPUs can run the OpenCL or DirectCompute port of Physx. I don't see where the additional licensing fees for ATI supporting Physx come in.

And that's great. It's better to try and educate yourself, instead of just making some wild claims as some do around here...And don't just take my word for it. Investigate by yourself at the source. Don't just stick with info found in forums...

Now going to your questions. PhysX is not an application per se. PhysX is a physics API, that can be incorporated in a game engine (for example), to apply physics calculations to the objects in the game's "world".
For ATI to be able to run PhysX on their cards, they still need that OpenCL or DirectCompute "code", which in turn will be translated to Brook+ in ATI's drivers. But to do that translation, ATI needs access to PhysX instruction set. And to get that, they need to license the tech from NVIDIA. The license is free, but there are costs in terms of source code for example.

Is it possible for Physx to have an OpenCL or DirectCompute port and still not be able to run on an ATI gpu? Can Nvidia do anything to prevent ATI from using physx if it does get ported?

Yes, it is, as I explained above.
As for the second question, only AMD/ATI can do anything about that. It's not a question of NVIDIA preventing anything, since translation is done in ATI's own drivers and NVIDIA has no control over those. But to have that translation, they need to license the tech first.

 

[b00gjuice]

Hey, PhysX, what's up?

Remember Glide?

PhysX? PhysX?! What's wrong, PhysX!!!! You don't look so good!

I remember Glide, it was propietary to 3DFX but still superior and for a few years several major titles were glide exclusive (for 3D), or used CPU. Everquest shipped with glide support and looked/performed better with glide than D3D. Glide was a big deal at the time. GLQuake basically killed the "Verite Cards" and made 3DFX dominant. Ultima IX was Glide, and it's Direct3D implementation sucked. Also, 3DFX wrapped openGL titles with glide, so they also dominated on OpenGL titles in the late 90's.

What killed 3DFX was 3 things, none of which was using propietary glide, which devolopers loved to use and hated D3D. 1) 3DFX buying STB going self-branding and locking out the oem's, pissing them off royally. 2) Nvidia's six-month new video card rotation, eventually catching up in processing power using D3D in just two years. 3) 3DFX bad business decisions (which I won't list, but they are many). It wasn't until GeForce hit the scene that a card running D3D could match what the Voodoo 2/3 could do with glide over the previous two years.

My point is, Glide was the only thing that kept 3DFX going for as long as it did because
3DFX was poorly run. Of course, hindsight is 20/20, the 3D industry was all new back then.

Nvidia needs PhysX to stay propietary and so what if they piss off the core ATI users that don't buy their cards anyway.

 

[shansoft]

PhysX is the most used physcis API by developers now, no matter the FUD you try and spin

prove it..

don't show that link again, because its nothing but a link with words..

I want to see how many games that have been release with physX implement compare to Havok....

and prove something that GPU physics is superior to CPU physics...

btw, the game you show is nothing ... I don't see anything that CPU/Havok can't do....and something is just too fake in there, you are creating infinite particle while hitting a object and the object never gets destroy LOL....
and please, do not show some weird rolling potato style fluid, it just not working or bring anything impressive but ugly and unrealistic.

 

[Parmenides]

Back on topic:
DX11 is not designed to be a Physx killer, just as Physx is not designed to be a DirectX competitor. While Physx is basically a giant CUDA app, and DirectCompute seeks to displace that, it's just not the same 'apples to apples' comparison, yet.

True, but the question is: with the advent of DX11, will proprietary Physx have much draw? Will it make much of a difference? Will it attract developers? Will Physx be necessary for cutting edge physics? If it won't then it will die.

 

[Atech]

prove it..

don't show that link again, because its nothing but a link with words..

So is your post...just empty words with no meaning.

I want to see how many games that have been release with physX implement compare to Havok....

Why?
It would just be "nothing but a link with words"

and prove something that GPU physics is superior to CPU physics...

When is too stupid too stupid?
Right now, you are more stupid than stupid.
Back in 2006 it was clear to everyone that the PPU was WAY faster than a CPU for physics:
http://www.youtube.com/watch?v=a7hIuytdRKw


Guess you ar more behind than I thought.

btw, the game you show is nothing ... I don't see anything that CPU/Havok can't do....and something is just too fake in there, you are creating infinite particle while hitting a object and the object never gets destroy LOL....

Show me...you have a nag of dismming links and posting nothing but FUD...surprise me.

and please, do not show some weird rolling potato style fluid, it just not working or bring anything impressive but ugly and unrealistic.

I can show you Havok CPU water:
www.havokCPUwater.com

Next time, bring something to table...besides your ignorance

 

[shansoft]

So is your post...just empty words with no meaning.

Why?
It would just be "nothing but a link with words"

When is too stupid too stupid?
Right now, you are more stupid than stupid.
Back in 2006 it was clear to everyone that the PPU was WAY faster than a CPU for physics:
http://www.youtube.com/watch?v=a7hIuytdRKw

Guess you ar more behind than I thought.

Show me...you have a nag of dismming links and posting nothing but FUD...surprise me.

I can show you Havok CPU water:
www.havokCPUwater.com

Next time, bring something to table...besides your ignorance

ignorance or you just simply can't prove anything?
playing word games to get off the question isn't going to help, you still haven't answer any of my question beside physX on CPU vs GPU, which is irrelevant..
we all know how shitty physX on CPU is compare to any other engine out there, using physX as a proof it doesn't prove anything.

and physX rolling potato style fluid? or just simply ignore it with "HAVOK CPU WATER DOT COM"..

 

[phide]

we all know how shitty physX on CPU is compare to any other engine out there, using physX as a proof it doesn't prove anything.

Ain't bad for being free.

 

[shansoft]

Ain't bad for being free.

free and proprietary....

stop messing with my post

 

[Netrat33]

btw, the game you show is nothing ... I don't see anything that CPU/Havok can't do....and something is just too fake in there, you are creating infinite particle while hitting a object and the object never gets destroy LOL....
and please, do not show some weird rolling potato style fluid, it just not working or bring anything impressive but ugly and unrealistic.

http://www.youtube.com/watch?v=u3EoXN8kdbE

sure it's a tech demo but.....yeah. possible obviously.

 

[kadozer]

So is your post...just empty words with no meaning.



Why?
It would just be "nothing but a link with words"



When is too stupid too stupid?
Right now, you are more stupid than stupid.
Back in 2006 it was clear to everyone that the PPU was WAY faster than a CPU for physics:
http://www.youtube.com/watch?v=a7hIuytdRKw


Guess you ar more behind than I thought.




Show me...you have a nag of dismming links and posting nothing but FUD...surprise me.



I can show you Havok CPU water:
www.havokCPUwater.com

Next time, bring something to table...besides your ignorance

ur link doesn't work Bro

 

[shansoft]

http://www.youtube.com/watch?v=u3EoXN8kdbE

sure it's a tech demo but.....yeah. possible obviously.

like I said, rolling potato style fluid.... I rather have static with realistic looking than this ugly potato

it can look nice with smaller particle, but none of the GPU in current market can render it smoothly during the gaming experience...

 

[eloj]

Man, this thread is retarded. What the PhysX-proponents for some reason aren't willing to acknowledge, is that accelerated physics in games won't provide anything but worthless eye-candy unless all the major hardware devices go through one common API, so being on "one side" cheering is stupid. How is division good for the gamer? It's not. The only reason to support the current state is if you're a true company fanboi before you're a gamer.

Since the big players are currently AMD, nVidia and Intel and they'll never be able to play nice in the sandbox, what will happen is that a) Microsoft will come in and "mediate" (DirectPhys) OR b) One of the implementations will go open and win over the closed ones.

Anyone who thinks the API itself can go on and SHOULD go on being a point of contention, is a first class idiot.

Let me be clear: whatever-API can only win if hardware manufacturers are free to implement backends for it. PhysX could be that API I guess, but seeing how nVidia are behaving it'll take a major cluestorm me thinks. The important thing though; when/if that happens, this discussion is dead, because then we gamers and developers will have what we want, and no one in their right mind will care who owned the trademark.

 

[Netrat33]

like I said, rolling potato style fluid.... I rather have static with realistic looking than this ugly potato

you think that's potato water?



.......

ok

I thought you were talking about the original demos back in 2006. But that looks pretty good. *shrug*