New Mafia II PhysX ON/OFF video.

Page 18 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Lonbjerg

Diamond Member
Dec 6, 2009
4,419
0
0
While I'm fairly certain you misunderstood Dribble, and engaged in an unnecessary attack on him, I'd still very much like to hear your level of expertise in compiling for PhysX. WITHOUT the nastiness if you please.

I first started using the PhysX SDK when I registered as an developer on AGIEA's site back in 2006....in May I think.

I have also used NOVODEX's Rocket since version 1.0
(A shame they dropped that one, becase it was a great entry tool to physics and the code/forces required to achieve the desired result when first entering the world of real-time physics)

I have compiled PhysX code (for my own intererst mainly) since 2006 too and have quite a large experince in using the the various SDK's.

(Now that irrelevant facts are out of the, care to point to the relevance?)

And anyone thinking that Mafia 2 on PC is a simple "Xbox port", can just look for themselfes:
http://www.youtube.com/watch?v=qzkIhgp27r8

Or anyone thinking that the eg. Apex dosn't add to game game can look here:
http://www.youtube.com/watch?v=ONQfbGBtitY

the little, subtle movement of cloth makes the scenes come more to life...oppose to the dead static gaming world we have as the norm now.

The PhysX level on console are similar to "software physics" (aka CPU) on the PC...there is no "GPU-PhysX" equvilant on the consoles, since all PhysX code on the consoles run on their CPU, NOT their GPU's...so comparing CPU physics to GPU physics the way Dribble did, only speaks of argumentation based on ignorance.

And that is really, really bad.
 

Pantalaimon

Senior member
Feb 6, 2006
341
40
91
If all of us are THAT into technology, then no matter what PhysX does not offer at this time, the next moves should be anticipated rather than bashed or a call for it's demise.

Not me if it means giving power and locking myself to one hardware vendor with sleazy marketing tactics like NVIDIA
 

Dribble

Platinum Member
Aug 9, 2005
2,076
611
136
You can't even rember your own words?:



That part make is sadly obvious that you don't know what you are talking about in regards to PhysX and how you compile for PhysX.

That for confirming my suspesions.

.oO(Sad that most at the bashing of PhysX, on a tech-site none the less, is based on simple ignorance o_O)

Actually I am a software dev, I am a senior part of a team that write large graphical software apps for a living. It's not for games but for the engineering market. It is however multi-threaded, multi-platform, and has plenty of 3rd party middle-ware. I have been doing this for a number of years - well before PC's had multiple cores, or hardware physics, back when SGI was *the* platform to develop on. I'm pretty darn good at it.

I'd don't develop using physx, but I do understand very well the sort of issues and priorities game devs will have developing these cross platform games and using 3rd party libraries like physx.

You however just sound like a fanboy, who desperate to defend what they love is going for the "try to confuse them by talking technical and sounding superior" method of arguing. This is often done despite not really understanding themselves - people who really understand tend to try and explain.

The irony of all this is I wasn't even bashing physx, if anything I was defending it, but you seem so on the defensive you didn't even read what I wrote and just took it as an attack?
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
I was hugely excited with the PPU when it was first announced, and I still beleive in the basic idea. We need an industry standard though, before game developers use physics in new and interesting ways.

I was also hugely excited when it was announced, and hugely disappointed...
but I don't see how industry standards magically fix the issue.
We do NOT need an industry standard, we need better middleware; PhysX, havok, etc... those are middleware, and none of them is revolutionary... there is no money for each game developer to write their own from scratch either. We need a company like unreal to incorporate a physX engine which utilizes the CPU AND various different GPGPU technologies (CUDA, openCL, whatever).
Until we get some awesome middleware we have to make due. I do not RESENT any of the middlewares for existing, I don't see why some do. But on the other hand, I can see why some might criticize some of those middleware, they are far from perfect. PhysX is the only middleware AFAIK that is CAPABLE of revolutionary first order physics... but nobody in their right mind will use it for that because that means nobody with an ATI card can play such a game...

BTW, I think gimmicky is a bad choice of words... because I wish we had more actual gimmicks.
We see uninspired things like more particles... bit woot. frankly I wish we had more gimmicks and less uninspired blandness... something that is a fundamental to the gameplay rather then a tacked on and superfecial eye candy. we need first order physics.

Problem is, innovation is dangerous, you can make a whole lot of money playing it safe... you can lose a whole lot of money innovating....
More than likely, we need an indie company to step up and make something hugely popular based on revolutionary first order physics, which means it would be a physX exclusive. Then all the top budget titles to follow will rip it off. When that happens companies will see the need, after a few nvidia only titles are released, we will see the birth of platform agnostic middleware (probable openCL based) which new games will then come out on.

Not me if it means giving power and locking myself to one hardware vendor with sleazy marketing tactics like NVIDIA

did the existance of eyefinity lock everyone into AMD only hardware? no, nvidia released their own version as soon as they saw it is good.
did the existance of 3dvision from nvidia lock everyone into nvidia only hardware? no, AMD released their own version as soon as they saw it was good.

Making a first order physX middleware that runs on both AMD and nvidia hardware isn't that difficult, its just not shown to have ever been profitable (only company that tried, Aegia, did not fare well... then bought by nvidia which can't seem to get the technology to flourish, yet).

We will not see a "lockin", we will see a FEW nvidia exclusive games... followed by an AMD version, and quickly afterwards a/multiple platform agnostic middleware.
 
Last edited:

Pantalaimon

Senior member
Feb 6, 2006
341
40
91
did the existance of eyefinity lock everyone into AMD only hardware? no, nvidia released their own version as soon as they saw it is good.
did the existance of 3dvision from nvidia lock everyone into nvidia only hardware? no, AMD released their own version as soon as they saw it was good.

I'm not sure what your point here is? The discussion in this thread is about PhysX, and I already said what I think of it and the company that owns it.
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
I'm not sure what your point here is?
I spell my point out in the next sentence.

The discussion in this thread is about PhysX
And I list other innovative technologies and their impact (or lack of) on the market in relationship to the theorized impact (nvidia lockin)

and I already said what I think of it and the company that owns it.
so? how does this in any way shape or form rebuttal my predictions?
 

Pantalaimon

Senior member
Feb 6, 2006
341
40
91
so? how does this in any way shape or form rebuttal my predictions?

I don't if it's a rebuttal, but considering the way that NVIDIA conduct their business strategy so far when it comes to GPU PhysX, there's no hope for it. At least IMO.

PhysX will be restricted to overdone gimmick effects, and will fail to set itself as a major middleware standard for GPU physics.
 

Ares1214

Senior member
Sep 12, 2010
268
0
0
I hope physx doesnt become standard, it takes things that might not look detailed enough since they wanted to add physx, and they then just make it so unrealistic looking.
 

Lonbjerg

Diamond Member
Dec 6, 2009
4,419
0
0
I hope physx doesnt become standard, it takes things that might not look detailed enough since they wanted to add physx, and they then just make it so unrealistic looking.

Do you think Havok's scripted wall-holes are better :confused:
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
The way I look at it is, it's good to see Physics moving forward. It's not really a battle of the CPU, GPU, Pc platform or console, or nVidia, Intel, AMD but trying to improve gaming to me. It's so important, the leaders are trying to promote their strengths on how they will show leadership -- which is a good thing because it builds upon improving physics.

The CPU and GPU are very important and would like to see developers utilize their strengths.
 

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
I don't if it's a rebuttal, but considering the way that NVIDIA conduct their business strategy so far when it comes to GPU PhysX, there's no hope for it. At least IMO.

PhysX will be restricted to overdone gimmick effects, and will fail to set itself as a major middleware standard for GPU physics.

Imho,

It may or may not be, but that's not the only mechanism for this PhysX GPU vehicle. I think the idea is to showcase what the GPU may do to enhance Physics and help build awareness so there may be a standard of some kind. If it is Bullet, PhysX, Havok, or someone else, as long as the GPU is part of it -- think nVidia would be pleased because it is about the GPU.

The ability to differentiate and provide value for the GeForce brand and their ecosystem is another point, and considering the costs, risk, responsibility and accountability, personally don't get too angry if there is some sort of differentiation or leverage. Don't expect idealism considering Physics hasn't matured and evolved enough yet -- where in its infancy, considering I believe Physics is the next frontier, so-to-speak.
 

Lonbjerg

Diamond Member
Dec 6, 2009
4,419
0
0
*snip*
....considering Physics hasn't matured and evolved enough yet -- where in its infancy, considering I believe Physics is the next frontier, so-to-speak.

My thoughts to the letter, back in 2006 when I heard about the PPU...followed by a:

/DOH Why didn't I think of that...
 

railven

Diamond Member
Mar 25, 2010
6,604
561
126
I'll take ANY hcange over the "*boom* - *debri* - *debri inactive* - *debri gone" we have now.
Scripted animations need to go for physics simulations ASAP.

While watching the video with the girlfriend we both noticed that exact thing you described:

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

Start watching at time mark 0:53 and look on the Apex On side. You'll see the debris shrink and disappear.

So, I guess outside of persisting a tad longer, it's relatively the same as your description.
 

Lonbjerg

Diamond Member
Dec 6, 2009
4,419
0
0
While watching the video with the girlfriend we both noticed that exact thing you described:

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

Start watching at time mark 0:53 and look on the Apex On side. You'll see the debris shrink and disappear.

So, I guess outside of persisting a tad longer, it's relatively the same as your description.

Some of the debri dissapear, but the majority remains..and are interactive.
It progress...but not perfect.
We won't get there in many years...kinda like with 3D rendering...incremental steps...but the CPU isn't the way to go for massive parrallel SIMD
 

Red Storm

Lifer
Oct 2, 2005
14,233
234
106
Real GPU Physics (as in game play affecting physics, not just visual effects) will really start to take off once the CPU/GPU chips from Intel and AMD go through a couple generations and become mainstream. Why? Because it doesn't require you to buy another chip/card. Everybody building/buying a new computer at that point will also be GPU physics capable, so developers won't have to worry about coding in content that won't be utilized by the majority. So you can see why nVidia has been very aggressive with their PhysX marketing. There is no way for them to legally build a CPU/GPU chip of their own, and it looks like that is where GPU physics is heading (it certainly seems to make sense at least).
 

klansek

Junior Member
Sep 14, 2010
10
0
0
I find it funny, that the video with physx on/off reminds me about 3d mark 2001 where in the hall scene there was similar effect displayed...bunch of wall debris and shells spraying around (i think it used havoc and it runs at 50fps on ibm t41 from 2004: single core 1.6ghz with ati radeon 7500 dx7 card). To think, that 10 years after, with all this power at hand, you need a special gpu to do almost the same, is just ridiculous. SW developers of mafia II could do the same and much better on the cpu...using havok or physx...of course if they would want to. In other words, the whole issue at hand is a PR. Anyone who believes different has a false consciousness...like believing that using shampoo against hair loss will stop you going bald. :D
 
Last edited by a moderator:

Kenmitch

Diamond Member
Oct 10, 1999
8,505
2,249
136
I find it funny, that the video with physx on/off reminds me about 3d mark 2001 where in the hall scene there was similar effect displayed...bunch of wall debris and shells spraying around (i think it used havoc and it runs at 50fps on ibm t41 from 2004: single core 1.6ghz with ati radeon 7500 dx7 card). To think, that 10 years after, with all this power at hand, you need a special gpu to do almost the same, is just ridiculous. SW developers of mafia II could do the same and much better on the cpu...using havok or physx...of course if they would want to. In other words, the whole issue at hand is a PR. Anyone who believes different has a false consciousness...like believing that using shampoo against hair loss will stop you going bald. :D

Welcome to the forum and great first post....Wraps it up pretty well.
 
Last edited by a moderator:

taserbro

Senior member
Jun 3, 2010
216
0
76
I never made it sound like a fact. My first had "it seems" not "it is: and my second post stipulated that it's just an observation. In any case, even though there might be some things that point to foul play it's impossible to prove. It's also impossible to prove otherwise. Obviously if there was no perception of Nvidia crippling hardware physX titles no one would be riled up. If there was proof then everyone would be riled up. Right now there is only an unprovable perception, so it makes sense that it riles up people but to a lesser degree, especially whith other physX drama

That's perfectly fine with me. I didn't insinuate that you did; simply that it's the kind of tone I get from almost everyone who chooses to bring it up. I wear way too many tinfoil hat layers to be pointing fingers here but I don't hold it against anyone unless I have proof shenanigans took place.

The car analogy was extreme, granted, but it was valid. Marketing your product as something that it isn't is false advertising, but it doesen't work the other way around. Not advertising your product's limitations doesen't magically get you off the hook in this kind of situation. Here's a less extreme example: Imagine if Half-life 3 came out without the ability to run on Windows 7, people bought the game assuming (as they should) that it would, and Valve's reponse afterwards was "well, we never advertised that HL3 would run on Windows 7".

I'm still not convinced. Do nvidia cards have a risk of failing catastrophically as part of a feature it was designed for? Because that's the analogy I'm looking for, some feature that's advertised as something it was designed to do but fails to live up to that promise (whether by exploding or just not doing anything), not something it happens to be able to do and then was disabled for whatever purposes.

Also, this is starting to stray a bit too far from the point but if half life 3 did come out without windows7 support as a conscious decision, it would be fine with me; your example raises eyebrows because windows is what runs games almost everywhere so people would expect it to run on windows that but had it been lack of support for linux, nobody would have cared because they wouldn't expect it like I personally don't expect nvidia to make something that'd directly benefit ati. I don't have a game box here but I went to check real quick and I'm pretty sure that my adobe cs5 box specified windows as a requirement but didn't directly state that it wouldn't run on a linux machine even if it does also technically have windows on the drive as well. There's plenty more of special cases like that don't warrant suing them for.

Yes it does. On the box it's advertised to perform PhysX. You don't actually figure out that it can't run phsyX coupled with an ATI card unless you search the nvidia website. And the justification on the website "complicated technical connections that only exist between Nvidia cards" is completely, demonstrably false.

Look, we could be arguing the wording and the scope of what they are obligated to represent but at the end of the day, the card is designed to run as a primary vga card. Saying it should also be able to do this because the wording has a loophole and that obligates them to do it is making an assumption that merits a bit of investigating before purchasing. Do ati 5xxx cards run eyefinity desktop windows like those fancy pictures on the back of the box while in a second slot while I run a game on my primary nvidia card? I'm not asking this to be cute, I honestly don't know. If they don't, that's a case for someone to go and look for fine prints too. I'm also pretty sure my new 1600 ddr3 kit won't run full speed when I pair them with the old 1033 epilda in the other 3 slots either, even though the package doesn't say anywhere that it can't run full speed in this special circumstance. Neither can my old motherboards boot on a usb drive, even though I'm somewhat sure all it'd take is a bios revision.

And of course, I'm not gonna be able to get a perfect example where it's 100% clear that it's possible but disabled on purpose, or not disabled on purpose but technically shaky but I don't think it means that nvidia has some obligation of covering every single special case in what would be a monolithic wall of fine print. Plus, as far as I know, nvidia cards can run as a physx card with an ati card as of now since nvidia did restore the beta driver that could do that; so strictly technically speaking, they do have the support everyone seems angry that they omitted to disclaim once in the past.

No. Not creating an "adaptor" for two things that otherwise *do not* function together is not the same as taking two things that *would* normally function together and sabatoging one of them so they no longer do. Not adding something != taking something away.

Sabotaging is a word that'll get you the chicks' attention but selectively disabling features isn't exactly new or illegal in this context as far as I know. What if the said car manufacturer had instead engineered the radio's shape so that it couldn't be retrofitted to improve its competitors' models then simply said that they don't fit, even though they could if they wanted to? I know a few companies that do that, albeit not actual car radios... but they do it solely to prevent people from buying their low end stuff and use it outside of what they meant to bundle it with.

Let's get things straight. I *do not* beleive there will be a lawsuit. To reiterate my stance, AMD could sue Nvidia anytime they wanted to but will not because Nvidia is shooting themselves in the foot by stiffling PhysX adoption. The last thing AMD wants is for PhysX to work with ATI cards.

That being said, if PhsyX ever does really take off, which is the giantest of long-shots, then you'll definitely see a lawsuit. I'm not a lawyer, but I do know that a common misconception is that anticompetitive law can only be used against big companies. From wikipedia because looking up the actual laws is outside the scope of a forum post:

http://en.wikipedia.org/wiki/Competition_law

"Competition law, or antitrust law, has three main elements: ...
banning abusive behavior by a firm dominating a market, or anti-competitive practices that tend to lead to such a dominant position."

That last part is why AMD could sue Nvidia at any time. If phsyX ever does take off they could sue nvidia under the first part (and obviously have an eaiser time making their case)

This is your interpretation of an incredibly vague line from wikipedia. I could sue the crap out of everyone and their pet turtles for anti-competitive practices if that's all it took.
It's really easy to build a case of outrage on that but once you're asked for some precedents in court, it might very quickly collapse. If so many people really feel this is a case, why haven't they sued nvidia for it already? It's not like they'd spare them because they appreciate their contributions to the field or anything and if they're so confident that they'll nail nvidia, why wait til someone else does it?
I'm still not convinced there's a case; maybe someone asked a lawyer friend or something and knows. I'd love to know as well. But until I do know, I wouldn't order a boat with my share of the damages if I were you.

Again, apples and oranges. Discontinuing support for a product is *way* different than continuing support but removing the functionality that was previously there to run alongside a completely unrelated part made by your competitor. It's not even apples and oranges. It's apples and chicken soup.

Would you still have the same opinion if it weren't nvidia we're taking about? Because to me, the case of my old laptop not being supported with newer apps anymore feels very much the same as if I were to have an ageia card that's somehow said to be supported but not for new titles (what does that even mean and what's the difference?): I still can't use those new titles be it because the guys who wrote the old drivers decided it'd be a waste of time, or because there's some evil cat-stroking guy deciding that he won't support old drivers. You can't take something away that never existed; if it existed, the aforementioned old drivers would be able to run the newer titles if you simply didn't upgrade them.

Or heck, it's quite easy to point and say 'you're doing it wrong' than find someone who's actually doing it right; riddle me this, which company still is supporting a product that it has previously acquired and works to improve one of its direct competitors' in a task the latter wasn't designed for? On the top of my head, I can't think of any but I'm hardly an authority on the matter and I'm perfectly willing to change my opinion if I find out this is a required practice but sincerly, those altruistic apples that would give up marketshare for the sake of their competing produce are starting to feel like a one of a kind fruitjob in a world of ruthless oranges.
 

zerocool84

Lifer
Nov 11, 2004
36,041
472
126
Personally, I don't care what becomes the industry standard, I just want it to happen already. PhysX so far has been a great dissapointment, but I still beleive that the idea can lead to great things once we have an industry standard and developers embrace it.

This is exactly how I've felt. The demos we've had of Physx have been awesome and make you go WOW but we as gamers are still waiting for what it can really do. We were promised a while ago it was going to change the way we play games and for the little effects we get it kills even top of the line hardware. It's great that we have things like Physx and Havok but so far the only one that's made me care at all is Havok cus they've actually done something to affect gameplay.
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
This is exactly how I've felt. The demos we've had of Physx have been awesome and make you go WOW but we as gamers are still waiting for what it can really do. We were promised a while ago it was going to change the way we play games and for the little effects we get it kills even top of the line hardware. It's great that we have things like Physx and Havok but so far the only one that's made me care at all is Havok cus they've actually done something to affect gameplay.

they have? AFAIK neither physX nor havok affected gameplay any.