NVIDIA Launches New SLI Physics Technology

Thomas Hines writes "NVIDIA just launched a new SLI Physics technology. It offloads the physics processing from the CPU to the graphics card. According to the benchmark, it improves the frame rate by more than 10x. Certainly worth investing in SLI if it works."

You know what...

Sounds like an ATI-killer to me! What ever happened to the hype about dedicated physics chips?

Re:You know what...

"A monopoly is always bad for the consumer... this is one of the reasons socalism doesn't work."

You can have a socialist government, and market competition.
The USSR's "implementation" of socialism was flawed. Don't get that confused with actual socialism.

Re:You know what...

Your comment reminds me a bit of this article [uncyclopedia.org]. Concerning the reasons for the lack of success of the American Institute of Communist Studies' program for granting certificates certifying something that someone said is "communist":

"And lastly, for reasons unknown, the AICS decided that half of the advisory board would consist of Communists and half of Libertarians. Since Communists believe that practically no one is a Communist including each other; and Libertarians believe that just about everything is indicative of Communism including most extant forms of Capitalism, the board reached an impasse in about half a second. "

Re:You know what...

Soviet Russia was very technically progressive. While being bad for the consumer, as it were, communism or socialism isn't necessarily bad for innovation.

Re:You know what...

I agree with you. But the same goes for the US. In the cold war - we went from scratch to repeat moon landings in ten years. Not so now. My point is that democracy doesn't have an inherent monopoly on innovation.

"Physics"

This is a little misleading. The hardware is really just fast at computing, not specifically designed for "physics". For example it doesn't have a build in ODE solver.

Re:"Physics"

I dunno what company would release a game that needs to SOLVE ODE's on the fly...I imagine you'd solve the equations before-hand, and put them in a nice form where all you need to do is multiply/add terms. If a company wants a cloak to behave realistically in their game, I'm sure they just find the proper coefficients in development, and all the game has to do is crunch the numbers on the fly.

We already have that patent

We already have that patent. [animats.com] For some years, we were locked into a licensing and noncompete agreement, which is why we haven't done much in that area for a while except cash the checks. But that noncompete period is now over. Stay tuned for further developments.

Our approach produces better-looking movement than the low-end physics packages. We don't have the "boink problem", where everything bounces as if it were very light. Heavy objects look heavy. Our physics has "ease in" and "ease out" in collisions, as animators put it, derived directly from the real physics. When we first did this, back in the 200MHz era, it was slow for real time (a two-player fighter was barely possible) but now, game physics can get better.

Take a look at our videos. [animats.com] Few if any other physics systems can even do the spinning top correctly, let alone the hard cases shown.

Improves framerate by 10x

This physics system is used for visual physics (i.e., realistic graphical effects), not gameplay physics, which are still done on the CPU.

Therefore you get a 10x framerate increase over running massively intensive effects on the CPU.

This is good, because games will look nicer. But if you don't have the GPU grunt, you can simply disable (or cut them down) them in game - it won't affect the gameplay.

SLI?

Why does this require SLI? You can do stream processing on most relatively-modern accelerated 3d video cards.

Re:SLI?

It doesn't, according to the article.

Nice

This will be critically important as programs start to push particle and geometry modeling. I remember back when I had my Quadra 840av in 1993, I popped a couple of Wizard 3dfx Voodoo cards in it when they first started supporting SLI and the performance benefits were noticeable. Of course we were all hoping for the performance to continue to scale, but 3Dfx started getting interested in other markets including defense and then were bought by Nvidia making me wonder if SLI would ever really take off. It's nice to see that the technology is still around and flourishing.

co-processor

How does this work in relation to AMD's consideration of a physics coprocessor or another specialized processor? It seems like that solution is superior.

General purpose GPUs

I've been waiting for this for a while. It's the obvious next step in GPU design. I have a feeling GPUs are going to become more and more general, and eventually accelerate the majority of inherently parallel processes, while the CPU executes everything else. We don't even have to change the acronym. Just call it a "Generic Processing Unit"...

Press release.

Of course its nothing more than a press release but there are numerous questions it raises:

1) What limitations are there on calculations. A GPU is not as general as a cpu and it would probably suck when dealing with branches especially when they aren't independant.

2) How much faster could this actually be. Is it simply a matter of looking to the future? (ie: we can already run with Aniso and AA and high resolutions so 5 years from now they'll be "overpowered"). IMO the next logical step is full fledged HDR and then more polygons.

3) What is exactly expected of these. General physics shouldn't be, but i can understand if they do small effects here or there.

All the answers to your questions...

www.gpgpu.org

Before people get too excited...

I don't think this is a general physics processor. It seems to be aimed at "eyecandy" physics calculations - mostly particle systems - whose results don't need to feed back into application logic. Which makes sense, given than GPU->CPU readbacks are a notorious perfomance killer.

Potentially shiny, but not really revolutionary or new. People have been doing particle system updates with shaders for a while now.

Re:Before people get too excited...

given than GPU->CPU readbacks are a notorious perfomance killer.

That has not been true for a long long time. Since PCIe became a standard, bidirectional communication between CPUs and GPUs has been as easy as unidirectional communication.

not limited to NVIDIA chips

This neither requires SLI nor is it limited to NVIDIA chips. NVIDIA is just launching it publicly. ATI will be showing it off behind closed doors this week.

10x faster?

10x faster? They might as well just say it's infinity times faster so that we know they are bullshitting from the second we read it...

Re:10x faster?

10 times faster is not all that unreasonable.

I used brook to compute some SVM calculations, and my 7800GT was about 40x faster than my Athlon64 3000+ (even after I hand-optimized some loops using SSE instructions). So its perfectly understandable for physics to be 10x faster on the GPU.

I don't understand?

By offloading physics from the CPU to the graphics card, this improves frame rates?

Why would I waste precious GPU processing to process Physics? I mean, all the CPU does these days is handle AI, physics, and texture loading. If you offload physics to the GPU, then the CPU is doing less and your swamping the GPU with more work.

If it does increase frame rates, then I would suggest why not improve graphics rendering rather then physics processing. I find that for all the advances nVidia and ATI have made over the years, 3D gaming visual quality is still inferior to cinematic quality 3D rendering. I mean, playing F.E.A.R, a relatively new game on the market, with ALL the settings to maximum, while I get 12 FPS the image quality just isn't that great on a current generation card.

I would prefer if nVidia and ATI actually focused on bringing cinematic quality 3D rendering to gaming, instead of just claiming they do. I want smooth high-poly models with realistic lighting and 60fps. I could care less about a game running at 120fps that looks bad. All 3D games suffer from a kind of mundane pseudo style of 3D modeling that leaves relatively well designed models playing in big rectangles with high-res texture cheats. Give me more lushes organic environments. Bring nurbs into the mix by creating actual curved surfaces into real time 3D rendering instead of just lots of triangles mimicking a curved surface.

So, while nVidia may have its heart in the right place, the last thing people need is their GPU being taxed with physics processing. Isn't there supposed to be a physics add-on card entering the market soon anyways? Won't multi-core CPU's offer better physics performance then a single GPU? Instead of trying to compete against add-in cards and multi-core CPU's, nVidia should just focus on improving 3D rendering quality and actually start delivering on their promises of offering cinematic 3D rendering to each new generation of video card they hype about.

Re:I don't understand?

Well, I for one, want to have a smarter AI in all games. Unloading the "mundane" physics engine to the graphic card will hopefully spare more CPU cycles for the AI. After all, it's not graphics that matter in games. It's the gameplay.

Re:I don't understand?

By offloading physics from the CPU to the graphics card, this improves frame rates?

Yes. Why does that surprise you? When you do incredibly complicated physics simulation, things can be very parallel and consequently GPUs outperform CPUs.

Why would I waste precious GPU processing to process Physics? I mean, all the CPU does these days is handle AI, physics, and texture loading. If you offload physics to the GPU, then the CPU is doing less and your swamping the GPU with more work.

You seem to be under the impression that your GPU cycles are more important than your cpu cycles. This is done with SLI for a reason..

If it does increase frame rates, then I would suggest why not improve graphics rendering rather then physics processing.

Because the quality of the render is controlled in software? Because hardware is currently limited by, ya know, physics and technology?

I find that for all the advances nVidia and ATI have made over the years, 3D gaming visual quality is still inferior to cinematic quality 3D rendering.

And in other news, offline processing is still more powerful than online processing. There's a shocker.

I would prefer if nVidia and ATI actually focused on bringing cinematic quality 3D rendering to gaming, instead of just claiming they do.

First of all, 99.9% of what nVidia and ATI do is exactly that. They are also starting to realize that the GPU paradigm, with minor modification, can be turned into a very powerful co-processor... and they are the experts at creating those types of chips. The market for them is growing... and they don't want to miss the boat.

I want smooth high-poly models with realistic lighting and 60fps.

And I want peace in the middle east. Give it 10 years, one of us may get our wish.

Re:I don't understand?

I want smooth high-poly models with realistic lighting and 60fps.

And I want peace in the middle east. Give it 10 years, one of us may get our wish.

Well, compared to 10 years ago, we probably HAVE cinematic quality rendering in games, and we definately have smooth high-poly models with realistic lighting and 60 fps. Trouble is that apart from 60 fps, every thing else in that statement is a very moving target.

Competition

Don't forget that http://www.ageia.com/ [ageia.com] is already doing this, and set to ship their cards sometime this year hopefully. Of course the significant difference between the two is that you would only have to buy one card for the SLI solution.

PCI Express

Why not have a complete physics card? It would be a nice use for that PCI express bus which only has video cards as an option right now. That way you could just buy the physics card, without having to upgrade the video card. Although this is all kind of weird. Start offloading everything off to specialized cards, you pretty much have a multiple CPU machine, where each CPU is specially tuned to do a specific type of processing. Might be the leap necessary to maintain Moore's law.

Articles with more bite

The one linked is a little bland for my taste

this one is better:

http://www.tgdaily.com/2006/03/20/nvidia_sli_forph ysics/ [tgdaily.com]

Or choose you own adventure via Google news:

http://news.google.ca/news?client=firefox-a&rls=or g.mozilla%3Aen-US%3Aofficial_s&hl=en&percentage_se rved=100&tab=wn&ie=ISO-8859-1&q=NVIDIA+SLI+Physics &btnG=Search+News [google.ca]

10x?

...it improves the frame rate by more than 10x.

Hmmm, so if I'm getting 10FPS in some game, then it'll boost it to around 100FPS? That can't be right...

Just more load balancing

A few years ago, they were being slammed for doing load balancing where they offloaded graphics processing onto the CPU when/if the CPU was less busy than the GPU. Now the GPUs are enough faster that they can frequently expect to be "ahead" of the CPU -- so now they're starting to work on doing the opposite, offloading work from the CPU to the GPU instead.

Of course, the basic isn't exactly brand new -- some of us have been writing shaders to handle heavy duty math for a while. The difference is that up until now, most real support for this has been more or less experimental (e.g. the Brook system [stanford.edu] for doing numeric processing on GPUs. Brook is also sufficiently different from an average programming language that it's probably fairly difficult to put to use in quite a few situations.

Having a physics-oriented framework will probably make this capability quite a bit easier to apply in quite a few more situations, which is clearly a good thing (especially for nVidia and perhaps ATI, of course).

The part I find interesting is that Intel has taken a big part of the low-end graphics market. Now nVidia is working at taking more of the computing high-end market. I can see it now: a game that does all the computing on a couple of big nVidia chips, and then displays the results via Intel integrated graphics...

I don't like it

While I agree that it will improve performance substantially, I fear that it will make games and 3D applications very video-card dependant. If nVidia's physics are slightly different than ATI's future implementation, then I think it will open up a world of new problems.

Math coprocessor?

I wonder if these cards could be useful for for Numerical computation? I could use extra cpu power for solving a linear system.

Why use a GPU, use a PPU

So you are scarificing Visual Quality for framerate and physics. Why not buy a dedicated physics card and get the best of both?

Re:Why use a GPU, use a PPU

I think the point is that this is for games where the bottleneck is in the CPU and the graphics card is sitting idle half of the time. By pulling 10% of the graphics card's resources to physics calculations, you could offload enough of the work from the CPU that it could keep the rest of the card completely fed and see a framerateimprovement with no additional hardware or loss in video quality.

This is a bad idea

From a consumer stand point, not many people are going to spend $600 to $1000 to get 2 video cards that can do really spiffy graphics. Its a very niche market which means no game company is going to spend an extra few months developing eye candy for a graphics card that a very small population uses. More likely we'll see ATI come out with a competing product that won't be so expensive and will attract more buyers.

Hardware is not the only preformance answer

Applications should be built to be more efficient, to handle modern hardware, instead of simply relying on consumers purchasing faster hardware.

parallel processing -- just add GPUs

I've used GPU for parallel processing of CPU heavy data in the past. You can use it for anything computationally heavy.

In my case, I've used it for stereoscopic image analysis in realtime. The best part about it is that you can add GPUs to a single box.

I don't know, but I would suspect that some of the bigger animation shops use multiple GPUs in a single machine for their rendering operations.

The PURE EVIL contained in modern graphics cards..

Modern graphics cards can be used to bypass security measures as an unprivileged user (reading kernel memory, say). Theo de Raadt of OpenBSD reminded [theaimsgroup.com] users how modern X works:

I would like to educate people of something which many are not aware of -- how X works on a modern machine.

Some of our architectures use a tricky and horrid thing to allow X to run. This is due to modern PC video card architecture containing a large quantity of PURE EVIL. To get around this evil the X developers have done some rather expedient things, such as directly accessing the cards via IO registers, directly from userland. It is hard to see how they could have done other -- that is how much evil the cards contain. Most operating systems make accessing these cards trivially easy for X to do this, but OpenBSD creates a small security barrier through the use of an "aperture driver", called xf86(4) (...)

While there are lots of funnies off of this...

... Most of you didn't get the point. It's not that you can access the GPU from userland (it depends on that access, but that's not the point). The main point is that that the current gen of programmable GPUs allow you to (theoretically) directly access kernel memory, as pointed out later in the thread by Theo:


> Are these new programable cards capable of reading main memory, which
> OpenBSD would not be able to prevent if machdep.allowaperture were
> set to something other than 0?

Yes, they have DMA engines. If the privilege seperate X server has a
bug, it can still wiggle the IO registers of the card to do DMA to
physical addresses, entirely bypassing system security.


Thus, a resourceful attacker theoretically could get access to kernel memory through anything which allows access to the video card. An unusual and probably difficult-to-exploit hole, but a possible hole none the less.

soooooooo

now the GFX card processes all the gaming and the CPU is only responsible for launching and saving your games? well.. they should call it the NVidia gaming system on a board.

Ageia making physics card

A company called Ageia is making a physics processing card [gamespot.com] that will handle physics calculations. It will be supported by City of Heroes/Villains [vnewscenter.com] when it is available.

NO SLI NECESSARY

from the article:

"Currently, the new Havok FX engine will support SLI physics even for a single NVIDIA GPU. But NVIDIA states that an SLI configuration is still the preferred mode since it allows the second GPU to be dedicated for effects physics processing." ...so, while yes, an SLI setup would be recommended for such intensive computation, it's not necessary. Older cards like the 6800GT should be able to handle it fine.

well, they COULD but...

Real-time cinematic quality graphics rendering = HARD.
Physics acceleration that allows for rather impressive collisions and water: MUCH EASIER.

Maximum output for minimum input. Having physics acceleration in the GPU makes sense as you don't have to buy an extra accelerator card.

Isn't da what the cell is supposed to do in PS3?

Mmmm ....

I thought that's what the IBM/Sony/Toshiba's cell broadband processor was supposed to do in the PS3? Oh well.... ...And what do we need that old x86, (dual, quad, whatever) for? Refresh my memory will ya?

Long live multicore and/or multiprocessor technologies!!!! But mercy on the po' programmers though ... Long live C/C++/Java multiprocessor extensions!

Time for a paradigm shift.

OK, I'm done. .smm.

Dedicated physics processing = overkill

My gaming rig has an FX-55 to do all of that physics number crunching--and it does quite a good job of it, too.

Maybe the most hardcore PC gamers or professionals out there could use a PPU so they can get the absolute most out of their hardware. For everybody else, it's just the latest "cool thing," and I really hope it doens't catch on.

Can't read the article...

I can't read the article since it's slashdotted, but here's what I want to know:

First, what physics API are they using? This is, after all, a little like OpenGL vs DirectX. You need a physics API to do this stuff, and there are out there a *lot* of portable and high quality APIs. Havok, Newton, Aegeia (spelling?), and the open source ODE ( which I use ). The APIs aren't interchangeable, and aren't necessarily free.

Second, at least when I'm doing this work, there's a *lot* of back and forth between the physics and my game engine. Maybe not a whole lot of data, but a lot of callbacks -- a lot of situations where the collision system determines A & B are about to touch and has to ask my code what to do about it. And my code has to do some hairy stuff to forward these events to A & B ( since physics engines have their own idea of what a physical object instance is, and it's orthogonal to my game objects, so I have to have some container and void ptr mojo ) and so on and so forth. If all this is running on the GPU, sure the math may be fast but I worry about all the stalls resulting from the back and forth. Sure, that can be parallelized and the callbacks can be queued, but still.

Anyway, I want info, not marketing.

Oh christ, and finally, I work on a Mac. When will I see support? ( lol. this is me falling off my chair, crying and laughing, crying... sobbing. Because I know the answer ). Can we at least assume/hope that they'll provide a software fallback api, and that that api will be available for linux and mac? After all, NVIDIA has linux and mac ports of Cg, so why not this? I'm keeping my fingers crossed.

New server tech!

Microsoft has also announced a new server technology that unloads all calls from Slashdot to a seperate system avoiding the dreaded "Slashdotting" effect.

Liars

"..improves the frame rate by more than 10x"

Liars end up in Hell.

Check out gpgpu.org!

This is definitely not new news at all. The general-purpose graphics processing unit community has been active for quite some time, see gpgpu.org. I think Nvidia may be using this to just push their quad-gpu / SLI setup as it would be pretty hard for the main cpu to keep all of those gpu's busy 100% of the time. By offloading physics, and numeric solvers, etc, you can do some interesting things. Think of textures as your inputs, and render-targets as your feedback from the gpu. However, this is not new, as many of you have pointed out already.

How about a video card....

.....that makes games fun to play?

FPS are a perfect example of where this will sell

BF2 for example. This game is ALL physics... I love it personally, and one of the coolest (and crappy) things is when you get shelled by artillary 500 yards into the air. Your limbs are flying everywhere and the FPS decrease is noticeable when there are 10 or so soliders all ragdolled in the air (i'm sure particles have a lot to do with this also, but IAN a FPS programmer). FEAR is another great example. This is just the turning point in realism for FPS games, and that is very cool IMHO.

Forget 'physics' - give me a good math API

The guys over at http://www.gpgpu.org/ [gpgpu.org] have been doing various math calculations, including 'physics' on GPUs for a while now. One big problem is that the only real API is OpenGL. So not only do you have to be a smart math programmer (which is pretty rare to begin with) but you also have to understand graphics programming too and then figure out how to map traditional math operations onto the graphics operations that OpenGL makes available. It isn't that hard to do simple things like matrix math, but trying to really optimize it for really good performance requires almost wizard-level understanding of OpenGL and the underlying hardware implementation.

The cards' math capabilities would be so much more accessible (and thus used by so many more programmers) if Nvidia (and ATI) would come out with standard math-library interfaces to their cards. Give us something that looks like FFTW and has been tweaked by the card engineers for maximum performance and then we will see everbody and his brother using these video cards for math co-processing.

My prediction

By the end of next week there will be a game out on the market that will require one of these cards...

Er, Not Exactly

Certainly worth investing in SLI if it works.

How about, Certainly worth investing in SLI if it works on the specific game(s) you most want to play at higher speeds and/or resolutions.

Otherwise, not worth investing in at all.

Confidence

I love how all of those tech-review pictures had "Confidential - Do not distribute" on them. Oh well, NVIDIA deserves to get their crap looted, since they're stealing the entire idea as a way to undercut and steal Ageia's idea.

Unintended Gameplay Effects

The (next) holy grail of games is the fully destructable environment, where damage isn't a sprite on a plane, but actual particle rearrangement of walls and buildings and such. This is a pretty strong step in that direction.

The problem is that the internal logic of games like Quake and Half-Life is that, if the environment reacts correctly with respect to the physics of weapons, any game environment will quite quickly be reduced to piles of rubble and little more. Think of the pictures of Europe in WWII, and then imagine the map after a few rounds using the BFG. All those clever hidey-holes and corner sneaking tricks become useless when the terrain is quickly reduced to vaguely hilly, rolling rubble landscapes.

There'll be some dodges, like "this weapons does radiation damage, not explosive, so that's why the gyprock walls of the apartment building aren't destroyed while everything inside is pulped," but fundamentally, we're headed for a design crisis where the only remaining plausibility will be found in historical combat sims.

SLI stands for... ???

So I was SOL trying to find out what SLI was. I tried to RTFA but it was /.ed and gpgpu.org was NFG either. When I was in school, it was mandatory to use the Full English Version (FEV) of an acronym the first time used in an essay. Is it such a unreasonble request?

Old School is new again?

While I hate to ride a horse into the ground and then feed off its bones, every time I hear something like this happening I immediately think "Amiga". Why? I would guess that it is because the Amiga had a CPU and then it had dedicated chips to handle other functions - math, graphics, sounds, etc. This arrangement created a computer system that did not get surpassed for MANY years after its demise and, some would say, it still hasn't been bested in many areas (multi-tasking is one of those).

Each time I hear that an "advance" has been made and I read that it is basically re-integrating various components back into the primary system or tying those components tighter to the CPU then I can't help but scream "AMIGA!" Of course, this leads to co-workers walking wider paths around me while having avoiding eye contact '-).

Still, all of these advances lead me to believe that we might going back to a dedicated chip style of computing BUT what I am also hoping for is a completely upgradeable system that I can pull the, say, physics processor out and plug a newer version or better chip into without having to replace the entire motherboard or daughterboard. Which, of course, leade me right back to that whole screaming scenario :) The Amiga style of computing may yet live again.

BS

It offloads the physics processing from the CPU to the graphics card. According to the benchmark, it improves the frame rate by more than 10x.

So in the best case, physics is currently taking at least 9x more computation than all other graphic card activities and offloading to the CPU reduces the cost of physics to 0.

Now in the less than best case...this claim is bullshit.

Why SLI?!

WTF does SLI have to do with this?

Surely if they can do this on two video cards, they can do it half as fast on one (framerates 5x faster are fine by me...).

10x Faster?

I can already get Doom 3 running at ~70 fps with pretty high graphics settings - are they seriousyl trying to tell me that by adding a second video card and twaeking a little I could get 700 fps?

I'll believe it when I see it (which I won't).

A previously-announced physics processor unit?

Didn't a company called Ageia (?) design a PCI-express addon (or PCIx or wtv) that was basically a separate chip completely dedicated to physics calculations (ragdoll thingies and that sort of stuff)?

In fact, wasn't the PS3 supposed to have said chip from Ageia (or wtv)?

This would be cool, but i wonder how many would actually flock to it (if cheap enough (~40) then probably it would lead developers to assume its existence, and if not to default to using good old ix86).

Calculate physics? Impossible, I dare say!

Next week's headlines:

Nvidia solves Chaos Theory!

waste of money

Certainly you do not need a second 7900gtx to compute physics effects in games. This is just nVidia "putting their hands on your stack." GPU's are already incredibly good at doing these sorts of computations. A $100.00 video card would probably suffice. nVidia has designed SLI to work with symmetric board configurations but I don't see why they couldn't do a redesign to offload just physics computations to the second GPU (since they state that the intended configuration is to work with single gpu "next gen" configurations). Of course this would repurpose obselete and low end hardware which is something nVidia most definitely does not want to do. However, ATi's 'crossfire' seems to work with mismatched boards so when they implement this it will certainly be more cost effective.

Re:Investing in? huh?

Anything that can return more than you paid for it is potentially an investment. Cars and jewelry can both fall into that category. As could something like insurance, art or starting a company.

-bZj