Nvidia Releases Hardware-Accelerated Film Renderer

snowtigger writes "The day we'll be doing movie rendering in hardware has come: Nvidia today released Gelato, a hardware rendering solution for movie production with some advanced rendering features: displacement, motion blur, raytracing, flexible shading and lighting, a C++ interface for plugins and integration, plus lots of other goodies used in television and movie production. It will be nice to see how this will compete against the software rendering solutions used today. And it runs under Linux too, so we might be seeing more Linux rendering clusters in the future =)" Gelato is proprietary (and pricey), which makes me wonder: is there any Free software capable of exploiting the general computing power of modern video cards?

Comments

Spelling...

[B4RSK]

Gelsto is proprietary (and pricey), which makes me wonder: is there any Free software capable of exploiting the general computing power of modern video cards?

Gelato seems to be correct...

I like this...

[rjw57]

This is a reversion of the norm :) [from the page linked to in the story]:

Operating System

* RedHat Linux 7.2 or higher
* Windows XP (coming soon)

Re:I like this...

[Mister Coffee]

It does not happen often that a hardware manufacturer has Linux support before it has Windows support. At least I have never seen it before.

Re:I like this...

[Anonymous Coward]

now thats more like it

Re:I like this...

[arvindn]

Naturally. If the product is intended to be cross platform right from the beginning, then the developers would prefer to work on linux and port it to windows rather than the other way round, so you can expect the linux version to be released slightly earlier. What is perhaps a little surprising is that they announced it before the Windows version was ready.

Re:I like this...

[Anonymous Coward]

When thinking back to the earlier story about unsupported sound cards on Linux, does this mean that Windows XP is not ready for the desktop? ;)

Re:I like this...

[killmenow]

Nvidia couldn't be a little pissed they're out of the XBOX2, could they?

Re:I like this...

[Anonymous Coward]

I believe that they're happy not to be shouldering Microsoft's xbox losses.

Re:I like this...

[Minna Kirai]

At least I have never seen it before.

Look at the AMD 64 ("Opteron", etc) CPU. Linux support is here, but native versions of Microsoft Windows are still yet to be released.

Re:I like this...

[ameoba]

It makes sense, really... If you're building an app that's intended to be used by clusters, why would you write it for XP? Having to spend an extra $100 per node really starts adding up when you've got serveral hundred or thousands of machines...

3D graphics cards aren't relevant

[Anonymous Coward]

Sadly, the hardware accelerations that consumer 3D graphics cards do aren't useful for the high quality renderings that are needed for film and television. The needs of games are just different, parially because of the need to render in realtime. So I doubt whether there's much scope for free software to make use of them for that purpose...

Re:3D graphics cards aren't relevant

[niheuvel]

Perhaps you should read the Nvidia FAQ? This topic is covered. From what I can tell, they don't use the GPU in the traditional way, they just use it as a co-processor.

Re:3D graphics cards aren't relevant

[mcbridematt]

But, NVIDIA's Quadro lineup *ARE* PCB Hacked consumer cards. Some PCI ID(or BIOS for the NV3x cards) hacking can get you a Quadro out of a GeForce easily, minus the extra video memory present on the Quadro's. I've done this heaps of times with my GeForce4 Ti 4200 8x (to a Quadro 780 XGL and even a 980 XGL) and I believe people have done it with the NV3x/FX cards as well.

This film renderer is different. It uses the GPU and CPU together as powerful floating point processors (not sure if gelato does anything more than that).

Re:3D graphics cards aren't relevant

[Oscaro]

This is not really correct. The graphics cards Gelato uses are consumer hardware. This doesn't mean that the image is generated directly by the card! The 3D hardware is used as a specialized fast and parallel calculation unit, used especially for geometric calculation (matrix per vertex multiplication, essentially) and other stuff. This (of course) means that the rendering is NOT done in realtime.

Re:3D graphics cards aren't relevant

[WARM3CH]

Actually, there has been reports of using such hardwares to produce the similar results of the high-end, software based methods like those used in films. The trick is to break the job (typically the complex RenderMan shaders) to many passes, and feed them to the graphics card to process. By many passes, I mean 100~200 passes. The outcome will be like rendering a frame in a few seconds (we're not talking about real-time renderings here) which is MUCH faster the software based approaches. The limit in the past was that the color representaion inside the GPUs used a small number of bits per channel and by having a lots of passes on the data, round-off errors would degredate the quality of the results. But now, nVidia supports 32 bit floating point representaion for each color channel (i.e 128 bits per pixel for RGBA!) and this brings back the idea of using the GPU with many passes to complete the job. Please note that in the film and TV business, we're talking of large clusters of machines and weeks of rendering and bringing it down to days with smaller number of machines is a very big progress.

Re:3D graphics cards aren't relevant

[Anonymous Coward]

Pinnacle have been doing graphics card assisted effects for a long time under Adobe Premier see

http://www.pinnaclesys.com/ProductPage_n.asp?Produ ct_ID=19&Langue_ID=7

for details

Re:3D graphics cards aren't relevant

[Anonymous Coward]

Have a look at

http://www.pinnaclesys.com/ProductPage_n.asp?Produ ct_ID=19&Langue_ID=7

Re:3D graphics cards aren't relevant

[Anonymous Coward]

That was true a couple of years ago, however Geforce 6800 and Quadro FX 4000 have support for:

- 32 bit per component precision in the pixel shader
- dynamic branches in the pixel and vertex shaders
- unlimited instruction count for pixel and vertex shader programs

These features are very useful, they make possible to render frames using GPU at the full quality. Considering that GPUs have HUGE amount of processing power, this will make the rendering much faster.

For example, above mentioned new GPUs from NVidia have 32 FPUs in the pixel pipeline only. Each of those FPUs can perform 4 FP operations at 32bit precision per clock. That's 128 operations per clock or 51.2 GFLOPS in the pixel shader alone.

Re:3D graphics cards aren't relevant

[shplorb]

Exactly, floating-point colour channels provide a larger dynamic range, which prevents the banding and saturation you see when doing multiple passes with say 8-bits/channel colour. This has been the crucial development to enable this. Conditionals and looping for shaders wasn't generally required - as Renderman-style languages can be decomposed into a series of standard OpenGL operations. (I remember researching all of this for an essay I wrote on this sort of stuff at uni a couple of years ago - interesting to see some of my predictions being realised.)

Who cares if this system takes five or ten minutes to render a frame? That is a hell-of-a-lot more productive than taking 8 hours to render a frame via software!

An interesting bit regarding that is that I read somewhere that the compute time per frame for every Pixar film has remained about the same (8 hours). I wish I could remember the source for that. Imagine what could be achieved with 8 hours of hardware accelerated rendering time!

Imagine having a computer using PCI Express and whacking a gfx card in each slot - renderfarm-in-a-box!

The future is indeed exciting.

Re:3D graphics cards aren't relevant

[Totally_Tux]

How do you know this? Did you perform benchmark comparing it to a real Quadro?

A couple of years ago I got a GeForce4 4800 and a Quadro4 900 XGL. I performed the required resistor mod and flashed the GeForce4 with the Quadro4's BIOS.

Sure the GeForce4 got recognised as a Quadro4 900 XGL in the Windows display control panel, but when you run benchmarks like SPECViewPerf it was obvious the modded-GeForce4 did NOT perform like a real Quadro4 900 XGL. Capabilities like the HW-accelerated clip planes did not seem to present in the GeForce4, and this made a big difference to the scores I was getting.

Re:3D graphics cards aren't relevant

[HD Webdev]

Please note that in the film and TV business, we're talking of large clusters of machines and weeks of rendering and bringing it down to days with smaller number of machines is a very big progress.

It is a lot of progress.

On the other hand, it'll be interesting to see whether rending times come down, or animations get much more complicated.

Re:3D graphics cards aren't relevant

[dave1g]

"Imagine having a computer using PCI Express and whacking a gfx card in each slot - renderfarm-in-a-box!"

"The future is indeed exciting."

Well that WOULD be awsome, but we all know that no motherboard maker will have the guts to produce a board with more than one 16x(or is it 32x?) I would be surprised if we ever see more than one or two 8x slots. (has any board ever had more than 1 agp slot to have 2 real graphics cards running a dual monitor set up. or 3 for that 120 degree look in fps games)

But I sure do hope I am wrong and you are right!

Re:3D graphics cards aren't relevant

[Anonymous Coward]

geforce3 was the last card with identical silicon as a quadro. The 5800fx (the dustbuster) is also supposed to be identical as a quadro.

I'm running a softquadro'd geforcefx 5900. It benchmarks 20% faster in Maya (spec and homegrown benchmarks), and much faster in spec viewperf.

I don't have a quadro fx to compare against, but supposedly I am missing hardware acceleration for antialaised lines, aa line gama-correction, clip planes, and overlay planes.

Re:3D graphics cards aren't relevant

[Paladin128]

I don't know... remember, NVIDIA makes AMD chipsets. I could see NVIDIA making a successor to the nForce3 line that is made for a render-cluster node: Dual (or quad, or even 8-way) Opteron, 8 PCI Express X16 slots... If this tech can do what they are saying, it could be incredibly profitable. How many outfits would pay top dollar to have their rendering time cut by a factor of 10 or greater?

This is just the next logical progression.

Re:3D graphics cards aren't relevant

[Anonymous Coward]

I think everyone here I a little confused. Realtime is just a speed. Raytracing can be done in realtime BUT you usually only see clusters achive this kind of speed. I have personly seen realtime raytracing.

Now onto more of opinion rather than fact:
I hope to see consumer cards trade Raster methods with raytracing as power of these cards climbs. It can be done. Nobody has made the plunge yet though. Next gen console, PS3, has been rumored to be looking into raytracing tech. Graphics are at the point where we keep adding hack above hack to get this next special effect working with standard rasterization. With raytracing this all comes naturally. Really. No joke. That is why some of the first the systems where ray tracers but games needed a (hopefully temporary) shortcut in speed.

This is just the first step to realtime consumer use. It just depends on where things like games go. And for that we usually look to the next gen. consoles. I dont think it will dominate totally with this generation but I think we will see new things possible.... anyway I would like to see if ATI can offer some ray tracing tech.

Re:3D graphics cards aren't relevant

[Anonymous Coward]

The reason you don't see more then one AGP slot is that AGP is not a bus per say (AGP == Accelerated Graphics Port), it is a point to point protocall. Adding in a second AGP slot would require a redesigned north bridge with two ports. This would be costly and difficult to produce. I'd guess that you're going to see multiple 16x PCI Express slots on atleast some of the high-end server boards. (There was no reason for a server to have multiple AGP slots since you couldn't stick a RAID controler in a AGP slot)

Re:3D graphics cards aren't relevant

[Anonymous Coward]

Heard that 8 hour figure to in reference to Finding Nemo said by the mighty John L of Pixar. I think it was on TV (Entertainment Tonight?) or part of the behind the scenes special on a Pixar DVD. I think they also mentioned that the bathrooms at Pixar are centrally located to better aid colaboration between the high and low employees's. Great stuff from a great company.

Re:3D graphics cards aren't relevant

[shplorb]

As another person has already said, you can only have one AGP slot. PCI Express is the next-generation, high-speed replacement for PCI. Remember how the first couple of generations of 3D accelerators were PCI-based?

Besides, you don't need killer bus bandwidth with this because you're not trying to pump out 100fps using a couple of hundred megs of geometry and textures on a card with only a hundred or so megs of memory. (That means you have to send loads of data over the bus 100 times each second.)

The power here is in the parallelisation and incredible performance delivered by highly-specialised processors. Graphics cards have phenomenal memory bandwidth - nVidia's latest has something like 32GB/sec (big B!) - compare that to say a Dual 2GHz PowerMac G5, which has 6.4GB/sec of memory bandwidth. New graphics chips are heading towards the usage of memory paging (3D Labs P10 already does this I believe) So with this and high-end cards with 256 or 512MB of RAM you won't need much bus bandwidth because you'll just page in little bits of geometry and textures as each processor needs it, rather than having to upload huge textures everytime an entire one gets trashed to make room for another one.

So once again, the key thing to remember is that you're not trying to push 100fps. Most of the time spent rendering a frame will be in the GPU shader units, not uploading data to the graphics chip.

Rendering artefacts between cards?

[Anonymous Coward]

The rumor on the street is that a Soho based SFX house tried this when they had a deadline that standard software rendering couldn't meet.

So they wrote an engine to do renderman->OpenGL and ran it across many boxes.

Problem was that they got random rendering artefacts by rendering on different cards - different colors etc, and couldn't figure out why.

When working on one box they got controlled results, but only had the power of one renderer.

Re:Rendering artefacts between cards?

[XMunkki]

Problem was that they got random rendering artefacts by rendering on different cards - different colors etc, and couldn't figure out why.

I have seen this problem in software renderers as well. The problem seemd to be that part of the rendering farm was running on different processors (some were Intel, some AMD and many different speeds and revs) and one of them supposedly had a little difficulty with high-precision floating points and it computed the images with a greenish tone. Took over a week to figure this one out.

Re:Rendering artefacts between cards?

[Doh!]

Gelato won't be doing real-time hardware rendering the way you would normally use OpenGL for. It seeks to use the processing power of the GPU for calculations that would otherwise be done by the CPU. The goal here isn't fast, real-time output, but rather high-quality film-res output that can still take hours to render, so theoretically it's designed to give consistent results across all cards that support it.

Re:Rendering artefacts between cards?

[DrSkwid]

I used to see that with 3ds4 as well when I was rendering this. One was a pentium and one was a pentium pro.

Ah those were the days. We were on a deadline and rendered it over Christmas. After four hours the disks would be full and it would be time to transfer it to the DPS-PVR. I spent six days where I couldn't leave the room for more than four hours, sleep included. Was pretty wild !

VH1 viewers voted it second best CGI video of all time, behind Gabriel's Sledgehammer so I guess it was worth it!

Re:Rendering artefacts between cards?

[Jeff DeMaagd]

My question is, if a GPU on an AGP card can send the render results back to the system, what is the point of PCI-Express? That was supposed to be one of the "enhancements" of PCI-E. PCI-X was said to have the same limitations as PCI and AGP.

Re:Rendering artefacts between cards?

[rendermaniac]

This was done by Hal Bertram who is now at MPC. He presented it as a Stupid RAT trick at last year's Siggraph. Hopefully his notes will get up at http://www.renderman.org/RMR/Examples/srt2003/inde x.html at some point. I think he basically dumped out the positions of all the grids that prman makes and then loaded this into his renderer. I am guessing he just used the processing power of the card to push out Tiff files although I guess this would involve some OpenGL at some point. Simon

Re:Rendering artefacts between cards?

[andr0meda]

We also experienced this, when simply the boxes would have slightly different localisation settings, and different motherboards (& CPU's). Apparently, our software rendering tool interpreted some data for certain plugins differently, and because of this, which sometimes resulted in layers of different sizes and resolution, alignment problems, color differences and so on.. the whole farm had to be taken down bit by bit to figure out why. In the end we ended up rendering on only a 3rd of the capacity because even ghosting the good systems did not help.

Re:Rendering artefacts between cards?

[cubic6]

An AGP GPU can send the results back, but the bandwidth from the GPU to the system is *really* slow compared to the bandwidth from the system to the GPU. It's asymmetric. That ends up being the bottleneck, because the GPU can compute frames much faster than it can send them back. PCI-E has equal bandwidth both ways, so it should theoretically be able to keep up.

Fab for machinima

[Paul Crowley]

For some possible applications, check out machinima.com - film-making in real time using game engines.

Re:Fab for machinima

[dnoyeb]

I think you make the best point on the board today.

The opening quote of the article poster is ignorant. Movie rendering has been done in hardware forever. He seems to be mixing up doing rendering in hardware with rendering on the fly in a video card.

What we have here is a slight mix of the two, but by no means anything new on the market. Its only letting you use your quadro if you already have one for movie rendering acceleration. I certainly would not buy one for this purpose. I imagine its still incredibly more profitable to use a CPU than GPU. Also, note that render farms computers rarely have video cards. The video part would be wasted.

But for in-game recording by home users and non-studio $$$ having artist, this will likely be a welcome addition. (Especially for those who can turn a GeForce into a Quadro) Though I have to wonder, doesen't the Quadro cost quite a bit of money!?

Still think its nice technology. I wonder if PCI-Express is allowing them to get this data off the GPU and back into the hands of the CPU?

Re:Fab for machinima

[Viking Coder]

RTFA.

You're the ignorant one. Movie rendering has been done on the CPU forever. This for the first time is doing final movie rendering on the GPU.

This is definitely something new on the market. Point me to another product that does final movie rendering with hardware acceleration provided by the GPU, and I'll eat my hat.

I imagine its still incredibly more profitable to use a CPU than GPU.

Why? Because it's faster? Bzzzt. That's the whole point. Take a look at the transistor counts for the latest and greatest GPU, and then look at the transistor counts for the latest and greatest CPU. GPUs are clearly outpacing CPUs, and for tasks such as rendering, will soon be able to run circles around software renderers (if they're not doing it already.)

Time is money. If it turns out you can render the same frame in 95% of the time using this technology, you'd be an idiot to not buy a quadro. A 5% speed-up over a period of months-to-years adds up.

But for in-game recording...

This has absolutely nothing to do with in-game rendering, let alone in-game recording. This is an offline (read: not real-time) renderer, suitable for such things as rendering final frames of a movie, not in-game rendering in real time.

I wonder if PCI-Express...

Video cards have been capable of sending final renders back to the CPU for a really long time. PCI-Express just makes it faster.

Re:Fab for machinima

[BFaucet]

No, that's not an application... Gelato is definitely NOT realtime.

The applications are in final rendering for films/video, NOT realtime rendering in game engines.

Re:Fab for machinima

[Eraser_]

Video cards have been capable of sending final renders back to the CPU for a really long time.

Just to prove that, see also the "screenshot" button in your game. It pulls the current frame straight out of the vram and writes it to the hard drive as a bitmapped image. As the author says though, PCI-Express should hopefully make it a more DMA like action, so as not to slow to a crawl when you pull data out.

Eat some gelato

[SpikyTux]

Gelato (Italian) == Ice cream

Re:Eat some gelato

[Vampyre_Dark]

And gelato is what you turn into into when you have too much. :)

Re:Eat some gelato

[piper-noiter]

With a cost of $3200+ (license plus annuity), this better be the best damn ice cream I've ever consumed. I expect chocolate, caffeine, and it had better be able to rinse out the bowl and stick itself in the dishwasher. That?s all I have to say.

Re:Eat some gelato

[neurojab]

Gelato is not exactly "Ice cream" in Italian. Gelato is the Italian version of ice cream.

If you go to italy and buy a gelato from a street vendor, you'll notice that it's not quite the same.. it's denser and more "gooey". Gelato is often served with two or more scoops of different flavors.

http://www.slowtrav.com/italy/food/gelato.htm

It's worth the trip to Italy to try it. Sure you can buy "gelato" here, but it's just not quite the same.

Quick question...

[Noryungi]

Is there any Free software capable of exploiting the general computing power of modern video cards?

Well, since they released "a C++ interface for plugins and integration" for Gelato (ice cream in Italian, btw), this probably means that free software can (and, eventually, will) support all these high-end functions... or am I completely wrong?

For instance, just imagine Blender with a Gelato plug-in for rendering... hmmmm... Now I understand why they named it "Gelato"...

Re:Quick question...

[sshtome]

Am I wrong in thinking that maya runs on Unices already??

Some folk just can't afford it?! (read everyone) but I think that it's good to have proprietry software on Linux. If you make 3D graphics for big money you can afford to pay for the dogs in 3D. And Linux the smart choice for your render farm...

yes you or I get Blender. You and I couldn't do 3D graphics for Harry Potter 4.

Re:Quick question...

[Anonymous Coward]

Maya was under Irix as far back as '96 --- this I know after doing no research as I ran Maya on an Indigo2 Extreme.

Don't you mean..

[Viceice]

Nvidia releses Hardware-Accelerated video renderer?

Re:Don't you mean..

[DrSkwid]

or even releases

btw. it's punishment

Re:Don't you mean..

[cubic6]

Just to be pedantic...

You don't render video, you render frames and assemble them into video, then display the video. The title is indeed correct.

Re:This would be more useful

[master0ne]

dont forget blender for windows aswell

the problem is in the Bus

[rexguo]

The AGP bus has assymetrical bandwidth. Upstream to video card is like 10x faster than downstream to the CPU. So you can dump tons of data to the GPU but you can't get the data back for further processing fast enough, which defeats the purpose.

Re:the problem is in the Bus

[BiggerIsBetter]

Not if the purpose is to output to a recording or viewing device, but they're probably planning to use it with PCI-X anyway.

Re:the problem is in the Bus

[snakeshands]

The purpose might mostly be to show people why they need to run out
and get PCI Express hardware; it completely addresses the assymetry
issue.

I'm guessing the main reason Gelato is spec'd to work on their
current AGP kit is to encourage the appearance of really impressive
benchmarks showing how much better performance is with PCI Express.

They have a good idea, and they're rolling it out at a good time,
I think.

Some folks were trying to do stuff like this with PS2 game consoles,
but I guess now they'll have more accessible toys to play with.

Re:the problem is in the Bus

[WARM3CH]

I believe the problem is not with AGP bus, but rather with the GPUs that are NOT designed in the first place to transfer anything back to the memory. In normal 3D applications, you just feed the graphics card with all sorts of data, like the texturs, geometry, shaders... and the result goes out throug the VGA connector! You don't need to give it back to the CPU or the memory. The GPU and the memory architecture of the graphics card is simply designed to just recieve the data with highest speed from the CPU. However, this does not mean you CAN'T get a high bi-directional throughput in AGP if you're going to design a special GPU for offline renderings.

Re:the problem is in the Bus

[Anonymous Coward]

I know a researcher who is working on this problem.

What he does is to take the video output, in DVI, and to feed it back to some 'homemade' PCI card, with 2 FPGA on it. Then he can benefit of the full bandwidth.

But it is quite pricey...

Re:the problem is in the Bus

[Trogre]

If it's I/O bound, yes.

However if the GPU can be left to crunch for most of the time and return say a row of pixels at a time I doubt the 1/10th speed of the AGP bus downstream would be a big problem. For complex scenes the input (textures, geometry, shaders) may well exceed the output (pixels) in terms of data.

Re:the problem is in the Bus

[Kris_J]

Hello? Digital out? You can plug it into something other than an LCD monitor.

Re:the problem is in the Bus

[Divlje Jagode]

that's for video processing (impressive nonetheless), what we're talking about is bus data transfer. The problem seems to be that you can push arrays (texture data) to the card's memory but cannot get the processed data back nowhere fast enough...

A couple of years back there was an article on /. about this. Some guy (team?) had designed a benchmark to show the assymetrical throughput. There was a program you could download to check it out for yourself (windows)

Re:the problem is in the Bus

[mike260]

So you can dump tons of data to the GPU but you can't get the data back for further processing fast enough, which defeats the purpose.

Fast enough for what exactly? What purpose does that defeat?
Compared to the cost of (eg) sending a frame across a LAN, the time taken to pull a frame across the bus would be utterly, utterly insignificant.

Re:the problem is in the Bus

[marcelo.mosca]

You are missing the point that nVidia will be using a bridge on its first PCI Express setup. The chips will basicly talk to the bridge in AGP16x and will suffer from the same asymetry problems today agp cards suffer.
Only the second generation PCI Express cards from nVidia will be native solutions and will use the bridge the other way arround (to usa a PCI Express chip inside an AGP system).

Re:the problem is in the Bus

[Paladin128]

Hello? We're not talking about real-time rendering here! We're talking about final renderings for movies at way higher res than most TMDS transmitters can handle, and WAY higher color depth than is supported by DVI (128-bit floating-point)

Re:the problem is in the Bus

[Performer Guy]

Wrong, you can do a lot of work on the card and store the results in memory, the final readback of the production ready image is a relatively small requirement. Besided the bus spec is not asymmetric just the typical implementations. Sure most people use PCI reads but you should be able to DMA read with plenty of bandwidth.

Re:the problem is in the Bus

[Kris_J]

Okay, so which is it? Is the bus too slow or is the DVI too fast?

Re:the problem is in the Bus

[Paladin128]

Neither. DVI would be impossible, and we're not talking about real-time stuff. The AGP/PCI Express X16 bus would be fine. The GPU would be treated like a parallel vector math unit, and be able to process large groups of pixels, and store them in VRAM, which would be used more like conventional RAM than a frame buffer. When the frame (or part of it) is done, it's sent back to main memory or hard disc. AGP can actually do this easilly through sideband addressing (accessing main memory without arbitration of the CPU. Not sure about PCI Express, but it offers a hell of a lot more bandwidth.

DVI can handle 8/16/32-bit color at upwards of 1600x1200. For film, you'd want way higher resolution (at least 1922x1080) stored at at least 64bit floating-point color depth for post-processing.

Teh horror !!!

[Anonymous Coward]

"Gelsto is proprietary (and pricey)"

A company that wants to be payed for their work, weird !

You will see more, allot more, of this for the Linux platform in the near future.

Software may be released with source code, but no way that it will be released under GPL, most ISV's can't make a living releasing their work under GPL.

And please the "but you can provid consulting services" argument is not valid, it dont work that way in the real world.

Re:Teh horror !!!

[Anonymous Coward]

>And please the "but you can provide consulting
>services" argument is not valid, it dont work that
>way in the real world.

Indeed. If you are entering the product support consulting business the extra expense to create your own product to support only pays off if the product is a whiz-bang success.

They could support themselves by banner ads on their home page. That is a tried and true business model.

Re:Teh horror !!!

[Cobron]

Still I think it's a good idea to mention if software for Linux is proprietary.
That just saved me the trouble of clicking the link and checking it out ;) .

ps: I recently visited a project trying to "harness the power of GPU's". I think that project was something like seti/folding/ud/... but tried to have all the calculations be made by the GPU of your 3d card.
If someone knows what I'm talking about: please post it, I can't find it anymore ;)

Re:Teh horror !!!

[Blackknight]

If I had mod points I'd mod you up. I'm getting tired of the slashdot mentality that everything has to be free. If a company makes a professional grade product, and there is demand for that product, they should be able to be rewarded for their efforts.

Re:Teh horror !!!

[Anonymous Coward]

maybe they can sell the hardware?

Re:Teh horror !!!

[Hast]

Probably something you can find on the General Purpose GPU site.

I've toyed with shaders some and implemented a system for image processing on GPUs. Quite a lot of fun really, though we didn't do any comparisons with CPU to see how much faster it was. (That project isn't published anywhere though.)

Re:Teh horror !!!

[Anonymous Coward]

They could support themselves by banner ads on their home page. That is a tried and true business model.

Don't give them any ideas! If my favorite sites like Slashdot or Fark started to put advertising up I'd be pretty annoyed. Besides, with adblock for Mozilla I don't see ads on annoying sites anyway. How can anyone possibly make money off of banner ads anymore?

Re:Teh horror !!!

[orasio]

The GNU/Linux platform, thank you very much, I don't think you could build much on top of just a kernel.

By the way, GPL has nothing to do with price, you can charge it and still keep it GPL'ed.

Re:Teh horror !!!

[blackdragon7777]

But then people can distribute it for free (ie piracy) and it would be fine under the GPL. The GPL does not work for a business that wants to develop software (not support). I want to make money writing and designing software not consulting.

Re:Teh horror !!!

[danila]

I wonder how soon will companies realise (they will eventually) that they can easily combine the benefits of open source with proprietary profitability. If nVidia realises Gelato under a GPL-like license with one addition - to use the code you need to pay $$$ to nVidia. Everything else would be the same as in GPL, all changes must be shared and made available (to those in the club). This would allow nVidia to create a small open-source community among CGI programmers, while ensuring they continuously get paid to pay for their own R&D.

BTW, the same approach can be used in game programming. Games would benefit enormously from open source collaboration and if the licence included a payment scheme, the model might be feasible.

Re:Teh horror !!!

[Karellen]

No, but NVIDIA is not an ISV. It's a hardware company ferchrissakes!

It doesn't matter if I have a copy of their great software to render movies with if I don't have one of these whiz-bang cards to render it on. How on earth am I (or L33t J. Hax0r) going to copy a piece of hardware? The premise you quote is shaky at best (RedHat, among others, makes a profit from selling support & consulting services) but completely pointless when directed at a hardware company!

Look at IBM. They're making loads of money with Linux, not because of Linux sales or support or consulting, but because they're using it as a way to help sell their hardware.

Re:Teh horror !!!

[orasio]

Piracy is attacking ships, killing, raping, robbing.
Distributing things for free is usually helping a friend, or at least helping people.
By the way, if you want to make money writing and designing software, and selling licenses, good luck. I think that is a dying trend for anyone smaller than MS.

I make a living designing and writing software where I deliver sources, and don't claim exclusive rights, much like the BSD license, because if the customers pay for the software, I think they should have it anyway they want it.

Re:This would be more useful

[BiggerIsBetter]

Alias has Maya for Linux. Newtek has Lightwave rendering node software for Linux. There are a few other 3D packages like AC3D too.

BURN!!!!!!

[Viceice]

I wonder what market segment nVidia is gunning for. Are they after some of Discreets market share or trying to offer a hardware solution that will beat the crap out of Adobe After Effects.

It would be really cool to have a hardware solution with Combustion like features for the price of After Effects.

Re:BURN!!!!!!

[TheOldFart]

Why not both? After Effects will do what Burn does but very slowly. Burn will do what After Effects does really fast but very, very expensive. Combine these new hardware accelerators with some cleaver code and you get all that for a fraction of the cost.

Re:BURN!!!!!!

[Anonymous Coward]

No, they want that film manufacturers will use Quadro GPUs instead of CPUs in their render farms.

Re:BURN!!!!!!

[Hast]

Play around some with pixel/vertex shaders, they are quite easy to get the hang of and plenty powerful. (Even if you don't have the latest and greatest gfx cards.)

Could make a nice addition to GIMP (if there isn't one already).

'pricey'

[neurosis101]

Um... depends on what you're looking for/expect. This isn't intended for you to buy and use at home. This is more likely for smaller developers (big developers write their own usually... think Pixar). Professional grade equipment is all expensive. The first common digital nonlinear editor was the casablanca, and with an 8 gig scsi drive ran close to a grand when it was released. This was just a single unit.

I bet the type of people that buy this are like big time architects that have a few machines set up to do renders for clients, and want to perhaps do some additional effects for promo/confidence value, that likely already have people running that type of hardware.

Then again all those Quadro users could be CAD people and they've got no audience. =)

Re:'pricey'

[Dynedain]

I work for a smaller graphics house that is part of an architecture firm and does mostly (but not all) architecture work.

This product is competing against other rendering engines like MentalRay, Vray, Renderman, etc. And at $2750 per license it's deffinately not for smaller developers or architects. There are plenty of other rendering engines out there that are significantly cheaper and dont require a video card that costs as much as an entire x86 rendernode.

Re:'pricey'

[RupW]

Professional grade equipment is all expensive.

No, you can get raytracing hardware for less than the software and a Quadro FX would cost you.

For example, there's the ART PURE P1800 card which is a purpose-built raytracing accelerator. It's a mature product with an excellent featureset, speaks renderman and has good integration into all the usual 3d packages. It's generally acknowledged as a very fast piece of kit with excellent image quality, and plenty of quality/speed trade-off options. And if you've a deeper wallet they do much bigger network-appliance versions.

Re:'pricey'

[Tim C]

Then again all those Quadro users could be CAD people and they've got no audience. =)

Not just CAD - I do server-side Java programming, and we've all recently been bought new PCs. The spec we went for included a Quadro FX 500; don't ask me why, it just did... (it was that, or a similar machine with a GeForce - I didn't make the choice)

Re:'pricey'

[urbaneassault]

oh god, please don't remind me of the cassie...in high school we built our first studio around 6 of them. I don't know, between their uncanny ability to constantly crash, or the excessive render times, they were definitely worth the price :). made me feel less dirty when we dumped them all for 3 edit* boxes. ahh, i miss the old discreet*

Re:'pricey'

[malducin]

This is more likely for smaller developers

Well who knows. The press release does mention ILM, certainly not a small fish.

big developers write their own usually... think Pixar

Yes though that's a special case since they sell it as well as is fairly open (API wise). What you probably mean is those with propietary solutions, in particular Blue Sky Studios, PDI and Rhythm and Hues. But even the big houses who use PRMan and Mental Ray, write customized renderers for specific problems, even if they use a commercial solution like PRMan and MR, like Digital Domain's VoxelBitch, Weta Digital's had a special renderer for Massive (though Massive spits out RIBs), ILM had (has?) propietary hair and particle renderers, etc.

The first common digital nonlinear editor was the casablanca, and with an 8 gig scsi drive ran close to a grand when it was released.

Though that was nothing compared to EditDroid that was the first film nonlinear editor, in which film was scanned and transferred to laser discs. And it certainly cost more than a grand ;-).

The press release does mention Frantic Films as well which is a smaller house. I guess they could use it directly for final renders. No idea how it would fit on a pipeline like ILM's though.

I guess this is what BMRT has turned into...

[Anonymous Coward]

It looks like this is what NVIDIA have done with BMRT when they bought it: look at what has become of exluna.com

If anyone's wondering, a couple of the latest releases of BMRT (Blue Moon Rendering Tools) before NVIDIA pulled the plug on them are available here

General computing on graphics hardware

[attaka]

I have been reading interesting stuff about this lately. Take a look at this Stanford project: BrookGPU

This might also be interesting: GPGPU /Arvid

Re:General computing on graphics hardware

[Anonymous Coward]

GPGPU...

Hmm...

So could you stack your computer up with multiple PCI video cards to effectively make a multiprocessor PC?

NOTE: I did *NOT* say a Beowulf cluster of video cards.

Re:General computing on graphics hardware

[BiggerIsBetter]

Yes. I've done it with the BrookGPU library and two Nvidia cards (1 AGP, 1 PCI). The benchmark wasn't that much faster though, presumably because the datasets weren't big enough to just leave the cards running -I suspect the main CPU was too busy feeding them to get any real work done.

Linux software

[HenchmenResources]

Is there any Free software capable of exploiting the general computing power of modern video cards?

Take a look at the Jashaka project. It is a real time video editing suit and the designers have been working with and have supposedly been getting support from Nvidia, so they may have had access and I would imagine certainly will have access to these video cards. I can't imagine them not taking advantage of this technology.

The other nice thing is if memory serves me correctly this program is being designed to work on Windows, Linux and OS X, so good news all around.

Using GPU for signal processing

[PastaAnta]

is there any Free software capable of exploiting the general computing power of modern video cards?

A quick Googling revealed the following:
- BrookGPU
- GPGPU

Re:Using GPU for signal processing

[Kris_J]

Is there a D.Net client for my GeForce4 yet?

New Headline:

[Anonymous Coward]

Windows XP's Achilles Heel Apparently Revealed

ExLuna, take 2 :)

[Anonymous Coward]

Seems like ex-Exluna staff (bought by NVidia) is going to kick PRMan's a$$ on hardware level: they tried it on software level with Entropy, but got sued into oblivion by Pixar, now it's time for revenge?

M4 open GL VJtool.

[kop]

M4 is a free as in beer movieplayer/vj tool that uses the power of openGL to manipulate movies,images and text.

I think it'll start happening a lot more

[PlatinumInitiate]

Not only with hardware manufacturers/drivers, but also general software. ISV's are getting annoyed by Microsoft's dominance of the desktop market, and through that, their (heavy) influence on desktop software. It's not inconceivable that in a decade, Microsoft could control every aspect of the standard desktop PC and desktop software market. At the moment some of the only really strong ISVs in their respective areas are Adobe, Corel, Intuit, Macromedia, Oracle, and a few specialized companies. Expect a big ISV push towards a "neutral" platform, like Linux or FreeBSD. Windows is too big to stop supporting, but ISVs will be smart to at least try and carve out a suitable alternative and avoid being completely dominated by Microsoft. All that most ISVs might be able to hope for in a decade is being bought out by Microsoft or making deals with Microsoft, if things don't go the way of creating a vendor-neutral platform.

Re:I think it'll start happening a lot more

[pVoid]

Who are you?

How do you know ISV's are getting annoyed? Do you go to lunch with ISV's every other day?

Not only are you crudely generalizing, I think your point is actually not sound at all. You think Adobe cares about Microsoft dominating?

The much more plausible explanation is that they (nVidia) already had the drivers/software for the architecture on linux (read previous posts, they bought the card).

Another yet plausible explanation is that drivers are more difficult to implement in Windows (because of kernel mode constraints).

And before people start flaming, I'll tell you right away:

I've written drivers for both. A seg fault in a linux kernel driver will generate a message saying "oops, you segfaulted your kernel", whereas a seg fault, or even an invalid heap memory tag in Windows will instantly halt the system with a blue screen. Don't even get me started about IRQL's in Windows - they are essential for proper SMP multithreading, but really are very difficult to work with.

Re:I think it'll start happening a lot more

[PlatinumInitiate]

How do you know ISV's are getting annoyed? Do you go to lunch with ISV's every other day?

No, but working for a medium-sized ISV who deals with Microsoft (we buy bulk embedded XP licenses for use in custom gaming machines), I can tell you a few things about how Microsoft deals with customers. They have actually tried to offer us better deals if we discontinued our Linux solutions and marginalized our dealings with our Russian partners who produce hardware and software for use with Mandrake Linux 9.x in gaming solutions. (Sounds impossible? Think again). I can only imagine how much more underhanded Microsoft are when dealing with bigger ISVs.

Not only are you crudely generalizing, I think your point is actually not sound at all. You think Adobe cares about Microsoft dominating?

I'm sure Netscape and Sun didn't care either, until Microsoft took them out of the market. You are really insulting the intelligence of the Adobe executives if you think that they haven't considered this possibility or what they could do to avoid something similar happening.

Re:I think it'll start happening a lot more

[pVoid]

I'm sure Netscape and Sun didn't care either, until Microsoft took them out of the market.

I know this is an age long dispute, but no, I do not think I'm insulting Adobe's execs. Why? because they are intelligent enough that they know they have to maintain a good level of product.

That's why they bought Cool Edit, and turned it into Audition. It's the same shell as before, just the splash screen changed.

My point, the age long point of discussion being that Nutscrape LOST its market, it wasn't run over by microsoft.

Microsoft can not come up with something to topple Photoshop even if it spent 4 years in production. By that token, how come dreamweaver is still around? By your argument, front page should have wiped them off the face of the earth by now.

Re:I think it'll start happening a lot more

[Crazy Eight]

I would think that the first plausible explanation is simply that this product would appeal most to those already utilizing Linux in renderfarms.

Re:This would be more useful

[EnglishTim]

Insightful my ass.

Maya, Houdini and XSI are all available for Linux, and they work well.

Free software is a product of a lifecycle

[heironymouscoward]

Software, like much technology, follows a classic cycle from rare/expensive to common/cheap as the knowledge and means required to build it get cheaper.

"Moore's Law" is simply the application of this general law to hardware. But it applies also to software.

Free software is an expression of this cycle: at the point where the individual price paid by a group of developers to collaborate on a project falls below some amount (which is some function of a commercial license cost), they will naturally tend to produce a free version.

This is my theory, anyhow.

We can use this theory to predict where and how free software will be developed: there must be a market (i.e. enough developers who need it to also make it) and the technology required to build it must be itself very cheap (what I'd call 'zero-price').

History is full of examples of this: every large scale free software domain is backed by technologies and tools that themselves have fallen into the zero-price domain.

Thus we can ask: what technology is needed to build products like Gelato, and how close is this coming to the zero-price domain?

Incidentally, a corollary of this theory is that _all_ software domains will eventually fall into the zero-price domain.

And a second corollary is that this process can be mapped and predicted to some extent.

Re:Free software is a product of a lifecycle

[tehcyder]

Perhaps I'm not following, but doesn't this mean that Windows should be free by now?

Re:Free software is a product of a lifecycle

[heironymouscoward]

Windows a specific product, not a technology. The technology would be the operating system. An OS is based on a series of other technologies, most obviously compilers, networking, disk management systems, kernel models, etc.

Since these underlying technologies have been zero-priced since the 1980's (mainly thanks to Unix), the OS as a technology has indeed fallen into the zero-price domain as well.

In other words: a small team can today build a product that competes fairly well with Windows, using off-the-shelf technology. This was not true 10 years ago.

Windows is not free, but it obviously has found itself right in the middle of the zero-price zone.

Re:Free software is a product of a lifecycle

[Jace of Fuse!]

In other words: a small team can today build a product that competes fairly well with Windows, using off-the-shelf technology. This was not true 10 years ago.

In fact, thanks to Zero-Price(as you put it) software (and underlying technologies), Windows Binary Compatibility is becoming more of a reality with each day.

I suspect it won't be long before some of the "free Windows Compatible Alternatives" start becoming real threats to Windows market share. That might still be a couple of years off, but it is eventually going to happen.

Re:Free software is a product of a lifecycle

[haystor]

(-1 RMS)

These things haven't been made free "mostly thanks to Unix" but to people who weren't fond of paying for Unix or being locked into something they couldn't change.

Re:Free software is a product of a lifecycle

[Anonymous Coward]

Ironically Unix made it possible to escape Unix. Think of lex/yacc as examples of Unix technology that allowed people to escape the proprietary Unix world.

I think his argument is about technology, not specific products.

You might argue that "Unix" is a product not a technology but both views are valid.

Re:Free software is a product of a lifecycle

[dookiesan]

Incidentally, a corollary of this theory is that _all_ software domains will eventually fall into the zero-price domain.

Mathematics at its greatest.

Nice to see some good out of BMRT/Exluna.

[ron_ivi]

Renderman.org's summary of Exluna & BMRT describes where much of this technology probably came from.

For those who don't remember, BMRT was a really cool RenderMan based renderer that Pixar had some sort of love/hate relationship with. IIRC, they used it, yet they sued the company. At the end nVidia bought them, though it wasn't clear why at the time.

Re:Nice to see some good out of BMRT/Exluna.

[ron_ivi]

And a nice Siggraph presentation of some of the capabilities of BMRT.

Interestingly, BMRT was free as in $$$ but not as in Free Software. This was one of the first software packages where I first recognized how big this distinction is. (A free as in Free Software program probably would have continued on as people may have coded around some of the disputed intellectual property - a free as in $$$ program was possible to kill with the carrot and stick of a lawsuit and buyout opportunity)

Re:Nice to see some good out of BMRT/Exluna.

[Anonymous Coward]

Parent wrote: "where much of this technology probably came from"

Indeed. NVidia's FAQ for this group says " It is the evolution of NVIDIA's acquisition of Exluna in 2002"

Re:This would be more useful

[SlightOverdose]

Don't forget POVRay

GPU as 2nd processor (slightly offtopic)

[CdBee]

is there any Free software capable of exploiting the general computing power of modern video cards?

I expect that once it suddenly becomes clear that the GPU in a modern video card has serious processing power, that someone will release a version of the SETI@Home client which can use the rendering engine as a processor. Bearing in mind that most computers use their GPU's for a very small percentage of their logged-in life, I suspect there is real potential for using it for analysing on distributed computing projects.

Absolutely

[TheFr00n]

Check out www.jahshaka.com. It's an open source video compositing / FX package that leverages the 3D accelerator chip on your graphics card to do incredible things. This is one to watch, it's definitely going places.

You can download binaries for linux and windows (and MAC), and source tarballs are available for the savvy.

I know, it's not strictly a "renderer", but it employs many of the fuctions of a renderer to create realtime effects and transitions.

Not an issue, esp. for non-RT rendering

[Namarrgon]

You're right that the AGP port is asymmetric, but this is unlikely to be a bottleneck if they can do enough of the processing on the card.

For 3D rendering, especially non-realtime cinematic rendering, you have large source datasets - LOTS of geometry, huge textures, complex shaders - but a relatively small result. You also generally take long enough to render (seconds or even minutes, rather than fractions of a second) that the readback speed is not so much an issue.

Upload to the card is plenty fast enough (theoretical 2 GB/s, but achieved bandwidth is usually a lot less) to feed it the source data, if you're doing something intensive like global illumination (which will take a lot more time to render than the upload time). Readback speed (around 150 MB/s) is indeed a lot slower, but when your result is only e.g. 2048x1536x64 (FP16 OpenEXR format, 24 MB per image), you can typically read that back in 1/6 of a second. Not to say PCIe won't help, of course, in both cases.

Readback is more of an issue if you can't do a required processing stage on the GPU, and you have to retrieve the partially-complete image from the GPU, work on it, then send it back for more GPU processing etc, but with fairly generalised 32 bit float processing, you can usually get away with just using a different algorithm, even if it's less efficient, and keep it on the card.

Another issue might be running out of onboard RAM, but in most cases you can just dump source data instead & upload it again later.

Little value...

[winchester]

Almost every FX house worth its salt in the CG business uses Pixar's Renderman on UNIX or Linux machines. The reasons behind this choise are very simple.

Renderman is proven technology and has been so since the early '90s. Renderman is well known, its results are predictable and it is a fast renderer. Also, current production pipelines are optimised for Renderman.
UNIX and Linux are quite good when it comes to distributed environments (can anyone say Render Farm?) and handle large file sizes well (Think a 2k by 2k image file, large RIB files).
And last but not least, renderman is available with a source code license.

Hardware accelerated film rendering is in essence nothing but processor operations, some memory to hold objects and some I/O stuff to get the source files and output the film images. Please explain to me why a dedicated rendering device from NVidia would be any better than your average UNIX or Linux machine? Correct, there aren't any advantages, only disadvantages. (More expensive, proprietary hardware, unproven etc.)

Re:Little value...

[Rotting]

That is odd.

I am no expert in this matter but a friend of mine has worked in the graphics industry (mostly with movies and TV shows) for several years now and from what I have learned from him a lot of modelling is done using 3DS Max and as well a lot of work is done in Maya.

I could be wrong though :)

Re:Little value...

[evenmoreconfused]

Please explain to me why a dedicated rendering device from NVidia would be any better than your average UNIX or Linux machine?

Uhm... because one $300 ATI or nVidea GPU has many 10's of Glops available (someone above said >50 in the shader unit alone) and your average $500 Linux box has maybe 5 Gflops?

Re:Little value...

[Hast]

As someone pointed out before a GPU is pretty close to a processor or co-processor. Basically it's a really big, really fast vector processor. Your argument is not too far from claiming that you shouldn't use SSE2 for Intel chips or Altivec for G5 processors. Compared to those a new GPU has a shitload of performance though.

Re:Little value...

[mike260]

Please explain to me why a dedicated rendering device from NVidia would be any better than your average UNIX or Linux machine?

Why do you think 3D hardware exists at all, when all it's doing is a load of integer maths? Surely a Linux machine is capable of adding numbers together, right? Obviously, dedicated hardware is faster, sh*tloads faster. Durr. The benefits to the artist's equivalent of the compile-edit-debug cycle are fairly obvious here, and worth rejigging the production pipeline to accomodate.

Anyway, you can bet nVidia has a Gelato-based Renderman implementation in the pipeline. nVidia's long-term goals were always 1. OpenGL in hardware (done) and then 2. Renderman in hardware.

Re:Little value...

[Zak3056]

Correct, there aren't any advantages, only disadvantages. (More expensive, proprietary hardware, unproven etc.)

And, apparently, orders of magnitude faster.

Personally, I'd put that rather firmly into the advantage column, and for a number of reasons. You could either render your movie with a smaller farm (always a plus) or you could render even more complex scenes in the same time period--which is probably what most people would use this technology for. On the commentary track of Monsters Inc, the guys from Pixar note that despite having MUCH faster hardware (and alot more of it) the average time to render a single frame of Monsters Inc was just as long as a single frame of Toy Story. Why? Because the frames were FAR more complex.

I think this is a Good Thing(tm) at least for the people who have the imagination to use it.

Re:Little value...

[Serapth]

1) Renderman is anything but fast.

2) Renderman is by no means alone in the FX world. MentalRay and Brazil most immediatly come to mind. 3) As to UNIX and linux being quite good in a distributed environment... well duh... so are Windows and Mac boxes. As to file sizes... with the limit of 4gb on 32bit boxes, thats a pretty damned big file, and thats active in memory size limit. As to with files, the limit is more in terrabytes. As to Renderfarms, almost ever renderer under the sun handles this the same way... by having a render server that dispatches individuals frames across a number of machines. Even 3D Studios max default renderer supports this.

4) Who cares if Renderman is available with a sourcecode license... if you look at the Max or MAYA sdks... you would never need to go to the code level in the first place. Ditto for Mental Ray. The majority of dev tools in regards to a stand alone renderer boil down to import/export tools, and shader programming, and thats about it. Also, if your a major production house, Discreet, Alias or Softimage will all work to modify their tools to accomidate you. The new wave effects built into Maya are a result of the Perfect Storm, as an example.

As to why a dedicated rendering device from NVida would be better then your average UNIX or Linux box... I think your missing the point here. This isnt a dedicated rendering device... its software that works hand in hand with the existing Quadro already in your Linux box. As to the advantages... its a massive parallel processor boost to your box, with a chip that is a hell of alot better at managing floating point math, then your main processor is.

What I wonder is, if you have >1 Qaudro, do you get a boost perchip?

Re:Little value...

[Viking Coder]

Actually, with the transistor counts on the GPU these days, I like to think of the CPU as the co-processor. (I'm actually half-ways serious. My company does far more raw computation on the GPU than on the CPU.)

Re:Little value...

[CapnRob]

I doubt you would get a boost per chip on a single frame. However, the people who are gonna be interested in this aren't exactly going to be only rendering single frames, and therefore, if you have more than one Quadro, you'll be able to render more than one frame simultaneously.

Re:Little value...

[Serapth]

Actually to be honest, I doubt that is the case. I imagine the processing power added by the GPU will be used to render each single frame faster. Speeding up rendering of a movie by distributing frame rendering across multiple machines is relatively easy... Just add 1/box per frame needed and your good to go. However, speeding up the time it takes to do a single frame on a single machine is much more difficult.

Think about it this way... if hyper threading today doesnt allow two frames to be rendered at a time on a pc ( and it doesnt )... how is the GPU going to change that? However, if you speed up the rendering speed it takes to do a single frame, it will of course process more frames.

Solution: Use one model of card, one driver rev

[Namarrgon]

It's certainly possible that different hardware or even different drivers on the different machines doing the rendering can create subtle (or not-so-subtle) differences in each resulting frame, but standardising the hardware and drivers across machines should solve that completely.

Re:Solution: Use one model of card, one driver rev

[killmenow]

It would also make it much easier to "plug-n-play" a new box when one malfunctions. Imagine a 42U rack filled with identical 1U boxes. Still not a cheap solution, though.

Re:Solution: Use one model of card, one driver rev

[arkanes]

Also, it's common at animation shops to coopt desktop CPUs when people aren't working - you'd either have to give this up or standardize the hardware/software in your renderfarm with the hardware/software the desktops use (which would be silly).

Re:Solution: Use one model of card, one driver rev

[HD Webdev]

It's certainly possible that different hardware or even different drivers on the different machines doing the rendering can create subtle (or not-so-subtle) differences in each resulting frame, but standardising the hardware and drivers across machines should solve that completely.

Yes, but it's not a good idea to tie one's business to a particular hardware company let alone one product, and one driver.

Also, there are many people who make the money decisions that will balk at making particular changes to their their software or hardware without being forced to. Especially if they've been bitten before.

After all, as an example, many people still have to deal with the slick way in which WinModems/NIC's (software driven) pretty much secretly tied their machines down to using Microsoft Windows.

When I try to convert people in businesses to using Linux for stability, one sticky issue is when I bring up the fact that they will have to use particular modems for their boxes. They immediately think that I am trying to lock them into Something Bad when in reality, the Evil Deed has already been done.

I have to defend my position against "What happens when it breaks, do I have to buy from you? Who will help me if you aren't in business?", "Isn't Linux as capable as Windows?" as well as the silent "hmm, is this just the tip of the iceberg?" and "Will I be able to defend my job if this screws us in the future?"

I don't blame them. I feel the same way when I'm told anything that resembles "You must use brand X".

Re:Solution: Use one model of card, one driver rev

[Namarrgon]

A good point, but since most shops have multiple jobs in the render pipeline at a time, you could render one job on the GPU hardware, and another on the CPUs the old fasioned way.

Or, if the rendering is really just pure IEEE float calculations (and doesn't depend on e.g. differences in vendor sub-pixel accuracy), then it might be sufficiently similar and standard to be intermixed with CPU results after all.

Re:Solution: Use one model of card, one driver rev

[ErikZ]

Even if you don't like playing with Linux winmodem drivers, won't ANY external modem do?

Re:Solution: Use one model of card, one driver rev

[HD Webdev]

Even if you don't like playing with Linux winmodem drivers, won't ANY external modem do?

As far as I have seen, yes.

(I use external modems regardless of whether I am using Linux or Windows)

Re:This would be more useful

[afd8856]

I hope you know that all high-end 3d packages are available on Linux: XSI, Maya, Houdini, Real3d. And then you have some cool open source, like wings3d, that can cover some a lot of ground on the modeling field. Combine that with blender & yafray, and (theoreticly) you have a complete open source animation studio!

Free software ready indeed!

[Goeland86]

I think that indeed there is free software to do movies and rendered animations using raytracing. First, Cinelerra can use a linux cluster for movie rendering. Second, there's a whole bunch of modellers/raytracers out there that perform very well: Povray is the oldest and most advanced, and can run on a pvm cluster, yafray is relatively recent and can use an openmosix cluster for networked rendering, Blender now integrates a raytracer AND exports to yafray. Those are the 4 programs I know of that I use, but there are more, I just haven't looked for more. So, yes, there is free software for movie rendering already!

Re:Free software ready indeed!

[Shinobi]

Repeat after me: Studios don't use FULL-SCENE raytracing, but they use raytracing for certain things where raycasting can't do a good approximization. Hollow Man is one movie where raytracing was used. They used PRMan for ordinary rendering, and then BMRT was called upon for the raytracing.

And, the major studios want: Speed, quality and a good clean API(so they can add their own stuff too)

Re:Free software ready indeed!

[Viking Coder]

So, you're saying that Blue Sky Studios isn't a professional studio?

In general, you're right - nobody does ray tracing for final renders, but Blue Sky is the exception that proves the rule.

Re:Free software ready indeed!

[levork]

I'll repeat again: professional studios do NOT use ray tracing.

There are at least two full length animated films coming out this year which prove you wrong on this count. Pixar's The Incredibles is one of them. And we're not just talking about selective usage of raytracing either. Raytracing isn't just about reflection maps anymore - in particular, ambient occlusion and other global illumination effects can be done with raytracing in ways that are hard to do with other "cheats". And while possibly slower computationaly, it's easier in human set up cost, which is far more important nowadays.

Re:Free software ready indeed!

[malducin]

Although in a certain way they are the exceptions that prove the rule, or better yet they signal the shift to hybrid solutions (use the best tool for the job). If shots or elements can get away with using regular PRMan features (scanline REYES) with no raytracing I'm sure it'll get used.

But indeed more and more stuff is hybrid where for example PRMan and Mental Ray are combined (either in the same movie, same shot or same element). Some examples of this that spring to mind are the exosekletons in T3, Hulk, Dobby in Harry Potter 2, etc. The biggest counterexample is the Matrix sequels, since ESC uses Mental Ray exclusively. Not to mention the boutique shops that use stuff like Brazil.

This year we'll get even more exmaples, one that springs to mind is I Robot. In any case it's all a matter of context but I guess we should refer to it as hybrid approaches now.

Re:Free software ready indeed!

[Pseudonym]

Five years ago, you were right. Today, you're almost right.

It's still true that professional studios AVOID ray tracing if they can get away with cheating. However, it has always been the case that the cost of a human is greater than the cost of a computer (per unit time). Computer speeds have finally gotten to the point such that if the task is "small enough", it can be cheaper to spend all night using the render farm than to spend all day setting up reflection/refraction maps, tweaking the shaders and painting occlusion maps by hand.

We still won't see ray tracing any time soon on "big" shots (e.g. the room with all the doors in Monster's, Inc), but we will see it happening more often on "small" shots (e.g. the volume caustics in Finding Nemo).

math coprocssor

[PineGreen]

Instead of just using the native 3D engine in the GPU, as done in games, Gelato also uses the hardware as a second floating point processor.
Does this mean that I could eventually use my GeForce to do things like matrix inversion for me?

Re:math coprocssor

[JFMulder]

Some people are already using GPUs for things like Seti@home and collision detection in games, so yeah, why not?

Re:math coprocssor

[Hast]

You can do that now.

Check the GPGPU links.

Re:math coprocssor

[Anonymous Coward]

who is using their gpu for seti@home? i want some of what you're smoking

Re:math coprocssor

[Ride-My-Rocket]

Does this mean that I could eventually use my GeForce to do things like matrix inversion for me?

In Nvidia's Russia, you have the Matrix.

round and round we go

[renderfarm]

ART, OGL assisted, now gelato. Sure there is a place but how do I stick a FX card into my several hundred 1U racks either physically or financially. Have you seen the size of these cards anyway ? Sure some vendors (mental images) are leveraging GPU power and have done the same with OGL for some time but unless the GPU calls are handled by calls to the renderer so you hide behind a consistent API it's a waste of your hard earned time getting your pipeline into shape in the first place. Long live GPU but I don't want to be aware of it's presence. PS. I think ATI are actually the smart kids on the block but chose the wrong colours for their marketing hype.Maybe they can get their chips straight onto the motherboard (much smarter).

Re:round and round we go

[RupW]

ART, OGL assisted, now gelato.

Do you mean this ART? Is their stuff any good? I went to a lecture given by one of their tech guys a long time ago and it sounded pretty impressive.

Re:round and round we go

[renderfarm]

sure I meant ART.co.uk. The argument is the same again. Lock yourself into hardware and 2 seconds later you need to write your own shader and .... then what? And like ART these cards are not cheap. I feel sorry for the ART guys though as they have been up this street for ever and just because nvidia have got marketing clout every college kid with a gforce thinks they have a death star in their laptop. Get real. It takes time and memory to render quality and it has to be totally flexible as techniques change by the minute.

Video Cards as Renderers

[agby]

I was under the impression that it's hard to use a video card for general computing tasks because of the way that AGP is designed. It's really good at shunting massive amounts of data into the card (textures, geometry, lighting, etc) but terrible at getting a good data rate back into the computer. They're designed to take a load of data, process it and push the output back to the screen, not the processor. This is the major reason, IMHO.

Re:Video Cards as Renderers

[Anonymous Coward]

Reality check. How much a final, production quality image takes uncomressed? Maybe 200MB.

How many production quality images you can render in one minute? Maybe 0.1.

You need very little of GPU->system bandwidth, since you will use minutes or even hours render a single frame.

Re:Video Cards as Renderers

[fistynuts]

You're right. Hence PCI Express.

Try Blender3D

[pillendraaier]

Try Blender3D (blender3d.org) Found out about the product just a week ago. I think it rules BIGTIME. Ik uses the accerated OpenGL drivers of my nVIDIA chipset.

Re:'pricey' - but worth it?

[Namarrgon]

Certainly true - more than Entropy's $1500 (when they sold it), more than many others, but still cheaper than PRMan's $3500 + $700/yr.

I think the point is not that it can render just like other engines, but that it can do so at a far greater speed (with a lot more flexibility and features than the PURE card). That would indeed be worth the money to all but the smallest studios - much faster feedback at full quality is an artist's dream, quite apart from the (more expensive) option of using it to accelerate your render farm.

What they don't really say anywhere is *how much* faster it is. There are many factors involved, but if you basically have 16 * 4 * 2 FP execution units running at 400 MHz in a highly parallel configuration, backed by 32 GB/s of bandwidth, there is quite a bit of potential there (~50 GFLOPS vs P4's ~2 GFLOPS?).

For a farm... If, say, it renders 5x faster than a given render machine, then that's 4 machines (and engine licences) you don't have to buy, which would easily cover the cost.

Heeeey.. that dino looks familiar...

[invispace]

IF anyone's interested, the dino on the http://film.nvidia.com/page/gelato.html page was one of of Entropy's flagship images. Entropy was a pay-to-play renderer made under the renderman Spec by a they guy who wrote BMRT. Pixar sued the company that made both of the renderman compliant renderers, and basically forced them into business with Nvidia, who quickly snatched up the company and paid off Pixar. Nvidia had been trying to come up with a hardware shader language like that of renderman, and thusly came out with the shoddy and less than capable CG shader language. Unfortunately, no matter how good that card looks on screen, it's still only going to be a preview render. Straight 35mm film is rendered out at 2048x1556, and you wouldn't believe how tedious CG work is with every single person above you telling you to correct every little thing. The one thing this will do is help look-dev folk and shader-writers out. They get paid enough as is though. Oh.... you might be interested to know, that most of the renderfarms are now at least 1/2-2/3 x86 machines running Linux, and they have been for the past 3 years. No large studios are using SGIs anymore, but surprisingly a lot of the boutiques are using OSX. I guess that's what happens when Apple takes a hint from MS, and buys(Shake) what they want instead of making it.

linux/itanium project

[jonykaos]

Isn't Gelato the same name used for the linux/itanium (www.gelato.org) project ... me thinks this could be a problem ...

Is a beowulf cluster of blue moon ray tracers...

[leonbrooks]

...called a "Jupiter" or a "Saturn" of raytracers? (-:

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:'pricey' - but worth it?

[Hast]

Perhaps you could even use a render box with a bunch of PCI-X cards in it (not sure if PCI-X allows that, I sure hope it does). Give it a few months and the current top of the line cards will have halved in price and you can actually put together such a render machine for a reasonable amount of money.

Then you naturally have to build a Beowulf cluster of those. ;-)

+1, insightful tag line

[Anonymous Coward]

"Religion is the most malevolent of all mind viruses." (Arthur C. Clarke)

's'true. count the number of people murdered in the name of atheism. just russian and chinese communism between them account for more dead in the name of religion than even the roman catholic church has been able to chalk up in millennia of sustained effort. now that's what i call impressive. it's a pity old art went a bit senile towards the end there, it ruined his later books.

Re:+1, insightful tag line

[Anonymous Coward]

sigh.. Dont you know capatilism, socialism, democracy, communism.. they are all religions? Sure, they are based on logic, but as soon as you start associating democracy = good, commies = bad, you might as well just join the national church.

Have you donated to SlashDot?

[leonbrooks]

I'm getting tired of the slashdot mentality that everything has to be free.

If you haven't contributed financially to this blog then you're a serious hypocrite - aren't you? (-:<

Free-as-in-beer isn't absolutely necessary, but it does solve an awful lot of "parking meter change" style problems, and it's a highly viable services leader.

Re:Have you donated to SlashDot?

[Anonymous Coward]

"Free-as-in-beer"

Until the day when I can get everything (food, place to stay etc) for free, "Free-as-in-beer" wont work

Its only a great idea for folks thats paied by an entity that have other goals then making money.

Ie. Educational institutions

Re:Have you donated to SlashDot?

[Anonymous Coward]

Ha!! I don't know which institutions you went to, but colleges, especaily the private ones, are run as business, despite all the attempted apperance of having the best interest of the students in mind.

Linus gave Linux away, Nvidia benefitted.

[anti-NAT]

So why should Nvidia benefit from Linux, without some reciprical giving ? Hardware programming specs would be enough of a gift.

Re:Linus gave Linux away, Nvidia benefitted.

[Anonymous Coward]

"Nvidia benefit from Linux, without some reciprical giving "

LOL, yer too funny

NVidia just did that, gave something back, its not free, but its a good, supported system that if you need you can buy.

You cheapos need to calm down, everything cant be "with out charge" ye know.

Re:Linus gave Linux away, Nvidia benefitted.

[dave1g]

because there is no clause in the GPL that says

"If you use this, you owe me your soul and all possible profits made from the use of this product"

It actually says quite the opposite and encourages companies to try to profit from it.

Re:Linus gave Linux away, Nvidia benefitted.

[CatOne]

How is Nvidia benefiting from Linux?

News flash... they really don't CARE about the OS. They want to ship product, and if 100% of all sales could be on Windows, they'd probably be ecstatic, because they'd only have to write a product for a single platform.

The fact that there are 3 or 4 platforms they have to support only lets them do a lot MORE work to get the same amount of product sold.

So Linux is of no benefit to them at all -- it's a detraction. But if it's a necessary step to get revenue (say because Renderman runs on Linux and big shops ask for it), then they'll do the work to get the money.

But really, Linux is not everybody's savior.

Re:Linus gave Linux away, Nvidia benefitted.

[anti-NAT]

"So Linux is of no benefit to them at all -- it's a detraction. But if it's a necessary step to get revenue (say because Renderman runs on Linux and big shops ask for it), then they'll do the work to get the money."

Maybe I'm not understanding you correctly. One moment you seem to be saying they get no benefit at all, the next you say they are "only" supporting Linux because they make revenue from it. How can making revenue not be a benefit to an organisation whose goal is to make money ? Bare in mind that revenue, unless you make a profit from it after taking out your costs, is pointless.

I would have thought it would be obvious that any business with even a small amount of intelligence only does things which will make it money. If you don't make money, and your intention is to, you are out of business.

Re:Linus gave Linux away, Nvidia benefitted.

[anti-NAT]

My point wasn't in regards to the what the GPL says.

My point was that Nvidia benefits from Linux, yet gives nothing back. That is why I'd never buy an Nvidia video card, and never recommend one to run under Linux.

Nvidia would be valued much more highly in the open source community by making some sort of reciprical gift. The level is up to them - no demands are being made. However, they would find that their value as a contributor would increase as they give back more. Mind you, they aren't expected to give everything. It is up to them.

Nvidia would have plenty of other examples of companies who have realised that recipricating benefits both them and the community - IBM, Intel, AMD, SGI, HP, etc. etc. etc.

Re:Linus gave Linux away, Nvidia benefitted.

[HuguesT]

That's not quite true. Nvidia users benefit by being able to use their video hardware fully under Linux. That's not something to sneeze at. Most users don't care if the driver is closed-source, as long as it's still free.

ATI (to name but one) doesn't make the same level of effort towards Linux users and as a result ATI users are in for a world of pain and wasted hours trying to make their video card work correctly under Linux. I know, I've been there.

Everyone rails about Nvidia's level of support in Linux but in reality it's not that bad. The newer drivers support RandR, and works better with SMP. The level of functionality for the run-of-the-mill PC has always been exemplary. One just need to report bugs and be patient, exactly as in the open-source case (again from the perspective of a mere user, not a developer).

If on the other hand ATI realized that Linux users constitute a worthy user base and decided to open-source and support their drivers fully, then they would get a competitive advantage with Linux users. We can assume that their drivers would quickly become high-quality and their hardware would be the preferred solution for Linux users.

Nvidia would have to reciprocate and open their driver assuming they still value the Linux customers. Until then they are making the right decision (as a resource-conscious company) that costs them the least amount of money in the short term while not jeopardising the bottom line, given that their hardware is still the best supported under Linux, and that's all most users care about. Presumably open-sourcing the Nvidia drivers would not be trivial, might expose trade secrets and other unpatented techniques (or even unlicensed used of techniques used by others), and would generally require a higher level of commitment to Linux users.

Seeing however that ATI hasn't made that decision yet and that their driver is decidedly inferior to that of Nvidia, Nvidia doesn't have to make the expensive decision to support the Linux users in a different way that they support Windows users.

Nvidia doesn't owe Linux users or developers anything that I'm aware of. The fact is that if Linux didn't exist they would sell just as many video cards. Those users would still be running some other kind of supported O/S, probably Windows.

Conversely if Nvidia withdrew binary-only driver support for Linux then users would be temporarily inconvenienced but in the long term independed developers would reverse-engineer enough of the driver to make their hardware (at least the older versions of it) work well enough with Linux again. In effect their binary-only support is really not indispensable, so Linux users don't owe Nvidia a great deal either.

developer).

So what? No one is under any obligation, and that's fine.

Re:Linus gave Linux away, Nvidia benefitted.

[anti-NAT]

That's not quite true. Nvidia users benefit by being able to use their video hardware fully under Linux. That's not something to sneeze at. Most users don't care if the driver is closed-source, as long as it's still free.

And that is sad. Although technically, due to the GPL, they aren't stealing, they have the same mindset a thief does - take something of value, and provide nothing in return.

Nvidia and ATi (to a lesser extent, they have released some programming info for their older cards), have a similar mind set. They take advantage of the market that Linux has created, yet also don't provide anything in return to the community that created that market - the open source community broadly, and specifically the Linux / XFree86 (and X.Win) communities.

I'm no kernel hacker, yet I return the value I've received from open source software, such as Linux, by regularly providing help in the Linux/BSD forum at Whirlpool. I realise I don't have to, however that is how I "maintain" my membership of the open source community.

I don't really care too much for the issues ATi or NVidia users have with the corresponding binary drivers. Those problems can always be traced back to the closed nature of the proprietory drivers. I prefer to spend my money with vendors of hardware who provide open programming specificiations. Even better if they sponsor open source driver development. I will reward them financially for understanding, and participating reciprically in the open source community.

I'm sure you can guess, NVidia will never see a dollar of mine. ATi maybe, although it will be for "old", "cheap" hardware (ie. the 9200 series, for which they openned the programming specs) that they make very little profit on.

Re:Linus gave Linux away, Nvidia benefitted.

[cubic6]

Another way to say it is that they have the mindset of a profitable corporation: Take something of value, pay no more than is required, and show the resulting profits at the next shareholders meeting. It really doesn't matter what their mindset is. They're following the rules as layed out, and you're attacking them for not following *your* rules. This doesn't make them a thief.

The bottom line is that you can't say "Look, here's some free software, open for anyone to use" and then, when someone uses it successfully under those terms, say "Well, you have to contribute back now, damnit!". You can't steal something that is free. If you were to change the license to specifically state that users were required to give something back, then Linux would *not be free any more*.

Re:Linus gave Linux away, Nvidia benefitted.

[anti-NAT]

This doesn't make them a thief.

I sort of expected that accusation. Note that I didn't say they were a thief, just that the logic behind the decision not to provide some sort of contribution of value back is the same as the mindset of a thief. Another, much simpler word for it is "selfishness".

I don't have that much issue with individuals who take a copy of Linux, and give nothing back. I'd prefer them to do something like help out a friend with it, or contribute a helpful comment to slashdot, or report a bug. They then are recipricating.

It is corporations that take it, use it to make money, and yet feel no obligation to provide something, anything, in return. I agree, those non-contributing corporations have the right to use it. I, however, won't be rewarding them for doing so.

Re:Linus gave Linux away, Nvidia benefitted.

[HuguesT]

Thanks for taking the time to respond. Yes Nvidia is being selfish and so are the millions of users of Linux who are installing it on their machine because it costs nothing and works well enough for them, yet contribute exactly nothing in return.

Except that's not quite true. The millions of users contribute usage. That means recognition, bug reports, significant market share, complaints about usability which are eventually taken into account, articles in the mainstream press, visibility, even minor celebrity for a few, etc. This means there exists a market for Linux journals, for Linux distributions at your newsagent, for interest raised in big industry (IBM, HP, etc), concern at Microsoft, competition in the marketplace.

Nvidia contributes to that positive swell. They provide a high-quality, free-as-in-beer, same conditions as under Windows, driver for their hardware. ATI doesn't, just to name one (at least their drivers are not high-quality). This means joe random user can start playing some games under Linux (thanks J. Carmak) *which are not free either, not even as-in-beer*, people can start rendering movie quality stuff, 3D rendering works well etc. Linux gain yet more users. If Nvidia hadn't made that effort the acceptance of Linux would be even less than what it is now.

All of this benefits you as a user because the general quality of Linux improves as a result of all this. It's hard for a company like Nvidia to understand the new values in particular those of Free Software. You've got to give them time.

In the mean time I respect your position and I applaud you for sticking to it. Kudos.

Re:Linus gave Linux away, Nvidia benefitted.

[CatOne]

What I'm saying is, that given a fixed number of customers, Nvidia would like to write the least number of ports to get all those customers. So given a million users TOTAL who would buy their products (this number pulled from you-know-where), their ideal situation is for all million to be on the same platform (most likely Windows). If the truth ends up being there are 850,000 Windows, 100,000 Mac, and 50,000 Linux, they have to maintain 3 ports, but still get the same number of total users (customers), which is 3 times the work for the same revenue stream.

So they need to decide if the incremental business (50K users) is worth the expense of the port, to them.

Re:This would be more useful

[Anonymous Coward]

Post production studios are Linux based. It used to be SGIs, now it's Linux.

impressive feature set, but I'll wait to judge

[Anonymous Coward]

whether or not it will make any difference. Since I'm not qualified to make any useful judgement as to whether high end studios will buy a ton of these for their workstations. Assuming that it produces the desired results and allows animators to render the movements with greater detail than current workstations with normal Pro graphics cards, I could see it slowly growing. From what I know of animation shops, the animators normally use shaded view to work out the movements. The shaders and lighting are handled by guys who specialize in shading routines. Of course an animator could get some stock in-house setups to test a sequence, but they're used to viewing in shade mode.

Their ability to visualize motion and subtleties in their head is what makes them valuable. They know what will look good a lot sooner than someone unskilled in the art of key-frame animation.

3dfx

[TheRealMindChild]

A lot of these "features" were present in the 3dfx Voodoo 5 line... and since Nvidia aquired them, I would have to say this is their implementation.

Re:3dfx

[Anonymous Coward]

Do you mean that this technology was being developed by 3dfx?

Have you got more information about that?

I guess that one wond't have been propietary...

Seems to be Open now?

[leonbrooks]

See here and here. More and more pieces of the moviemaking toolchain are available Openly, only a matter of time before someone adds a GUI wrapper to integrate it all. Will they dare call it Raxip? (-:

Re:Seems to be Open now?

[WWWWolf]

Aqsis and other open-source Renderman renderers are not based on BMRT source code (which has never been released, I think) - they're all independently developed.

Renderman specification, as I understand it, has been "open" for a quite a time now. You can download the specs for both the protocols/APIs and the shading language. The main difference between commercial and open-source renderers is the supported feature set and the level of speed optimisations...

And yeah, it's nice to see Renderman stuff available as open source. Now, if only Blender folks would get us Renderman integration instead of three-quarters-convenient scripts, then we would also have a nice modeller as well...

Re:Seems to be Open now?

[orasio]

Please, stop whining, if Blender folks don't want to integrate with Renderman, there are many things you can do to change that. You can pay them, or someone else to do it, or code it yourself, or raise the money for the feature.
Anyway, those are things you can do with the thousands of dollars you are saving for not buying maya, and you could help others, whining and bitching doesn't help anybody.

Re:Seems to be Open now?

[WWWWolf]

Sorry, I wasn't whining, just hoping aloud. Apologies if it sounded like whining. Wasn't the intention.

I don't have much money (empty bank account and 5 Euros cash...) and I don't know much about Renderman's specifics, so all I can do right now is to express my wishes that these features are implemented some day. Hopefully, someone with either the money or expertise (or both) can hear this pathetic little plea and do something about it. As for myself, I sure don't mind waiting a few years, as I'm not paying so I'm not in the position to demand anything to happen immediately. Besides, Blender rules even without Renderman support - the internal renderer and Yafray appear to be improving nicely =)

Re:Seems to be Open now?

[orasio]

Allright, sorry, there are so many whiners and bitchers about great projects that I get a little oversensitive.
On realted news, it can be less than a few years, maybe 1 year, because I am starting to get the expertise in renderman to make such a converter, or improve one, and have seen the necessity myself for a good Renderman export. After I finish the university, it's on my to-do list, a Renderman export in blender, for use with a GPU enhanced renderer.

You're kidding, right?

[Namarrgon]

Almost every FX house? I don't think so.

PRMan is a fine product, but it has its limitations, as well as its price. There are numerous competitors, many of which use the same Renderman interface but offer more speed and/or more features at a lower price (BMRT and Entropy are[were] notable, and relevant, until Pixar squashed them with the threat of an expensive court case). Brazil, AIR, etc - these RIB-based renderers drop into the same place in the workflow.

Please explain to me why a dedicated rendering device from NVidia would be any better than your average UNIX or Linux machine?

Only if you explain why your average UNIX or Linux machine is better than a Commodore 64 or a PDA, which is also "in essence nothing but processor operations" etc :-) If you listed SPEED in there, you're on the right track.

A modern GPU has far more floating-point hardware than any general-purpose CPU, and it's all geared towards the process of rendering pixels. For certain tasks, one of those expensive dedicated rendering devices from nVidia could be better than FIFTY of your "average" UNIX or Linux machines! Is that enough of an advantage to consider?

Dang, I went and fed the troll, didn't I...

Re:You're kidding, right?

[malducin]

Almost every FX house? I don't think so.

It depends on the context. Most major and high end VFX houses heavily rely on PRMan, like ILM, Weta Digital, Digital Domain, Imageworks, Tippett Studio and more. Mental Ray is also being used quite a bit, ESC in particular is an MR only place. On the other hand the small and boutique houses tend to use stuff like Brazil, Lightwave, etc. Places like Cafe FX, Blur Studios, etc. For example the movie The Core was done mostly by small houses and had quite a bit of 3DMax use coupled with something (Brazil I think). But most major VFX movies have PRMan rendered in them simply because it's what the major use and they work on the big projects. A handful of places have propietary.

Just to witness some of the biggest VFX films this year will probably rely quite a bit on PRMan, like Van Helsing, Day After Tomorrow, Harry Potter 3, Troy, Spider-Man 2 and I Robot. Actually Pixar has a partial lists of where PRMan has ben used:

PRMan Movies

As far as price, well PRMan is a product for professional of a very niche market. As far as limitations, well almost every other renderer might be a bit more limited IMHO. PRMan is geared towards technical users.

Re:You're kidding, right?

[Namarrgon]

True, the heavyweights do lean towards PRMan (they can afford it), and PRMan has made good improvements in recent times. The alternatives have been quite functional (and at times rather more functional) and a good deal cheaper, which made e.g. Entropy very attractive to the smaller studios. Pixar certainly felt threatened enough to shut Exluna down.

I know The Core in particular had some shots rendered with Entropy, and houses like Blur prefer Brazil (perhaps since the author of Brazil used to work there). A lot of the smaller post houses don't feel the need to shell out the premium for PRMan, and have opted for one of the various alternatives.

PRMan might be The Standard, and they've brought it more up-to-date in the last year or two, but it's far from the only usable alternative.

Re:You're kidding, right?

[Animaether]

Heh.. funny.. some of those houses and some of those movies actually used Brazil r/s under 3ds max to an extent as well.
If you happen to be able to go to NAB, visit the Boxx Tech booth. SplutterFish are exhibiting there and have some nice promo material including recent movie stuff.

The big problem is that the bigger houses don't really want to let on what software they are using. ILM, for example, uses Brazil r/s - but that only became public in an obscure interview (I'd even have to dig around to find it.)

Re:You're kidding, right?

[malducin]

Well it wasn't that obscure. An article on CGW from last year about matte paintings specifically mentioned that. The Rebel Unit at ILM is mostly PC based and they do use Max, Brazil and AE if I remember right. More or less the same with Imageworks (the work on Bad Boys 2 was discussed).

Couldn't go to NAB but hopefully I'll see it at SIGGRAPH.

But ATI's solution is free

[Enbar]

Admittedly it's not exactly the same things as NVIDIA's solution, but the main component of breaking big movie quality shaders into multiple passes is in ATI's Ashli (http://www.ati.com/developer/ashli.html). The big plus is instead of costing thousands of dollars it's free. Also I noticed everyone is saying agp read backs make this sort of thing useless. The fact is that most of the scenes rendered will take seconds to hours on the graphics card (vs. minutes to days on a CPU). The slow agp reads aren't going to be much of a performance impact in these situations.

Re:But ATI's solution is free

[forkazoo]

I have mod points, and I really want to do a little bit of smacking down, but I'll just go for correcting instead. I phear the metamods, yo!

ASHLI is *not* a renderer. It isn't anywhere near doing what Gelato does. Gelato takes a scene file, and gives you a picture. It does it very nicely, using motion blur, programmable shading, and all sorts of fun stuff like that. It is written by the Ex - Ex Luna boys. (Larry Gritz, Matt Pharr, Craig Colb -- Three mofos who know their shizzle.)

ASHLI takes a renderman shader in RSL, and gives you a compiled shader for OpenGL or DirectX. It's then up to you to write the whole renderer.

It's cool, but if you are seriously writing lots of RenderMan shaders, you can probably just as well write them in GL Slang, for your in house customised OpenGL renderer. ASHLI's utility is limited. Frankly, it is pretty much a complete non sequiter in a discussion about a renderer.

You mean PCI Express?

[Namarrgon]

There are no PCI-X-based GPUs I'm aware of.

PCI Express is a possibility, if you can find a motherboard with more than one x16 slot (or a GPU that fits an ordinary x1 slot). Doubt you will for some time, though.

And you wouldn't have to use top-of-the-line cards, either. Something mid-range, with more bang for the buck would do fine.

But yeah, one day. Great to see how the demand for better games has resulted in cool hw/sw like this flowing on to my own industry :-) Now we just need to figure out how the pr0n crowd can drive up demand for systems with multiple gigabytes of RAM...

Yes, but it's not a problem

[Namarrgon]

See above.

General computing tasks, yes, AGP and GPUs are not so useful. For rendering though, they're near-ideal, even with AGP (though PCI Express would certainly be better).

Wax too

[magullo]

I'm really excited about the future of Jashaka, but for now I stick to Wax (Win only, but more stable and usable than Jahshaka at this stage).

Why would this be a useful app?

[Anonymous Coward]

Seems that if you wanted extra processing power it would in most cases be cheaper in terms of price/performance to just add another box to the job, or a(nother) coprocessor card.

Gelato is Software

[MikeD83]

Gelato is a $2,750 software package. It is intended to be used with an Nvidia Quadro FX 4000 workstation video card. The video card has not hit the market yet, but the Quadro FX 3000 goes for $1,300. Which brings the total cost for the package at between $4000 and $5000 per machine.

How Fast Is It?

[Anonymous Coward]

I see lots of posts implying Gelato is 10x or 100x or "Super-Duper" fast. But how fast is it? It is really even 2x faster than RenderMan? Amazingly since Entropy came along, the folks at Pixar seem to have taken a renewed interest in improving RenderMan in both speed and memory.

It will be interesting to see what this thing can really do and how well it actually uses the hardware co-processor, especially given the "Super-Duper" cost.

Wow, it's only 2004

[AugstWest]

and someone has finally invented the Toaster.

Re:Naming...

[Ignominious Cow Herd]

Between Gelato and Fluendo I think we've officially run out of good names for products. I suggest we start naming things like astronomers do. Lets call this one NGC 2607 or something like that.

Crypto Tool?

[Panaflex]

What gets me excited is this line:

Key to this doctrine of no compromises is the nature of how Gelato uses the NVIDIA Quadro FX GPU. Instead of just using the native 3D engine in the GPU, as done in games, Gelato also uses the hardware as a second floating point processor.

WOW. That would be a fast FPU, I'm supposing. How fast can it sieve?

Pan

blinksmart

[Doc Ruby]

Video coprocessors are really fast at multiply-accumulate DSP functions. Those are linear equations (y=mx+b), which underlie lots of computer modeling techniques that aren't graphical. For example, neural networks. And speech recognition. Where is the open source app that runs the video chip against a memory buffer of nonvideo data for other apps?

No, there's not.

[Moryath]

"is there any Free software capable of exploiting the general computing power of modern video cards?"

No.

The actual problem is in how the hardware today works. Video boards are REALLY, REALLY GOOD at getting an image moving in one direction - from computer, to vid board, to monitor.

They are quite LOUSY, however, in sending that video information back to the computer or the hard drive.

This is why even the gameplay movies you see today have been created with a simple workaround: hook up the board to a capture card of some sort (sometimes as simple as a vga->TV external converter) and capture away.

I'm guessing the big proprietary change in NVidia's model is an allocation of bandwidth back for sending the data the other direction - the rest of the stuff (raytracing etc) is just fluff. And of course it doesn't hurt in the slightest that it doesn't have to run at 60+ FPS all the time, so it can take the extra time to make frame, send frame back other way, make frame...

Re:No, there's not.

[Crazy Eight]

Video boards are REALLY, REALLY GOOD at getting an image moving in one direction - from computer, to vid board, to monitor.

Isn't this just a driver limitation? Does the AGP bus have a unidirectionality built into it?

There is some serious interest in utilizing graphics hardware for non-graphical computation. The GPU would be treated like a co-processor. It would function in relation to the CPU in a manner analogous to the way MMX, SSE, and Altivec, relate to standard instructions. This site is a kind of focal point for people researching the notion.

Re:No, there's not.

[cubic6]

Yes, the AGP bus has unidirectionality built in. PCI-Express has the possibility to change this, since (IIRC) it has equal bandwidth both ways.

They are pricey...

[peu]

today, but wait for the China clone flood and you'll have this technology next year in your desktop...

you can do it auto

[Anonymous Coward]

"You can do it Otto! You can do it Otto! Make that spare! I'll give you free gelato!"

Re:Guess what?

[Anonymous Coward]

Actually you are, you little fuckwad.

Combustion not "as good" as AFX?

[adamgeek]

"Combustion like features for the price of After Effects."

honestly, in my mind, combustion and aftereffects are tied [for my uses]. I end up using AFX more often, but that's simply because of the myriad of plugins available for afx that dont exist (or that i dont own) for comb. comb is a great tool, and in some ways signifigantly more powerful than afx.. but in many ways they are just two different flavors of the same ice cream. i'm not sure what "combustion like" features you're referencing that AFX can't replicate..

in my mind, the high end video market is grossly over-inflated (price-wise) currently.. where top of the line stuff is literally 20x the price of the product beneath it that does half as well. 2k and 4k realtime editing.. very cool, but charging $100k for that software, when one could do 2k and 4k "not quite so realtime" editing in a variety of $500-1000 apps.. annoying. course, i guess they figure that if you have the money to shoot on a 2k format, you have the money to fork out for their products. if this card helps to drag some of these prices back down to earth, it wouldn't make me cry =D

Linux Rendering Clusters

[dj_cel]

As Maya works on Linux, I beleive that Weta Digital (company responsible for vfx i LOTR etc) uses a massive Linux based render farm. At my school we use 3dsmax 4, but will now be switching to Maya 5 and supposedly running it on Mac hardware. I have put together a proposal for the IT dept for building Linux boxes to run Maya instead for a fraction of the cost of the Mac hardware which they seemed to like, but unfortunately the ignorent upper level execs did not favor a decision to move to Linux regardless of a pretty competent IT staff. They felt that students would not want to adopt a new os and teachers would be unable to solve technical issues on Linux boxes. I think that universities should start making a jump to using Linux based solutions, but realize there is a large gap in the administration because most of the teachers (not all) are not comfortable with a new os. I have used Maya on Linux Mac and PC and found that it performed equally if not better than Mac and far better than PC, mainly because it was originally authored for IRIX. I look forward to the possibility of using hardware in combination with powerful software to doing complex rendering that usually takes several minutes per frame. Hopefully the onset of this technology will bring more users as well as other publishers (please Adobe please!) to Linux.

Would be nice to have it running on OSX as well

[saha]

First I'd like to see a greater selection of graphics hardware on the G5 machines. The Radeon 9800 are good, but a Quadro FX 4000 or 3000 would make selection better. I don't think Apple should try to get 3DLabs Wildcat boards on the G5s. Applications like Gelato running on OSX would give the platform a boost as well.

For the kind of work I do, a Nvidia Quadro FX 3000G would be best for driving large displays.

Wish people wouldn't have NS collisions

[xenophrak]

Cool! My distro already has m4. But is has some bugs in it still.

m4 ./movie.mpg

=moovlmvhd-3A-3EX@"trak\tkhd -,-3EF@$edtselstF mdia mdhd -3A-3A+w-9hdlrmhlrsounappl?Module de gestion sonoreminfsmhd;hdlrdhlrali sappl@KGestionnaire dOalias Apple$dinfdrefalisHstbl4stsd$raw +wstts -àstsc

m4: ERROR: EOF in string

0.2 FPS/machine * 120 machines = 24 FPS

[tepples]

Gelato is definitely NOT realtime.

"Machinima" is CG puppetry. If it takes a CPU and video card no more than five seconds to render a frame at full resolution, and a production company using digital puppetry techniques has a server farm of 120 CPU + video card systems, doesn't that equal real-time at 24 frames per second? A matching 5-second delay on the audio would give the censors time to bleep whatever swear words the target audience's government bans from the air.

Jashaka in early stages

[HenchmenResources]

Yes unfortunately as magullo pointed out Jashaka is still in it's early stages, so at this time it is not a fully working program. the good news however is that although it is an open source project it does seem to be moving forward at a steady pace and seems to be staying on track with the designers time line. Iv also found that the news of the Nvidia release of Gelato was announced on Jashaka's .org page as well, but the note seems to indicate that the Jashaka project will not be using the Nvidia API but will still support the new cards. Unfortunately I am unable to clarify that point since the article discussion board seems to be discussing other things. worth a read all the same.

We need benchmarks! Benchmarks I tell you!

[Moekandu]

What it all comes down to is price/performance. Will you get faster rendering by spending the $4K on Gelato, or buying two or three more dual CPU boxen and ingnoring the GPU advantage?

Also, how does the performance of Gelato scale based on the CPU's in the box? Does it render significantly faster if you've got Opterons instead of Athlon MP's or Xeons? What about number of CPU's?

In any case, Gelato needs to be some truly slick stuff, or it won't be able to justify it's price.

Nothing terribly new

[Wolfier]

Take a look at these:

http://graphics.stanford.edu/papers/rtongfx/

http://artis.imag.fr/~Nicolas.Holzschuch/0610200 3/ Paulin/DensityEstimation.pdf

People have been (although fairly recently) toying with the idea of using GPU to accelerate global illumination for a while.

Props must go to nVidia tho for combining and putting all these ideas into a nice package.

Physics APIs- Direct KinetiX and Open/mgh

[Intocabile]

With physics in games becoming more and more advanced how long till we see an API for hardware accelerated physics. Not saying you'll be shelling out anymore cash for a ElectricForce FX physics card but have dedicated hardware on graphics cards for physics calculations. Not only could this boost performance but it could combined with pixel shaders and geometric transformations to increase what is possible. Who knows, the GeForce 6800 might be programmable enough to do this already, albeit at a performance hit to the graphics pipeline. There are some pretty incredible physics demos that can be found on the internet everyone's favourite monopoly(Crash Video) has one that showcases their new game development suite XNA. I don't see this level of physics being available in any game any time soon but with hardware accelerated physics who knows what's possible. Even without a standard API to build on developers might implement their own physics acceleration shaders(for lack of a better term) in cases where the CPU is the bottleneck. Ever since the first programmable GPU was release I imagined that they could accelerate more then just graphics and be used to increase the responsiveness of computers.

Sounds like an old (Lucasfilm?) Siggraph paper.

[Ungrounded Lightning]

[...] advanced rendering features: displacement, motion blur, raytracing, flexible shading and lighting, [...]

That sounds like an old Siggraph presentation I saw a decade or two ago when I used to go to Siggraph. Lucasfilm, I think. (The fine sample picture in the article showed a motion-blured image of a set of pool balls in motion.)

When rendering an image using raytracing, there are several effects that are achieved by similar over-rendering processes. I.e. you ray-trace several times varying a paramter:

- Depth-of-field (use different points on the iris of the "camera", blurring things at different distances from the "focal plane".)

- Diffuse shadows (use different points on the diffuse light source(s) when computing the illumination of a point.)

- Motion blur (use different positions for the objects and "camera", evenly {or randomly} distributed along their paths during the "exposure" - ideally pick the positions of the whole set of objects by picking several intermediate times, rather than picking the postion of each object separately, to avoid artifacts of improper position combinations.)

- Anti-aliased edges. (Pick different points in the pixel when computing whether you hit or missed the object or which color patch of its texture you hit.)

As I recall there were about five effects that worked similarly, but I don't recall the other(s?) just now.

To do any one of them requires rendering the frame N times {for some N} with the parameter varied, then averaging the frames. (Eight times might be typical.) Naively, to do them all would require N**5 renderings - 32,768 raytracings of the frame to do all five.

The insight was to realize that the effects could be computed SIMULTANEOUSLY. Pseudorandomly pick one of the N from each effect's set for each frame and only render N frames, rather than N**5. Eight is a LOT smaller than 32K. B-)

Sounds like Nvidia ported this hack to the firmware for their accellerator.

Re:Sounds like an old (Lucasfilm?) Siggraph paper.

[Wyzard]

The paper you have in mind is "Distributed Ray Tracing" by Robert Cook, Thomas Porter, and Loren Carpenter, all from Lucasfilm. It was published in the ACM "Computer Graphics" publication, Volume 18, Number 3, July 1984. If you have access to the ACM Digital Library, you can find the full info and a PDF link here; otherwise, you can download my copy from here. (This redistribution is permitted by the paper's copyright notice.)

Distributed ray tracing isn't a "hack", though -- it's a fundamental principle of modern global-illumination rendering systems. Kajiya introduced the rendering equation in 1986 (ACM citation here, my copy here), and Monte Carlo integration is used as a means of solving it. Monte Carlo integration approximates the value of an integral by sampling random points in the function's domain; in graphics, the domain is the illumination in the scene, the point samples are taken by tracing rays, and the random distribution is what originated in distributed ray tracing.

Re:Sounds like an old (Lucasfilm?) Siggraph paper.

[Ungrounded Lightning]

The paper you have in mind is "Distributed Ray Tracing" by Robert Cook, Thomas Porter, and Loren Carpenter,

Thanks for the reference.

(For those who don't want to follow the link and read the paper, the fifth item was translucency - which this approach models by using a probability function for whether the ray penetrates the surface, rather than splitting the ray into a pair of weighted rays and tracing them both - which rapidly explodes if you have several translucient surfaces in the image).

Distributed ray tracing isn't a "hack", though -- it's a fundamental principle of modern global-illumination rendering systems.

Yes, ray tracing is basic, as is Monte Carlo. Comgining them MIGHT be considered a hack, or straightforward synthesis. But that's not the breakthrough that I was referring to when I said "hack".

You need to remember the environment of the time.

This was twenty years ago - and compute power was twenty years farther back on the Moore's Law exponential. That's a LOT less bang for the buck. (Using the 1.5/year formulation you can do well over 8000 times as much computation now for the same dollar.)

Ray tracing was known to be "right". But it was also extremely compute intensive. So even basic ray tracing was considered dreadfully expensive. The bulk of the graphics research at the time was directed toward finding ways to get good rendering without incurring such a compute cost. Most of the work consisted of finding shortcuts, then patches to correct the artifacts of the shortcuts, then patches to correct the remaining artifacts of the shortcuts plus the previous patches.

Cook, Porter, and Carpenter bucked the trend and examined the combo of Ray Tracing with Monte Carlo, despite the multiplier over basic ray tracing. So far so good. And they found that Monet Carlo raytracing could do not one or two, but FIVE difficult "fuzzy" effects that were EACH considered difficult, producing images of breathtaking perfection. But that's STILL not what I called a "hack".

The breakthrough was recognizing that once you had incurred the small constant-factor cost over basic raytracing to achieve ONE of the effects, THE OTHER FOUR COULD BE HAD FOR FREE. THAT's the "hack".

Again consider the time. Look at the pictures in the Siggraph journal - where the best results of a dozen research projects might be published every month - with glitches, low resolution, visible triangles, and artifacts galore. Then comes this breathtaking shot of the billiard balls in action, with only its perfection to separate it from a high-resolution digitized photograph.

Over the next several years the combination of this algorithm with the explosion of computing power in the supercomputers of the time reduced video synthesis to a "solved problem".

Indeed, the explosion of supercomputer construction was largely driven by this very "killer app" - at one point well over half the Crays in existence had been ordered for entertainment and advertising video synthisis. (It was only later that the rampup of desktop power combined with this algorithm's parallizability that led to "render farms" of cheap, fast desktop workstations.)

Yes, it's a "fundamental principle of modern global-illumination rendering systems" - NOW. Put the emphasis on "modern". But (like the sealed-in-steel battery), at the time it was the brilliant breakthrough that, only after the fact, seems obvious.

Re:This would be more useful

[Wyzard]

POV is literally child's play compared to the software used to render movies. It isn't remotely photorealistic -- even the natural falloff of light intensity over distance due to the inverse-square law isn't properly simulated. (Instead we have an approximation using the fade_power and fade_distance keywords.) The shading model (based on pigment, finish, and interior) doesn't account for effects like anisotropic reflection and subsurface scattering that occur in nature, and the idea that RGB (1.0, 1.0, 1.0) is "white" discourages the use of high-dynamic-range lighting.

Don't get me wrong, POV is a great piece of software for learning, and I wrote a few features for it in years past (variable reflection, metallic reflection, part of conserve_energy, and a blurred-reflection feature that doesn't seem to have made it into the current version). But trying to use it for serious production graphics is like using a Talking Whiz Kid as a word processor to write a Ph.D thesis.

Ya'll want FUD with that?

[adolf]

How slow is slow?

I was under the impression that AGP is just like PCI, but with faster reads from system RAM and a dedicated bus.

Which is to say that it suffers the same limits that 32bit/33MHz PCI does. So it's good for 132 megabytes per second, less overhead and fudge. And since it's got its own set of wires, it doesn't even have to compete with such things as the NIC. Nice.

If it's actually doing the rendering on the card, all it has to do is push back finished frames. 24fps 640x480x24 video consumes a bit over 20 megabytes per second, which is well within the capabilities of what we've got with AGP.

Besides, with the application in mind here, realtime isn't even a pipedream: it's an impossiblity. Minutes of video take hours or days to render. Bandwidth requirements are really pretty minimal.

I think it'll be justfine, despite your FUD.

A good move

[ae-valkyre]

The push to hardware powered tasks is a good one.

Bite out of your hat ?

[Animaether]

Although strictly speaking I'm not sure whether it was used on any movie... ...there -was- a plugin for 3ds max, by Intergraph, called RenderGL. It used OpenGL hardware to accelerate some aspects of an otherwise software-based renderer.

Re:Bite out of your hat ?

[Viking Coder]

Maybe a nibble. =)

Also, there have been several OpenGL renderers for things like BMRT, but they were never for rendering final frames of a movie.

Again, maybe a nibble from my hat. =)

GeForce 6800

[deniss]

GeForce 6800 Ultra Review