NVIDIA's New Pro Graphics Quadro FX 3000 Reviewed

SpinnerBait writes "NVIDIA recently took the wraps off their next generation Workstation Graphics card, the Quadro FX 3000. This card is based on the same general GPU architecture as the NV35 but optimized for CAD and DCC applications. This article over at HotHardware shows what the new Quadro FX 3000 is capable of and it makes a strong showing. However, you've got to pay to play (or work) on this card, that's for sure."

Comments

Most high end graphics cards STILL suck.

[Hayzeus]

But once I see FOUR-dimensional acceleration, /then/ I might upgrade.

Re:Most high end graphics cards STILL suck.

[forgotmypassword]

You have to go faster to dilate time!

Re:Most high end graphics cards STILL suck.

[NanoGator]

"But once I see FOUR-dimensional acceleration, /then/ I might upgrade."

You'll get your chance when stereoscopic monitors come out.

Re:Most high end graphics cards STILL suck.

[Anonymous Coward]

Whassat? Did you say dilate time?!

(Sorry, you left it wide open.)

I don't need it.. yet at least

[qmrq]

I've yet to find anything that gives my Quadro4 a real workout.. I'll stick with what I have for now.

Optimized

[mopslik]

...but optimized for CAD and DCC applications.

That's funny, I thought nVidia was "optimzed" for 3DMark2003.

Re:Optimized

[Worminater]

optimized forwhatever looks better on paper:-p sigh, frustrating when deciding what to buy:-p

Re:Optimized

[Dark Lord Seth]

I only go for video cards that are optimized for my wallet.

Games!

[Ikeya]

Yeah, yeah... So it's good at doing work. That's boring.
What we really wanna know is how many FPS can we get in Quake?!

Re:Games!

[wmaker]

/timerefresh ... 544fps

Re:Games!

[Anonymous Coward]

Quake is old school now. HL2 and Doom will be the new standard.

I am going to go camp for HL2 right after this post. I better hurry, so I can be first in line.

Re:Games!

[synth7]

Next time read the article first. The information you wish is presented on this page of the review.

crazy

[Boromir son of Faram]

I can't believe how powerful video cards are getting. Is it just me, or has the computer industry gone a little too far with the whole "bigger is better/more-more-more" consumerist mantrae? I mean, currenty video cards render full color frames faster than the human eye can perceive. A cheap PC has enough RAM to store the entire contents of the Library of Congress many times over. Most commodity hard drives are in the 100+ GB range. And yet the hardware producers continue churning out bigger, faster machines, as if anyone will ever need a 10GHz CPU. When will enough be, finally, enough?

Re:crazy

[Lane.exe]

Who let the non-geek on here?

just kidding buddy. But you should know... you need p0w3r for t3h sake of p0w3r.

Re:crazy

[msgmonkey]

Yeah but they are also getting more and more expensive. On the other hand though somoene must be buying them.

Re:crazy

[pheared]

And yet the hardware producers continue churning out bigger, faster machines, as if anyone will ever need a 10GHz CPU. When will enough be, finally, enough?

It's because people keep using more and more things like perl and Java.

OK sorry, cheap shot. :-)

Re:crazy

[Shut the fuck up!]

A cheap PC has enough RAM to store the entire contents of the Library of Congress many times over.

I am not really sure how big in megs or gigs the LOC is, but I really doubt that this statement is true. LOC is always passed around as a joke, anyone know the actual size?

Re:crazy

[Verteiron]

You may not need a 10GHz CPU. I know I don't, at least, not yet. But as interfaces improve, as software becomes more complicated, hardware like this may well a requirement. After all, monitor images are still nothing like as detailed as a printed page, and even a very high-end monitor has only a fraction of the resolution that can be produced by a $100 printer. It's going to take serious hardware to acheive that kind of image clarity in real time. 10GHz CPUs may be just the beginning.

Re:crazy

[2nd Post!]

When someone designs something that uses all of it.

I own a Mac, so the situation is slightly different, but as a user of iMovie, iDVD, and Final Cut Express, I can use *all* the ram, CPU, and HD space I can get. Two hours of footage take 30gb of raw storage. Rendering 5 minutes of video takes 30 minutes. iPhoto takes 1.2gb of ram.

So for me, it's not enough. I need more.

Re:crazy

[hawkbug]

You say cheap shot, but I agree. Java, .NET, and any other programming language that runs by relying on a "runtime environment" is definitely going to eat up cpu time. Yes, it's easier to program with, but it's the reason hardware needs to be faster and faster all the time.

Re:crazy

[Anonymous Crowhead]

According to this, 20 terabytes. Now please direct me to the store that sells cheap pcs with 20 terabytes of ram.

Re:crazy

[randomdef]

books? come on now! 1 gig for the OS and 99 gigs for porn.

Re:crazy

[realdpk]

The big plus in newer cards, from my point of view, is that you can now turn on full-scene anti-aliasing (FSAA) and still be able to play games. That wasn't the case a few years back.

FSAA can make scenes *beautiful*. Image quality is improving at the same time as framerate (not necessarily in the same platform though..)

Re:crazy

[platypus]

I own a Mac, so the situation is slightly different, ..

Yeah, Mac owners can use up all the resources of their machines with a fast tetris level ;).

Re:crazy

[hackstraw]

A cheap PC has enough RAM to store the entire contents of the Library of Congress many times over.

From http://www.loc.gov/about/:

It is also the largest library in the world, with more than 126 million items on approximately 530 miles of bookshelves. The collections include nearly 19 million books, 2.6 million recordings, 12 million photographs, 4.8 million maps, and 56 million manuscripts.

If there is about a meg of info in each of the 19 million books, ....

Re:crazy

[niko9]

Is it just me, or has the computer industry gone a little too far with the whole "bigger is better/more-more-more"

I mean, currenty video cards render full color frames faster than the human eye can perceive

Currently, I am addicted to Desert Combat, which is a mod for Battlfiled 1942. I consructed a top of the line gaming system just for this game. 3.0Ghz Pentium and Tyan 9700 Pro, 1 Gig 'o ram.

The game is buttery smooth most of the time, even with AA and AF on and at a resolution of 1600x1200.

But, the producers of the mod have a habit of updating the in game textures with every release, which makes the game look even more gorgeous, but will eventually start to tax my system. Not only that, Battlefiled supports upto 64 players, and with the advent of broadband in almost every home, don't be surprised to see games with a max of 128 players. So asking your card to render moving players, tanks shooting shells, blood and gore, lets not forget the choppers and amphibious assault vehicles, your "I mean, currenty video cards render full color frames faster than the human eye can perceive" seems irrelevant.

When you see so called "gaming benchmarks", please remember they are on avergae. They rarely point out the times when there are so many things being rendered, that you will see a bit of a slowdown, even with top of the line cards.

Re:crazy

[GoofyBoy]

1. Games. Falls into new and shiny anti-aliased toy department.
2. Multimedia. Encoding and editing stuff still can make a powerful computer cry.
3. Compiling. Those things can never be too fast.
4. Poor programming. Some newer applications just are resource hogs and its partially due to crappy programming/design.
5. Never, ever underestimate the amount of pr0n which can be obtained. NEVER.

Re:crazy

[kasperd]

When will enough be, finally, enough?

Never. (Trivial answer, but true)

Re:crazy

[AvengerXP]

FSAA doesn't make scenes beautiful, it makes them blurry and slower. Sharpness > Jaggies.

Re:crazy

[Anonymous Coward]

Obviously the entire library has not been digitized but we can estimate:

19 Million books * 3MB = 57TB
2.6 Million Recordings * 100MB each = 260TB
12 Million Photographs * 2MB = 24TB
4.8 Million Maps * 15MB = 72TB
56 Million Manuscripts * 3MB = 168TB

Total - 581 TB

Those are really rough estimates with some pretty broad assumptions:
1) looking at my ebooks they are around 3MB each.
2) What does recordings mean and would they be wav FLAC ogg mp3. 100 sounded good to me as an average (some recordings may be whole albums - some may be just speeches).
3) 2MB is probably low for good-quality scans but who knows what size the photos are.
4) These would be along the same line as photos just bigger.
5) Manuscripts and books have to be around the same size?

Re:crazy

[keester]

I think you're at the wrong site. You were probably looking for a senior citizen's computer complaint board.

This is Slashdot. Round here, folks like bigger and better hardware and the possibilities they bring.

Re:crazy

[2nd Post!]

Hehe, but a PC solves none of my problems. A PC doesn't do Final Cut Express any faster, it can't handle my 2,200 iPhoto collection any better, it can't burn iDVD DVD-Rs faster, it can't run iMovie... so there :P

Re:crazy

[mikael_j]

At ~50 chars per line and ~40 lines per page and an average of 230 pages per book (low estimates) You'd have 50*40*230 ~= 460kB/book

Now add a 800x800 jpeg for each map, say 60kB/map.

Add also at least 800x800 for every photo, another 60kB/photo.

Even without counting the manuscripts you would then have:
19,000,000*460 + 12,000,000*60 + 4,800,000*60 ~= 9519531 MB just there.. I could be off in my calculations though.. :)

/Mikael

Re:crazy

[realdpk]

I dunno, when I compare a game with FSAA on vs off, I'm most impressed with the "on" version. They look way, way better. It's getting to the point that relatively inexpensive hardware can do it without letting the framerate drop to unacceptable levels (for me, 40FPS is the line).

Re:crazy

[z00z]

... as if anyone will ever need a 10GHz CPU. When will enough be, finally, enough?

We're still a long way away from there. One high-tech "leader" (driving so-called "innovation") once said:

"640K ought to be enough for anybody."

Needless to say, he's been proven wrong (and in many more ways than one).

People want to render perfect-quality movies in real time; to have robots that can compute with near human intelligence; to compile Linux from scratch in a fraction of a second (at least *I* want to do all three :)

Faster CPUs only open up more avenues. Don't be too constrained with today's applications.

Re:crazy

[dosius]

It's not just you. I find this move wasteful. I can remember when a 'nix would run in under a meg of RAM, when a GUI would run on a CGA or a Hercules (yes, I think even X), and when hard drive sizes were still in the double-digit-megabyte range.

That said I would like to see an OS that is relatively POSIX-esque, with X, running on a low-end 486 with 8MB RAM, on a stock VGA.

I'm sure it's out there.

-uso.

Re:crazy

[n.wegner]

>and any other programming language that runs by
>relying on a "runtime environment"

You mean like C and C++? That msvcrt.dll actually stands for something, you know. Microsoft Visual C Run-Time. IIRC Linux has glibc, Mac OS X probably has glibc too.

If you were to program a Win32 application without msvcrt, you'd probably find, say, your malloc (using GlobalAlloc directly) to be a bit slower, because msvcrt's malloc is a bit better optimized for common usage. That, and having a malloc implimentation per program would require more space, startup time, etc. than having a shared library.

Re:crazy

[AvengerXP]

Wrong.

http://www.widearea.co.uk/designer/anti.html

"But anti-aliasing is more than just making something slightly fuzzy so that you can't see the *jagged edges*: it's a way of fooling the eye into seeing straight lines and smooth curves where there are none."

Basically, it's a human-perception hack to make you believe everything is soft and rounded when it's not.

Re:crazy

[2nd Post!]

But that's not what I said :)

The original post asked, to the effect, of who would need all the GHz/GB that was being pumped out, and I gave an example that *I*, as a non pro consumer Mac owner needed all the GB and all the GHz I could get.

I never claimed I needed more resources than anybody else. I just claimed that I needed more than I currently had, and that I could easily use >> 2GB ram, >> 40GB hard disk, and >> 2GHz CPU.

Re:crazy

[Junks Jerzey]

You may not need a 10GHz CPU. I know I don't, at least, not yet. But as interfaces improve, as software becomes more complicated, hardware like this may well a requirement. After all, monitor images are still nothing like as detailed as a printed page, and even a very high-end monitor has only a fraction of the resolution that can be produced by a $100 printer. It's going to take serious hardware to acheive that kind of image clarity in real time. 10GHz CPUs may be just the beginning.

Maybe. But some technology is good enough, like terminals used to display flight information in airports and LCD games at Toys 'R Us. You could replace LCD games with much fancier alternatives, but they're cheap, low power, have a bit of charm, and people like them.

The general problem with PC technology is that it's too broad. We don't know what we want, so we expand in every direction. Now we have PCs to two pound heatsinks, four fans, 400W power supplies, and need giant, impossibly complex operating systems to run (in this regard, Linux is in the same category as Windows, especially once you factor everything on top of the kernel into the picture, like XFree86). In the end, we have boxes sitting on desks in insurance companies and corporate reception areas that have 80-bit bit FPUs (standard on the x86 since the 486), 64-bit data bases (standard on the x86 since the Pentium I), MMX instructions (standard on the x86 since then Pentium MMX), SSE2 vector units (standard on the x86 since the Pentium 4), and so on.

Re:crazy

[sketerpot]

And around here, you're also more likely to have people point out that your (this post's grandparent's poster's) sig is wrong: Boromir was the brother of Faramir, who became the Steward of Gondor.

Sorry if this is offtopic, but it's been nagging me.

Re:crazy

[jon3k]

Functionality costs bytes, its really that simple. There's plenty of stripped down *NIX distributions, Minix comes to mind right off the bat. Not sure if its still around to be honest with you. Check out http://www.distrowatch.com/

Re:crazy

[CrowScape]

And you are, in fact, correct, this is not a Mac vs. PC issue (but you specified "Mac" which is just asking for it here). Anyone in digital video editing needs as much power and resources as they can get their hands on, Mac or PC, especially since HD came along and screwed everything up again.

Re:crazy

[n.wegner]

Kind of off-topic, but seeing "impossibly complex" and X-Window in the same sentence reminds me of those one-line descriptions of XFree like "You need X like a fish needs a bicycle." Can someone please post a link to a comprehensive source for these? I think there are some for Emacs, among others, as well, so please post links to any similar set of quips if you've got them.

Re:crazy

[Thing 1]

Many others have replied to you, so I'll be brief: we need to constantly improve technology, making bigger-and-better systems, so that we will reach the singularity as fast as possible.

Why? Because 50 million people will die this year. They don't have to die; once we reach the singularity our computers will be able to solve problems faster than our humans (and the rate of problem-solving will increase, as those machines build stronger and faster successors), eliminating old age and disease and the need to work for a living.

Yes, a faster game may not seem like it's on the path to world salvation, but it's a necessary part of it.

Re:crazy

[Jeremy Erwin]

even Apple's new G5 models can only address 42 bits of memory (4 TB theoretically, limited substantially by the absence of 512 GB DIMMs)... But since he also speaks of 10 GHz CPU, perhaps he knows of a hidden stash of future technology.

Re:crazy

[yomegaman]

Excuse me? The sentence you quoted says exactly the same thing as the post to which you replied, so I don't see how you can say it was wrong.

Re:crazy

[mikael_j]

I was using low estimates to avoid having some slashbot bitch about how I'm overestimating the sizes of everything..

/Mikael

What is a DCC Application?

[Eros]

Okay, I know what a CAD program is but what is a DCC application?

Re:What is a DCC Application?

[Verteiron]

You know, for 3D accelerated direct client-to-client file transfers on IRC. Yeah. I can't wait for this card to be supported under Linux so I can view BitchX in all its full, 3D text glory (as seen in Hackers).

Oh wait.. Google says it stands for "digital content creation". Poo.

Re:What is a DCC Application?

[mopslik]

According to this article, DCC stands for Digital Content Creation.

What's the license?

[bizcoach]

Does anyone know how their drivers for Linux are licensed? Greetings, Norbert.

Re:What's the license?

[kasperd]

Does anyone know how their drivers for Linux are licensed?

The usual closed source stuff is the best I would hope for. Surely I would prefer a GPL driver, or at least just specifications released, so somebody else could write it. But I was afraid they might have gone the other way, and just release a closed source driver for Windows, and no specs. Luckily that seems not to be the case (yet). Still your question is important, much more important than so much other stuff, but unfortunately I don't know the answer. No matter how great the hardware is, it is not worth much without a proper driver. So the answer to your question really should have been in the resume.

Re:What's the license?

[be-fan]

The Linux drivers will be around for awhile. A lot of high-end companies (like ILM) run their graphics workstations on Linux/NVIDIA. Since NVIDIA is the only high-end graphics company with quality Linux drivers, they're pretty much the only option in the market.

Re:What's the license?

[localghost]

Binary only under a proprietary license. They're decent, though. On par with the Windows counterparts, and latest hardware tends to be supported. No (official) 2.6 support yet, though. There's a patch, but I haven't had much luck with it.

Re:What's the license?

[kasperd]

Since NVIDIA is the only high-end graphics company with quality Linux drivers, they're pretty much the only option in the market.

The quality is being questioned. NVIDIA drivers have been blamed for some kernel crashes. The lack of options is the major reasons we don't want to see those drivers go. But that is not necesarilly what counts for NVIDIA. How large a fraction of the NVIDIA cards sold are being used with Linux? That is important to NVIDIA.

Re:What's the license?

[be-fan]

In random configurations, yes, there are issues with NVIDIA's stability. In closed configurations, there is not. I've never had a problem on any of my systems with NVIDIA's drivers, so its quite possible to build a system that runs rock solid with NVIDIA's drivers. Companies that are using NVIDIA on Linux can use these configurations, so that's not a problem.

As for Linux's importance to NVIDIA, remember three things:

1) Being the only option for high-end 3D on Linux is a big boon to them. A lot of computer graphics companies are moving to Linux (ILM, as I mentioned, runs Linux on almost all of their graphics workstations). These users are a very profitable market segment for NVIDIA's high end hardware (like the Quadro this article is about).
2) Being the only option on Linux is great for publicity. Gaming geeks are at least aware of Linux, and many have fooled around with it. Running on Linux is nice PR towards this segment.
3) Maintaining these drivers doesn't cost NVIDIA much. Their driver architecture is highly modular, so being cross platform is relatively inexpensive.

Re:What's the license?

[kasperd]

In random configurations, yes, there are issues with NVIDIA's stability. In closed configurations, there is not.

That's strange, it is the first time I hear about random configurations. Most people I know use pretty deterministic configurations.

Re:What's the license?

[be-fan]

Ah, but you're forgetting about the hardware tunneling effect. In a closed, low-budget, IT system, hardware components will randomly tunnel from one machine to another :)

Workstation Class Cards

[Serapth]

I have never understood how this breed of cards exists to this day. Really... the difference between a "stock" GeForce and a workstation class Quadro GeForce... just doesnt justify the cost difference anymore.

When you go back about 3 or 4 years ago... when you contrasted a Oxygen video card, or a FireGL vs a TNT or 3DFX card, you could see where the extra money went. But now, todays commerical grade video cards are more then capable. In fact, alot of people I know that work as graphic artists, use traditional Radeon or GeForce 4's in their workstation machines. Outside of say... Pixar, I just dont understand people buying the workstation class cards.

Now, to go back to this arguement... its nice to see that nVidia managed to get rid of the vacumn cleaner sized fan!

Re:Workstation Class Cards

[GauteL]

Did you read the article? Some of the benchmarks show that the workstation cards do the job at a magnitude of 3 times faster in some test than the fastest gaming card from NVIDIA.

If you need this speed, then you pay for it. Paying say $2k for a video card is not much if it makes your employees more productive.

Re:Workstation Class Cards

[Serapth]

Yes, but realworld vs benchmarks just doesnt work in this regard. I can run Maya or Max ( probrably the two most common 3d apps today ), in OpenGL accelerated mode at full speed for the most part at 1600x1200, without any slowdown.

So really... whats 3 times faster then fast enough??? Also... "some" benchmarks... well... beyond texture throughput, and raw triangle blit speed, nothing really matters, in the current generation of 3d apps, anyways...

Re:Workstation Class Cards

[tinrobot]

Now, to go back to this arguement... its nice to see that nVidia managed to get rid of the vacumn cleaner sized fan!

What do you mean, the fan is huge on this thing. I've had one for the past month or so and it is a nice card, but man, it's a total hog. The card requires you plug it in to a hard drive power connector because it can't suck enough juice off the AGP slot. It also requires that the PCI slot adjacent to the AGP slot remain open for cooling.

I find that price is usually commensurate with the size of the card. I think there must be a secret law that any card that takes up two slots can't be sold for less than $1500.

Re:Workstation Class Cards

[CrowScape]

One word: Reliability. Professional workstation cards are orders of magnitudes less likely to crash on any demanding application that is being stressed out. You can "get by" with a gaming card if you're on a tight budget and aren't rendering scenes for the next Star Wars movie (although a gaming card was probably used in the first episode from the looks of it)

Re:Workstation Class Cards

[Serapth]

Have you seen the first generation's GeForce FX's fan? Believe it or not, the damned thing is actually alot smaller and more quite! ;)

Re:Workstation Class Cards

[Merlin_80000]

I would think that George Lucas, Pixar, and friends have something to do with that.

Re:Workstation Class Cards

[Jungle guy]

Do you really know what are you talking about? Workstation class cards are made for stability, game cards for performance or price/performance. See this review: using CAD applications with a game card will lead to crashes and freezes.

Comparing workstation graphic cards and games graphic cards is like saying you can put your white box PC on a high speed conection and use it as a server. Why do you think people still buy IBM servers with Penitum III when the Pentium 4 costs the same? Because they want stability, not only speed.

Re:Workstation Class Cards

[Homology]

I have never understood how this breed of cards exists to this day. Really... the difference between a "stock" GeForce and a workstation class Quadro GeForce... just doesnt justify the cost difference anymore.

Just take a model of some hundred thousand triangles, and render them as wireframe. The gaming card ("stock" as you call it) will just slow down to a crawl. Add a few clipping planes, and the frame rate gets even worse. But the Quadro cards does indeed cater this common usage in CAD, and charge accordingly.

Now, you may say it's pretty much the same hardware, and you might not be that far wrong. A couple of years ago you could "upgrade" a GeForce card to a Quadro card just by hacking some software. Don't think that is possible anymore.

Re:Workstation Class Cards

[afidel]

For a little perspective here's a quote from one of Nvidia's driver authors:

I use a Quadro FX 3000 with 256MB of VRAM, which is the absolute high end of NVIDIA's professional graphics series. It's basically a souped-up FX 5900 core manufactured for the CAD, DCC and visualization markets. There are many decent used vehicles that cost less than a Quadro FX 3000.

Hmmm. A car or a video card.... For me it's an easy question to answer, but for work the equation works out a bit differently. There the CAD engineers are paid well into six figures, the annual liscenses for the CAD apps are almost six figures, so anything that is going to increase those engineers productivity is WELL worth the cost. If you have an equipment budget of around a million a year for engineering PC's a couple thousand for a video card that will make the engineers more productive is a no brainer, their workstations already cost around $10k, what's $2k more?

Re:Workstation Class Cards

[Serapth]

Actually, the video card has absolutely *ZIP* to do with the actual rendering.

To my knowledge the only big difference ( in silicon ) between a Quadro and GeForce is that the Quadro had hardware for accelerating line rendering... and you used to be able to patch a GeForce with software to make it equivalent to a Quadro... and benchmarks would show the results... wish I had the link now, but it was called SoftQuadro I believe... havent used it myself though ( never had the need ), so this is all hearsay...

Re:Workstation Class Cards

[CrowScape]

I didn't say they had to do with rendering, but if you're trying to set up a scene with lots and lots of actors and then, heaven forbid, you do an OpenGL or even a wireframe fly through of it to preview it (which you do quite often in DCC) then you will run into a scinereo where a gaming card can crap out. What, you think the designers just set up the keyframes and then let it render for the next several hours (or days) tying up all their system's resources and hope it comes out right?

Oh, and with hardware rendering in Maya, graphics cards are now able to have SIGNIFICANT impact on actual rendering.

Re:Workstation Class Cards

[Serapth]

When did Maya gain that ability? I havent seen it in action since version 4, so it may be a new feature. If thats the case, thats very cool...

As to setting up a large scale scene, your right, you need a meatier machine to render more complex scenes ( meaning render to screen as opposed to render to file... ). However, I would argue the gain you would get from moving from a highend consumer GeForce to a Quadro, would be negligable. I would say spending that money on more RAM, or a better CPU ( such as an Intel with HyperThreading or in the future, and Opteron ) would be money better spent. The gain is just way too minimal to justify the cost difference. We did a test inhouse comparing the GeForce to the Quadro, for use this Catia and ProEngineer at my work place, and saw no real functional difference... perhaps your millage may vary. Btw... when we are talking about IGES files, we are talking about ULTRA dense meshes in this case... well in excess of a million polygons.

Then again, if money is no object, I suppose any gain is a gain, right? Im not saying that the Quadro isnt faster... it probrably is. I just dont think its enough faster for *most* people in the industry to justify the additional expense. That said... I would love to have one! ;)

Re:Workstation Class Cards

[NanoGator]

"Actually, the video card has absolutely *ZIP* to do with the actual rendering."

No, the video card doesn't provide any bits that get put into the final image. However, a very significant portion of rendering is the setting up of the scene. You can spend hours upon hours getting your animation just right for render. That's where the video cards help 3D animators with productivity.

By now you've probably already heard something to this effect heh.

Re:Workstation Class Cards

[CrowScape]

I'm not arguing for speed, but reliability in this case. You don't want your video card booting you out of Maya or locking up your system because it can't begin to handel the textures or polycount (DCC and CAD certifications do exist for a reason). Right now, the Quadro FX 3000 IS only for those people for whom price is no object. The Quadro FX 1000, at around $850, or the Quadro FX 500, at $400, are practicle buys. I'd like to think when an FX 4000 is released the FX 2000 will become practicle as well.

Re:Workstation Class Cards

[wfberg]

I have never understood how this breed of cards exists to this day. Really... the difference between a "stock" GeForce and a workstation class Quadro GeForce... just doesnt justify the cost difference anymore.

Well, a specialist model requires retooling of the manufacturing process, which costs a fixed amount of money, and you sell fewer of them, so the added cost per unit produced is higher than you'd expect from the extra features to be gotten.

Of course, if nvidia would churn out Quadro models by default, the price per unit would only go up by a few bucks and everyone would have a gee-whiz superfast card. But that would be Communism. So what they do is a thing called price differentiation. Essentially both the consumer and pro market are buying the same product, but the pro's recon they can't live without some of the bells and whistles, and that they have no problem spending some extra cash, so they buy the pro version at a higher cost. This way everybody pays the most they can afford.

Price differentiation is also the driving force behind mail-in rebates; if you don't care too much about the cost you're not likely to mail in for the rebate cash, so everyone pays roughly what they were planning to spend (especially if the rebate undercuts a competitors price; since you don't expect to reimburse it 100% of the time you can offer your product at a price-after-rebate that is suspiciously low). Some companies take this a bit too far, and you can only succesfully claim the rebate if your time has no value at all and you are prepared to spend weeks jumping through hoops and reading fine print.

Re:Workstation Class Cards

[Fulg]

...so anything that is going to increase those engineers productivity is WELL worth the cost.

Not to mention the support you get from the CAD/DCC vendor.

For most CAD/DCC apps, the IHV has to go through a very rigorous certification procedure, ensuring that professional users don't have to deal with crappy drivers. Everything just works... I've been on the receiving end of some of these certifications (writing an OpenGL driver) and let me tell you, those app vendors don't fool around. There is a very narrow margin for driver errors to slip through.

Most consumers would rather (understandably) have a cheaper card, but if I was doing any professional CAD work I wouldn't tolerate anything less. You're paying for a lot more than just the name...

BTW that quote is just too cool to pass up. Welcome to my new .sig :)

Re:Workstation Class Cards

[Taos]

"No, the video card doesn't provide any bits that get put into the final image."

That's changing. Mental Ray is supposedly working on it. Larry Gritz's company, Exulna was bought out by NVidia after Pixar sued them into oblivion for patent infringement. Hmm... wonder what NVidia wants a top of the line Renderman complient renderer for? (Yes, they are working on it)

There are some concerns with the concept I have, but I get the chance to ask them Tuesday.

Re:Workstation Class Cards

[NanoGator]

"Mental Ray is supposedly working on it. Larry Gritz's company, Exulna was bought out by NVidia after Pixar sued them into oblivion for patent infringement. Hmm... wonder what NVidia wants a top of the line Renderman complient renderer for? (Yes, they are working on it)"

That would be pretty cool. Imagine buying multiple 'renderer' cards to put in a machine. Imagine being able to use your computer for other stuff while the cards are rendering. That's one of the reasons I work on dual proc machines. ;)

"There are some concerns with the concept I have, but I get the chance to ask them Tuesday. "

I dind't understand this comment, clarify pls?

Re:Workstation Class Cards

[Anonymous Coward]

It depends on your application.

I have a GeForce Ti4200. One of my mechanical engineers has a Quadro 4. Autodesk Inventor is *much* faster on his system. The GF is good enough for me, because I'm the PHB and only fire up Inventor now and then. It's not good enough for him.

I'm ordering a new system for one of the other engineers. I'm leaning towards the Quadro FX 1000. The FX 500 might be good enough, and the FX 2000 and FX 3000 are overkill for the work that we do. But I can see where other people have work that justifies spending for the fastest cards.

Re:Workstation Class Cards

[Taos]

>>There are some concerns with the concept I have, but I get the chance to ask them Tuesday.

>I dind't understand this comment, clarify pls?

I get to go to a demo of the NVidia renderer next week. Grill them on some issues that I think might crop up when using this stuff in production.

damnit.

[TrippTDF]

Why is it that I have have to pay for for a video card alone that I have to pay for an entire gaming console?

I know PC gamers are very die-hard (I'm a little more casual than most), but do you ever stop and look at the damn COST of the hobby? It's ridiculously expensive.

I can understand why people want to have the best boxes out there to play UT2003, but do you ever question how much you are spending? You've got to stop and think about it sometimes.

I'm just figuring off the top of my head, but I'd bet you'd have to spend an average of $250 a month on hardware alone to keep from being obsolete.

Any hardcore gamers like to give their opinions? I'm very curious why people would drop SO MUCH money for cards like this...

Re:damnit.

[frkiii]

(Slaps forehead)

No wonder I have three computer systems in my house!

1) 1.1 GHZ, 500 MB RAM
2) 2.2 GHZ, 1.5 GB RAM
3) 4.4 GHZ, 1.0 GB RAM

The funny thing is, I have never been into console games. Reason is, I also do a lot of net research, computer graphics, coding JavaScript (C and Java). System number 3 is my main gaming system, I use my others for my research and other work. Hence, I have tended to "accumulate" systems since the early 1990's. Not just for upgrading for games, but also for faster compiles, better handling and processing of larger higher-resolution graphics files, etc.

In looking at this card though, I wouldn't touch it. I can get more than adequate graphics cards at the local hardware vendors I frequent, for very reasonable prices.

Regards,

Fredrick

Re:damnit.

[DeathPenguin]

>>Any hardcore gamers like to give their opinions? I'm very curious why people would drop SO MUCH money for cards like this...

RTFA, this is not a gaming card, it's a workstation card and the drivers are tweaked for such apps as 3DSMax and Photoshop rather than Quake and UT. It's NOT for the average gamer and it's NOT for the average desktop user. It's for a very niche market, nothing more. That question is sort of like asking why people drop $10k+ into an SGI workstation that is clocked lower and performs worse in everyday apps than white box x86 computers.

Now, why nVidia jacked up the cost of the Quadro when they've historically been known for solid-performing value cards in this segment (As opposed to exotic multi-GPU solutions from 3DLabs, for example), I don't know.

Re:damnit.

[afidel]

I personally never pay much more than $100 for a video card. I look at what's available at that price point and decide if it fits my needs, every time I've looked it has =) Besides I buy games that keep me interested for years instead of months like most console offerings. I have played Moo3, Diablo2, Diablo, Civ, MoM, HoMM3, and several other computer games many years after their release, Soul Caliber and Dragon Warrior are the only console titles I can say that about. My annual outlay for my PC including games and hardware upgrades is still less than what my friends spend on consoles and games despite the fact that I use my PC for work, education, and other leisure activities besides just playing games on it (for instance I VJ at local underground parties). Besides this is NOT a game card, it's a CAD/prerendering card.

Re:damnit.

[ShieldW0lf]

Why is it that I have to pay more for a video card alone that I have to pay for an entire gaming console?

Yeah, but the games for a console are expensive, while games for the PC are free!

Re:damnit.

[bfischer]

4.4GHZ? What kind of processor is that?

Re:damnit.

[frkiii]

Ugh, my mistake, dunno why I had 4.4 GHZ on the brain.

It is a Pentium 4, only 3.2 GHZ CPU. Sorry for the error.

Wasn't near my home system when I wrote my post, and I believe I had read an article about a 4.4 GHZ Intel and AMD chip being worked on (not released yet) within the last few days. Lack of sleep, lack of coffee and new EverQuest expansion woes on the brain, definitely on the brain.

Regards,

Fredrick

$3k? When $500 is almost as good?

[FileNotFound]

Having RTFA I am surprised that they liked the card.

I mean how can you say that the Quadro is a good card when it costs 6 times more than the competition and is less than 10% faster?

6 times for for 10% faster? No thanks.

Re:$3k? When $500 is almost as good?

[stratjakt]

People like Macs for the same reason.

Call them fans/zealots/cheerleaders, whatever.

I'm sure someone will tell you how super-fantastic nVidia's drivers are and how mega-gay ATI's are. (In my experience it's been exactly the opposite, but such is life)

Just like someone will no doubt tell me how rootin' tootin' fantastic OSX is, and how it's worth a $2000 difference in price.

Re:$3k? When $500 is almost as good?

[FileNotFound]

Erm...since when is there a $2k difference betwen Macs and PCs.

More to the point, it's not the same ratio. A Mac is not 6 times more than a PC.

Re:$3k? When $500 is almost as good?

[BurritoJ]

Read beyond the first page. On some tests it did twice as well. Particularly on the ProEngineer benchmark. If you've ever tried to spin a model and endured it herky-jerk it's way into position, you know how much a video card is really worth.

Joe

Re:$3k? When $500 is almost as good?

[Merlin_80000]

from what I read, the $3k card is designed for render farms, if I understood correctly (i'm not 100% that i did, but i at least understood some of it)the cards talk directly to each other for fast mutli-machine rendering.

This is an enhanced version of the QFX3K, with a daughter card that sits on top of the board, mezzanine style. This board provides such functionalities as "Framesynch" multi-system synchronization and "Genlock" for synchronization to an external source input. The Framesynch connections are the sockets that look similar to Ethernet jacks. These provide the ability to have multiple systems working on a single model and rendering contiguous pieces of it, for example, while displayed at the same time on a larger screen format. The Genlock connector is the coaxial plug on the card. However, we're looking at the standard version of the Quadro FX 3000 here for our test purposes, so that's what we'll be focusing on in this showcase.

that means that when Pixar /AAA Architecture and Design decides its time to finish up with their rendering project, it cuts render time down significantly, and since there are so few new cards out there that do exactly this, yea the price is going to be steep. and hell this is the kinda marketing that empires are built on. in order to get a real return on the investment, you need to buy more than one (let's say 32) and then the fastest agp8x machine you can get your hands on, and don't forget software licenses. so even though the card is $3k, to use the card well, you need to spend around $250K, $96K of which goes to nvidia...this is all top-of the head/google calculator math, but i'm just trying to get a general idea across....this card isn't for gamers, like at all....they probably want gamers NOT to buy this card....this is for the movies and the designers/architects.

Re:$3k? When $500 is almost as good?

[JungleBoy]

There IS a reason, but its a high cost setup. The Quadro FX 3000 can do Dual DVI at 1600x1200 on each screen. Very few cards can do dual DVI at this resolution. In fact the only other one I know of is the Quadro4 XGL 900 (which I use to drive my viewsonic vp201mb LCDs). Many high end game cards can now do insane dual monitor resolutions, but only dual analog is supported; or if they do support dual dvi, they only support it at 1280x1024. I think the Matrox Parahelia can do Dual DVI @ 1600x1200, but I bet UT2k3 doesn't run as well as on my Quadro. And with 3200 pixels of width, you can set a nice high field of view.

Re:$3k? When $500 is almost as good?

[gmhowell]

Don't worry. Sooner or later, somebody will post about how many bazillions of seconds that cuts of render time at his company, and how that saved gazillions of dollars per year.

Ignoring the fact that his company would get better productivity gains if he weren't screwing around on /. on company time.

Re:$3k? When $500 is almost as good?

[SquadBoy]

The software this card is supposed to be used for runs into the 10s of thousands for one seat. So yes that performance increase is well worth it.

Re:$3k? When $500 is almost as good?

[jpc]

no, render farms dont use graphics cards. the genlock is for when you have multiple machines rendering the same image eg on multiple projectors

Re:$3k? When $500 is almost as good?

[pipingguy]

There IS a reason, but its a high cost setup. The Quadro FX 3000 can do Dual DVI at 1600x1200 on each screen.

Are CAD and DCC users that have two monitors going to be using LCDs instead of CRTs? Perhaps LCDs are much improved now, but I remember "color fidelity" issues and response time being considerations.

Re:$3k? When $500 is almost as good?

[mczak]

Pretty much all cards with dvi connectors can do 1600x1200 per connector nowadays, though it should be mentioned that all consumer ati radeon cards only have one dvi connector. However, all newer Matrox cards (Parhelia, P650, P750) as well as some nvidia consumer cards, including dirt cheap GeForce4MX (not all of them have two dvi ports of course, but some do) can easily do 1600x1200 with both dvi connectors.
However, the FX2000 (and FX3000, but no older nvidia workstation graphic cards) indeed have an advantage in that area - all other cards have dvi ports which can do a maximum of 1600x1200 (at 60Hz), because they all use so-called single-link TMDS transmitters. The FX2000 and FX3000 however have one single-link and one dual-link port (thus 1600x1200 on one monitor but 2048x1536 on the second monitor).

Re:$3k? When $500 is almost as good?

[jafuser]

Ok maybe I'm a noob, but what good is a FPS game with two monitors? Wouldn't your view be centered where the two montitors meet?

It would seem to me like only an odd number of monitors would be good for playing games...

The Police Quest test

[lateralus_1024]

One has to wonder how well it can run Police Quest. Time will tell. I'm sure NVidia is dodging this test.

Re:The Police Quest test

[sn0wcrash]

Please.. the real "hardcore" test is Leisure Suit Larry.

Re:Dual Out

[shish]

4000 x 25000? now *that's* widescreen!

(Or am I showing ignorance? is this really not a typo?)

Re:Dual Out

[niko9]

Dual DVI output: Drives two independent digital displays at 1600 x1200, or one at 3840x24005.

Nice.

Sounds like the perfect card for the perfect monitor

9.2 million pixels at 0.1245mm stripe pitch, 3840x2400 native resolution.

Now, that my friends, is nice.

Re:Dual Out

[pheared]

Yeah I think it must be. ;-)

What about Half-life 2?

[az]

The review is good and well, but what framerate do I get when i run Half-Life 2?

Re:What about Half-life 2?

[stud9920]

Approximately 40 spf

Oh, yeah...

[fuzzix]

I like it a lot!

Re:Oh, yeah...

[fuzzix]

I mean, it must do at least 36,000,000 hentai a second.
You guys sure it's not called The Pr0n Graphics Quadro FX 3000?

Re:Dual Out

[FileNotFound]

Actually that'd be tall screen or something.

The first value is the horizontal.

Yeah, but it's faster!!

[spineboy]

Another way of looking at it.
If the competition card is rated at a spped of 10, then this card goes to

11 !!!

The movie was Spinal Tap, for all those wet behind the ear kiddies..

Re:Yeah, but it's faster!!

[FileNotFound]

Thing is, this is the business market.

It's often ok for home users to pay $400 for a card that's only marginaly faster than a $150 card because they want "The fastest card in the world" in their PC.

This is not the case in the business world. Nobody cares if your employees have "the fasterst card in the world" in their PCs. They care if you have "the best ROI in the world!"

Re:Yeah, but it's faster!!

[GigsVT]

In the business world, $2000 buys you a whole lot more support than $500.

Yeah, but if you need support on a video card, then someone fucked up somewhere big time. That's one of those things that's supposed to "just work".

Re:Yeah, but it's faster!!

[DrJohnnie]

Ok... ROI - At 10% faster....

The burden rate for our CAD designer is ~$60/hour
(We use Pro/ENGINEER - could take advantage of the card)

2000 Hour/year x $60/hour x 10% = $12,000

20% of the designer's time would have to be spent waiting on the graphics to pay for the card.

Re:Yeah, but it's faster!!

[afidel]

Yeah, but if you need support on a video card, then someone fucked up somewhere big time. That's one of those things that's supposed to "just work".

Bzzzzt.
Not only do CAD companies dictate a short list of supported graphics cards but they often dictate a specific set of one or two driver revisions that they support per application revision. Trust me when you are spending near six figures for an annual liscense and paying a CAD engineer well over six figures you simply don't CARE what the card costs, if it is the best performing card on the supported list you will probably buy it.

Re:Yeah, but it's faster!!

[FileNotFound]

Personal opinion here, but I think that counting ROI based on such tiny savings in time doesnt' work.

For example, say that the designer has a slow as hell computer and waits 5 minutes per render. He uses that time to take a smoke/get a drink/bathroom break/cube chat/whatever. Now imagine that the render is instant.

The designer still takes the same ammount of breaks just without the excuse. Maybe a few less, but the result is hardly the exact ammount of time saved.

Re:Yeah, but it's faster!!

[GigsVT]

And they do that assumedly so they can work around any bugs in the driver, you don't need support from the video card manufacturer.

BTW- WTF is a "CAD engineer"? Is that like "sandwich artist"?

Re:Yeah, but it's faster!!

[afidel]

An engineer who specializes in computer aided design and manufacturing as apposed to say designing circuits, power supplies, or any of the other myriad of engineering positions a company may have. Basically at my last gig they had engineers whos job it was to take the specs and designs from the other engineers and work them into real products which could be efficiently manufactured given the limitations of the assembly plant. They specialized in process flow, materials, and some other stuff as apposed to other areas of engineering.

Re:Yeah, but it's faster!!

[DrJohnnie]

I agree, I was just throwing in some numbers in for an "ideal" case. The real cost saving over a year wouldn't cover the card.

I have found over the years with each upgrade the designer makes bigger (longer to render) models, and the software get bigger (OS, MCAD, PDM, ...) The ROI doesn't come from modeling faster, the ROI comes from better models. That is impossible to measure (and my not have any real ROI)

By the way... when I'm waiting for my model to render - I read slashdot ;)

BUT look at what you get for that chunk of change!

[jbottero]

A ***way cool*** heat sink and chip fan! Like out of a jep plane! Way cool dude!

Re:Impressive.

[hawkbug]

You spent 3 grand on a CAD card so your brother could play Dark Age of Camelot? I think not.

If you read the article....

[spineboy]

It's not a gaming card, but it's primarily for CAD/AM stuff. It's loosely based on an older gaming card, and that's why they decided on a whim to try out UT2003, etc.

Re:Dual Out

[shish]

Damn it, I *am* showing my ignorance. This is not my day. Wait a minute... /me checks flies

DAMN!

Now the *interesting* thing is...

[SoTuA]

...how does one justify spending that much in a Graphics Card when a "gaming" one gives almost the same performance at 1/6 the price?

Years ago, the optimizations where noticeable and a workstation card was worth it. With these days of insane GPUs and cards (I think my system memory is actually slower than my video card memory!) the optimizations don't look that impressive, and sure as hell aren't worth 6x more.

Impressive? No, especially the review.

[LxDengar]

Max 4.2 as a benchmark? Please, move up to 5.1 at least, and get XSI and Maya in there (Although admittedly Maya can be a troublesome beast for new graphic cards). Its a workstation card, run it against workstation apps.

I appreciate the effort with the SPECopc benchmarks, but review sites need to put more effort into testing a card like this. I cannot see what the reviewers were thinking by not putting this card up against what it's price point is aimed at - specifically Wildcat cards.

Re:Impressive? No, especially the review.

[afidel]

They likely couldn't afford the X thousand dollar liscense for Maya, let alone one for Autocad, Solidworks, etc. The best place to get info on what video card to purchase is from you software vendor. All of the CAD/DCC companies have a short list of recomended cards which are the ones they test the most with and which are best supported by their products. You're spending big cash on the software and the talent, through a little time and money at getting the right hardware.

Re:Impressive? No, especially the review.

[LxDengar]

While I understand what you're saying, XSI and Maya come in versions (a.k.a, free) that would be wonderful to run this card against. I tend to stay away from my software vendors and perfer the word of real life people when it comes to video card experiences. Vendors are... well, vendors. My point is, if you're going to take the time to review a $3k video card, Unreal 2k is not what I want to see.

Re:Dual Out

[shish]

> Last I checked my screen resolution didn't change the size of my monitor.

I was assuming some weird CAD hardware or something. What size monitors do CAD guys have for their 2000 x 3000 pixel screens anyway?

Re:Impressive.

[fuzzix]

I bought my brother a Cray X1 to play minesweeper. No biggie.

What need next

[r_j_prahad]

OK, graphics cards are finally where we want them to be for price-performance point. Now what I am needing is the same value in video projection systems, like 2000 lumens for $200. Because until you have experienced X-Plane in 8-by-10 foot format, you have not experienced X-Plane.

Porn would also be incredible experience in that scale, yes?

Re:What need next

[afidel]

Nah, almost all P0rn is shot with equipment like Sony Handicam's. Low budget rules the sleazy underworld, so projecting a lousy picture on a big screen just makes you realize how poor the quality is. It's even worse with Divx rips, ugh. But I have to agree with you, 2,000 lumens for even $500 would be great.

Re:Support companies that support OSS!

[wmaker]

Yes, actually it does support linux... OPERATING SYSTEMS Windows(R) XP (WHQL-certified) Windows 2000 (WHQL-certified) Windows NT(R) Windows 98, Windows 95 Linux--Full OpenGL implementation, complete with NVIDIA and ARB extensions (complete XFree 86 drivers) Professional CAD and DCC Certifications

Re:Support companies that support OSS!

[dosius]

IMHO they aren't "supporting" Linux unless they release source code for their driver.

-uso.

Ever use a mid to high range 3D CAD program?

[bombom]

I work as a CAD developer and you won't belive how much Quadro cards can outperform regular Geforce cards (don't get me strarted about ATI and thier crappy drivers for the FireGLs).

The single biggest limitation of the GEForce cards is they are optimized for 1 window. A Quadro card OTH can have open multiple windows in a 3D cad program . (e.g. Geforce 4's choke after having 3-4 moderatly complex parts open and a good Quadro 4 can handel 12-15 windows no sweat). AutoCAD is pretty light on the GFX card so a older quadro might suffice but something like Inventor or Solidworks or Catia can really benefit from these cards. If one of your engineers sees a 20% speed up in generating drawings of a complex assembly, or a 10% speed up rotating a gear assemlby, 3K is money very very well spent.

Results for Soft-Mod'd Radeon 9500

[Pup5]

I suppose that I can just assume that the 9800 results will be fairly similar, but I'd like to see numbers for my softmod'd Radeon 9500 by W1zzard

3DLabs?

[kyoko21]

Did these people forgot that 3DLabs is still around and still make some excellent cards for these CAD applications?

Re:Dual Out

[afidel]

21-24"
The Sony Trinitron CPD-G500 21" monitors can do 2048*1536 @ 75Hz. Their 24" widescreen does 2304 x 1440 @ 80Hz. Someone else may have a better monitor but I haven't seen em =)

Using graphics cards as high-speed processors

[Thagg]

At the latest Siggraph conference in July, there were quite a number of papers describings systems that used these new graphics cards in novel ways. There is a tremendous amount of horsepower available, but it is often not obvious how to apply that horsepower to your particular computation.

Graphics cards seem to be increasing performance on a faster curve than CPUs these days, so people hope that the effort made in restructuring programs to use these graphics cards will pay off big down the road a year or two.

The biggest problem with using these cards is that they are changing so fast the documentation and tutorials are woefully out of date. Still, the promise is so great it's often worth the effort.

thad

Re:yeah but at a refresh rate

[Jeremy Erwin]

but does an $8000 plasma screen have the resolution?

why *STILL* 2 slots?!

[andrewleung]

why does nVidia's BEST card STILL take up 2 slots in a machine while ATI AND 3DLabs can make a card with near performance using 1 slot?!

sad... sad. these guys are just mocking consumers that buy into their "marketting"... why do we even comment?!

Re:why *STILL* 2 slots?!

[paradesign]

you do realize that most people that use this card will never know what it is, other than tehe fact that it works. they could care less if it takes up 12 slots, cuse theyre never going to take the time to look at it.

this card is a tool for professionals, nothing else. professionals have IT departments to deal with all the crap associated with running a computer.

but, nice troll anyways.

Re:yeah but at a refresh rate

[Anonymous Coward]

We use IBM T221 displays at work in conjunction with Quadro video cards. The cards we have can drive the display at 25 Hz. Guess what? No migraines. Guess why? It's an LCD screen. The only thing the refresh rate affects is how fast the screen can be updated. So, you can have some problems with mouse pointer display lag, but you don't have flicker problems like you would on a CRT. So, it wouldn't be very good for full motion video, but it's fantastic for displaying static high resolution images.

Couldn't agree with you more...

[cpopin]

Yah, I looked at the high-end cards and looked at what software I would be using that would push it, mostly popular first-person-shooter games, and decided on the NVidia GeForce4 Ti 4200 128MB. It's the most bang for the buck. Hell, running on a P4 2.0 GHz with 1GB RAM, any game would be happy.

Re: (ACTUALLY!)Workstation Class Cards

[voxel]

There is a difference. If you've ever programmed OpenGL or DirectX and played with rendering settings, you'll notice rendering in Wire Frame mode is *extremely* slow on a consumer card, but blindingly fast on a Pro card like the Quadra.

3D modelers spend alot of time in this mode, so it only makes sense to spend extra transistors on this area versus in the consumer market.

Re:Imagine...

[BiggerIsBetter]

FPGA

for Huge resolutions, Sony GDM-F series, QUXGA

[spage]

Sony's own GDM-F series is far superior to the CPD-G series because of its finer dot pitch (GDM-F520 0.22mm aperture grill vs. 0.24mm for G- 520) and higher maximum refresh rate. Yes you can get 2048x1536 out of the G series but you'll have fuzzy dots.

As I wrote in my Guide for size queens, find a PHB with an F series monitor and do the midnight monitor swap with a generic 21-inch monitor. S/he'll never notice on her/his 1024x768 default desktop :o)

I have so-so eyesight and run my GDM-F400 at 1600x1200 and GDM-F500 at 1800x1440 or higher. It takes a lot of display appearance tweaking, Mozilla Ctrl-mousewheel to zoom sites, Apple [Cmd-shift-+] to enlarge apps, and bitching to sites and apps that don't scale, but it's well worth it.

[IBM T221 LCD] at 9.2 million pixels at 0.1245mm stripe pitch, 3840x2400

Yes the Quadro cards are designed to drive the IBM and Viewsonic "Quad Ultra XGA" LCD panels, but it appears you have to use an interleaving technique which reduces the refresh rate substantially. As I understand it, at that resolution you're pumping out more pixels than the DVI spec is capable, even with dual TMDS transmitters.

3840x2400 still sounds damn intriguing, but it's impossible to get a demo in the real world. Maybe Apple's rumored next-generation Cinema display will bring ultra-ultra-resolutions to end-users who don't have an assigned workstation sales rep.

Re:yeah but at a refresh rate

[Metalbunny]

The refresh rate does have an effect on how fast the card can spit out frames ... or atleast it used to with the older ones ... Now, the quadro is for office use, so running on TFT screens doesn't really matter ... but for the avid gamer, they're still too far behind the cards to be of any real use ... Pounding out 240 fps in good old Unreal Tournament will make it unviewable on most TFT screens ... even the really good ones... they simply can't refresh fast enough to keep up. That said ... I've yet to actually see a TFT running of the DVI port, so this is based on the analog input ones ... the DVI ones are supposed to be able refresh faster and a bit more clearly...

Re:Dual Out

[AbRASiON]

IIyama make 22" monitors with 141khz horizontal bandwidth.

That equates to 2048x1536x85hz and something ludicrous at 1600 (120+ or so?)

Very very nice monitors, currently the best consumer level priced monitors out there.

bold[ and ONLY bloody made for the northern hemisphere so frigging useless to me in Australia............. sigh ]bold

Wrong, wrong, wrong

[theskov]

The maximum horisontal refresh rate is 67 KHz, but that's not the interesting figure. The interesting figure is maximum vertical refresh rate, which is 41 when running at max resolution.

And no it doesn't give you a headache, but still makes the monitor unusable for games. It's even slow enough that the cursor updates notably more sluggish, which makes it less suited for CAD and the likes.