Now days applications have evolved a lot. Maya, max run great on our normal gaming cards. Instantaneous renderers like vray, mental ray etc,. are directly dependent on computational performance (in other words number of cores). They don't bother to check the card is professional or not.
I happened to work on a workstation carrying quadro fx 4800 which alone costs ~900$ and I didn't find any difference. I don't know why they are over hyped.
you obviously have not tried to work with a large scene/assembly, the difference between a pro card and a gaming card becomes very apparent when you do. Most CGI folks only spot render on the local machine, all production renderings are sent to the farm so for day to day working a pro card pays for itself, unless your just working on small stuff, at which point whats the worth of using Maya/Max......
I paid a very high price for a pro card once. I presently do not see myself doing that again.
I'm not a pro, however, my opinion on this is that unless you are doing serious work in the pro field, pro cards have little value over consumer cards. Even if you are not a pro, one solution to rendering large models is to take a hike while it renders. I often did exactly that. On a large or potentially long-running job, I let the computer run while I did other things.
[citation][nom]SteveLord[/nom]The people that have no idea what these types of cards are designed for always make me laugh.[/citation]
Same here... What do you mean its slower than the Radeon 7990, its a Pro card isnt it?
I could see this being quite useful for entry level Hyper-V with RemoteFX as i paid a bit more for a v4800 and v5800 in a couple of situations and it definitely helped virtualized performance/end user experience
[citation][nom]ProStuff[/nom]@lockhrt999you obviously have not tried to work with a large scene/assembly, the difference between a pro card and a gaming card becomes very apparent when you do. Most CGI folks only spot render on the local machine, all production renderings are sent to the farm so for day to day working a pro card pays for itself, unless your just working on small stuff, at which point whats the worth of using Maya/Max......[/citation]
Actually I have worked on large(as heavy as few hundreds of MBs and sims of multiple of TB) scenarios and still I couldn't find that much of the difference.
I agree that with pro cards textures look more accurate in hardware view. But then it doesn't bring much to the table. It doesn't accelerate rest of the things. For Maya, max except rendering rest of the tools are single or double threaded.
Seriously autodesk need to get their a$$es kicked.
Some other applications like Vray, Houdini deliberately use graphics cards for specific use. But they score more
with gaming cards as they do have more CUDA/stream cores than pro cards. Here more means performance escalation in 3-4 folds. If I have 590 I get get 4x more performance than qaudro fx 5800 and still save my limbs.
The difference is antialiasing, precision, better color and contrast, virtually "0" artifacts and faster work on heavy polygon scenes and on viewport 2.0, at least in Maya. The Quadro looks nice with no artifacts but the GeForce looks rough with slower menus and windows redraw. The geForce cards lack precision and when the scene gets very complex with a lot of curves, particles, textures and geometry you want this precision working for your eyes.
An entry level Quadro 600 costs $150-170 and works perfect so there is no excuse to use an expensive $400-600 gaming card for professional applications and you can use both (Pro & gaming cards on W7) on the same machine. This is what I am doing and also my experience over 2-3 years.
If you see no difference with a Quadro is because you need to adjust the settings on the Nvidia control panel. Also Nvidia develops special drivers for every major Pro app so please don't say there is no difference with all that dedicated engineering and programing going on.
I'm on the fence these days about Pro cards, but not about this class of pro card. They're very much not worth the money on the low-end, and diminishing returns as you approach the high-end. The Quadro 6K costs roughly half its namesake, and for what? An underclocked gaming card GPU with more RAM. Why not just make RAM modular and let us add our own?
Jecastej, I've heard similar remarks from a few people (BoostAbuse, a few other professionals in the industry) about the pro cards, but have yet to see one outperform a Geforce with the overrides enabled, in the Nvidia control panel. I've clocked most of them with Everest now, as well. Nvidia doesn't have drivers for special apps, however - you're just plain wrong there. They don't have drivers for Maya specifically, nor for Rhino, Mudbox, Silo, etc. They may have driver profiles, but that's not the same thing. Those are just override profiles you can access already in nvcpl.
The problem is that it's simply hardware locking from the driver end. The Quadros sport the exact same chips as the Geforces - but you pay the premium for better QC and also for professional-grade support and replacement should one fail. So it's justifiable if your product fails or you have problems, sure.
Maya is such a specialized app and with so many, many variables and issues with the application itself (it's my bread and butter), it's almost impossible to actually diagnose how a card will perform unless you use it, tweak it, try various drivers (the Microsoft-suggested or Nvidia-suggested or Autodesk-suggested ones are never, ever the best ones; it's always the slightly older ones).
I'm pushing 20M polys at 15fps with a GTS250 in Maya, and almost twice that with a GTX460. There's simply no reason to purchase a FireGL or Quadro at this price point for Maya. I spent a year chugging along on a 380FX and it was just pathetic; the GTS250 demolished it and was an excellent replacement, even if it is just a retooled 8800GT.
I want to open up a full assembly model of a car with all the threads on every screw shown and maximum image quality opened into Solidworks 2011 and see if I can at least rotate it with a framerate greater than 1fps with this new FirePro V3900 lol would be fun to see
[citation][nom]juncture[/nom]I want to open up a full assembly model of a car with all the threads on every screw shown and maximum image quality opened into Solidworks 2011 and see if I can at least rotate it with a framerate greater than 1fps with this new FirePro V3900 lol would be fun to see[/citation]
Tell me, which graphics card you have used can do such a thing? It's actually poor implementation all around if the viewport can't handle the LOD. SolidWorks does pretty well. Other apps do much better. But even so...
...who decides what "hi resolution" is? If a screw (to use your shitty example) is triangulated at viewport-rendertime, on a decent 22" or larger screen, how many pixels of the screw are being rendered? If the answer is less than half a pixel (which it is, period, if you're viewing the entire car on any modern monitor, period) then it needn't be rendered at all.
I can answer your non-question idiotic statement: no, it cannot. And before you say, "My Radeon/Geforce can!" let me answer to you: that's because you suck at modeling. Precision modeling (real precision) for CAD/CAM has nothing to do with viewport display except for error-checking and machine prep.
It's like you just dropped the word "Solidworks" expecting people to think you're cool. You're bragging about something you had no hand in creating, and "dissing" something you know full well can't do what you're bragging about.
[citation][nom]lockhrt999[/nom]Now days applications have evolved a lot. Maya, max run great on our normal gaming cards. Instantaneous renderers like vray, mental ray etc,. are directly dependent on computational performance (in other words number of cores). They don't bother to check the card is professional or not. I happened to work on a workstation carrying quadro fx 4800 which alone costs ~900$ and I didn't find any difference. I don't know why they are over hyped.[/citation]
Instantaneous renderers? Number of cores only? It's blatantly obvious you know nothing about rendering. Just like the dude above dropping "Solidworks" like he's got a pair, you dropped the word Maya pretending to know the topic. Except you did fumble your way into being correct: software renderers don't care about your GPU hardware. But VrayRT, iRay, and the other GPU renderers care. And the Quadro 4800 is simply an overspec'd and underclocked GTX280 which listed at $1,999 originally, in 2008 when it came out. It's a decent card compared to gaming card because it is a gaming card.
MY point is this: this card (article's topic) sucks, alright. But it sucks because at this price point the gaming cards demolish it for performance. Comparing it to a $2K card is absurd, son.
Solidworks will do LOD for the display. A better example would be moving a two million polygon object in Maya in real time, with no LOD generation done yet. Moving the THG logo object around in Maya is an exercise in frustration...
Using a pro card has more to do with supply chain and support than anything else. These debates about performance are an irrelevant issue. If you are OEMing a card, especially for a medical product, a pro card is the only choice. Forget about the other stuff.