What Quadro outperforms geforce in apps

I'll use the Geforce gtx 460m for example.

I'm trying to figure out what quadro video card will perform just as well (or close too) as the 460m when it comes to applications, NOT GAMING (this is for my bro, he Don't care for FPS in games or gaming at all).

Since quadros are better for apps, ex

Quadro FX 580 vs GeForce GTX 285 in Ascon Kompas 3D on Vimeo
YouTube - Quadro 4000 vs GeForce 8800GTX ( After Effects )

I'm trying to determine which quadro card = gtx 460m when it comes to apps, then I can determine which card will be best suited for him.

 

As of right now, the only card in the Quadros line that is Fermi-derived is the 5000M and that sucker is equivalent to the GTX 470M and it is NOT cheap, since it is only available on the flagship Dell Precision M6500 and the HP Elitebook 8740w and the options with the 5000M can hit over $3k FAST.

That being said, it is rumored that the successor units for the M6500 and the 8740w will be carrying cheaper options like the 3000M, which should be on par with the GTX 460M, since the Quadros 5010M is replacing the 5000M as the flagship and that will be based off of the GTX 485M. The 4000M should be based off of the GTX 480M and the 3000M based off of the GTX 460M.

NVIDIA Quadro 3000M - Notebookcheck.net Tech

 

This can actually depends on the app as much as the card. So it'd be nice to know exactly which softwares he'll be using it with. Using it for CUDA? Or for CAD applications?

 

Actually, the 5000M is much faster than the 470M, but a bit slower than the 480M. It has 320 SP vs. 288 and 352 respectively. The reason why it's much faster than the 470M though, is also since it shares the same bus as the 480M (256-bit) vs. the 470M's 192-bit.

But yes, in new systems, the 5010M will replace the 5000M with a 384 SP part (which is the same as the GTX 485M), while the lower end parts have no equivalent as of now.

The 4000M is a 256-bit 336 SP part with GDDR5 (yes, above even the 5000M) and the 3000M is a 256-bit 240 SP part with GDDR5 (a bit below the 5000M).

While the mainstream 2000M is a 128-bit 192 SP part with GDDR3 (which is a bit above the 445M w/ 144 SP) and the 1000M is a 128-bit 96 SP part with GDDR3 (the latter shares many similarities to the GT 525M/540M/550M line, which is similar to the GT 420M/425M/435M line, all have the same 96 SP).

 

he will be using 3ds max, maya, corel paint, and some type of video editing software like cs5 and something for special effects.

Since he now has a good camera, he will probably be doing a lot more video production from now on.

The Quadro 5000m is too much on the expensive side. So I doubt a purchase with the new fermi technology will work (just learned about it, thx).

So I dunno, I guess I need to know of a quadro that won't shoot the laptop price to anything above $2600... maybe $3000

 

Unless you want to wait for the next-generation of corporate workstations to come out, you can get a Quadros-equipped system with at least a FX 2800M (GTS 360M) or a FX 3800M (GTX 260M) for around that.

 

If he needs the OpenGL oomph of a Quadro, then any Quadro will outperform your GeForce by a wide margin. If he doesn't use much of that, then there is no advantage to a quadro.

 

Well you do make the right point on Quadros trumping GeForces on OpenGL Rendering but not ALL Quadros can kick ALL GeForces. A FX 880M can't pwn a GTX 485M in OpenGL as the FX 880M is a lower-end card in hardware terms vs. a top-of-the-line powerhouse.

That said, I am sure a FX 3800M can still hold its feet against the 485M in OpenGL.

 

thx a lot StrikeSaber.

+ rep

The FX 2800M is looking nice but well see how much he wants to put into it.

What about Ati? ex, the FirePro M7820? Its much cheaper for some reason (seriously, whats with the price difference?). It seems to be better for gaming but how does it compare to say the.... Quadro FX 3800M as far as apps go?

@Pirx - I take it that max and perhaps some cs5 apps work better in openGL?

edit - does firepro use opengl as well or doe it use something else?

 

firepro uses directcompute, I'm not sure if they can use opengl or not? The price difference is because it's nvidia, and because people will pay it; though to be fair if a business user needs CUDA (some do) they will have no choice but to pay for it.

 

It technically does as well as using DirectCompute as Hakira said as well as OpenCL. However, in terms of pitting against a Quadros, it is still the inferior choice as the drivers I have used so far has yet makes use of DirectCompute or OpenCL. Even though ATI states they do, GPU-Z shows they don't and ATI is slowly trudging over the laptop drivers to ensure they do (or I last heard).

Most of the professionals that do OpenGL or CAD work still ended up buying the Quadros over the FirePro. The FirePro is there as an excuse for us "gamers" to game lol (since it is a re-BIOSed 5870M).

However the FirePro is not so bad in CS5 and Stream though. I ran it through video conversion and encoding as well as some CS5 with GPU enhancement, it does seem to run those renders faster than using a DX-biased card.

However, if your brother is a seriously die-hard user in professional rendering applications, I would still have to say go with the Quadros. Their cards are often converted with more than just a BIOS swap. Many recent Quadros got some hardware differences that makes them very difficult to become GeForces again and vice-versa. With FirePros, ATI hasn't done much other than just flashing a Radeon into a FirePro with some BIOS and then eventually get drivers to activate the remaining features. Furthermore, nVidia is always better with the driver support .

 

thx a lot, I learned a lot today. I had no idea before that there was such a difference in quadro and geforce until now. And I took some time to look more into openGL.

I do have one more question but I think I'll start a new topic instead.

thx again.

 

actually, one last question, I might suggest to him the 485m. Which would u get for those apps I mentioned before, 485m or the Quadro FX 3800M?

 

I think the 485 is more powerful than the q3800, you'd want a q5000m to top a 485 I think.

 

In gaming or typical powerhouse benchmarks yes. HOWEVER, for OpenGL applications, the FX 3800M might be better optimized than the 485M. For DirectX, then yeah you need 5000M power to rival the 485M (and it would still lag behind a bit).

However, if you want hardware to hardware raw power (ignoring OpenGL or DirectX), then the 485M equivalent is the 5000M as of today's posting.

 

k cool. ok, this is really really the last question xD

Between the geforce and quadro, will there be a difference in appearance when looking at a rendered image from say 3ds max or something? Detail? texture? Will the card determine how it looks or is it purely artistic skill?

 

There will be little to no visual difference; the entire point of a business class card is its raw number crunching abilities and processing algorithms (which are sometimes hardware based like some of the recent quadros, and sometimes not like the firepros). You buy a business class card so that it doesn't take infinity+1day to render everything in real time - they are purely for production and speed-crunching.

 

Are you more concerned with performance while creating/modifying objects, scenes, video.... or in speed of renders & encodes?

create/modify: 3800
render/encode: 485

 

I will have to say rendering still favors the 3800M. Some owners of the Elitebook 8740w uses it for video production and they prefer the Quadros over the GeForces on compiling video.

 

I would hope so as expensive as those Q's are

Well, looks like both 485m and 3800m is out, too expensive for him.

lets crank it down to say.. 460m vs FX 2800?

 

For professional stuff, stick with the FX 2800, even though it is a more dated card hardware-wise. However, for professional rendering, the Quadros is still quite a beast.

 

vray/iray use CUDA cores as CUDA cores, regardless of driver support.
More CUDA = faster renders (this does not translate into viewport performance)

Even the iray FAQ says the only benefits to quadro/tesla over geforce is stability and available memory.


I watched a webinar by autodesk re; XBR project (for 3ds 2012). Apparently quicksilver was a prototype viewport client, and will be integrated as such into 2012.
(so more CUDA will then most likely equal faster viewports).

 

However if he needs OpenGL support on top of it, the GeForces in the similar caliber will crumble over the stability and optimization of the Quadros.

 

The programs he listed aren't going to use ogl unless he forces them to. D3D by default.

 

Yea, I normally see him sometimes switch between the 2. I started switching between the two yesturday for a bit only to see if I could feel a performance difference but not yet on the rendering.

One thing that was annoying me with OGL was the rectangle select was very glitchy.

well, my bro wants one now so hopefully gentec or xoptic has one laying around... if not, he will go with 460m since he doesn't want to wait.