I am starting to do GPGPU programming with CUDA and am building a new rig. I am trying to do figure out the difference between the Quadro 4000 and the GeForce GTX 480 that justifies the price difference. I will be doing heavy scientific computing, things like performing SVD on MRI images. Something that is very parallel, very intensive scientific computing but not really graphics intensive. Which card would be better for me? I can get two 480 for the price of one Quadro, but the 480 has more cores. This doesn’t really make sense, and as far as I can tell they both use the fermi architecture. Could someone help me understand the difference, so that I can know and justify the price to my manager.
Hmm, no reply since yesterday. May as well give my 2 cents.
Quadro is more the professional graphics workstation card offering from NVidia. It is VERY good at graphics related stuff and CUDA capable, but not quite specialized for it. For scientific applications you should look at the Tesla series instead.
As for the differences between the Tesla and 480? A quick search showed this:
[url=“The Official NVIDIA Forums | NVIDIA”]http://forums.nvidia.com/index.php?showtopic=165055[/url]
So faster double computations, ECC memory, larger memory and possibly most importantly, increased support are what I consider the highlights.
Hmm, no reply since yesterday. May as well give my 2 cents.
Quadro is more the professional graphics workstation card offering from NVidia. It is VERY good at graphics related stuff and CUDA capable, but not quite specialized for it. For scientific applications you should look at the Tesla series instead.
As for the differences between the Tesla and 480? A quick search showed this:
[url=“The Official NVIDIA Forums | NVIDIA”]The Official NVIDIA Forums | NVIDIA
So faster double computations, ECC memory, larger memory and possibly most importantly, increased support are what I consider the highlights.