Tesla C2070 Display resolution question.

Resolution question. All of the Quadra FX and GeForce Cards show a digital resolution max but the tesla’s do not in the product specs.

I understand tesla’s are primarily used in high end computation and not as display cards… BUT when using cuda cores to do GPU based 3D rendering the model loads into video memory based on the lowest video ram capability of the cards you have installed. So if you have 2 Geforce 470s and a Tesla C2070 you only really have 1G available to load in your mesh to your rendering program. In some cases you can turn off cards but not always. So even though GeForce cards are much more cost efffective for GPU Cuda rendering if your models are large you have to go Quadra or Tesla. QuadraFX5800 ram-to-cuda core ratio is poor… so the Tesla starts to have some appeal… lots of ram and a decent number of cuda cores.

The next question is resolution. If you are doing print resolution for large posters you need to be able to spec large native render sizes.

So finally my question:

What is the max display resolution of a tesla C2070

Are they for sale yet? Anyone know how much?



The first question isn’t really answerable, since Teslas don’t have video out. But they could display much bigger than a 2GB GTX285, between 2x and 4x as much, I’d guess, if they could.

Not for sale yet. They are still working out the bugs for the commercial gaming chips, and probably trying to find those few chips with all 512 cores functional.

How much? A real lot. I’d guess between $1500 and $5000, probably near the higher end of that range. But nobody who isn’t working for nVidia knows.



The Fermi-based Tesla C2050/2070 certainly do have a video out.

The 2070 is only 448 cores, like the 2050. The difference is 6GB of RAM versus 3.

The release date is just for the next generation of DDR5 RAM chips with doubled density to become available.

Hynix is promising the 2Gb GDDR5 for the second half of 2010, which matches with the S2070’s expected delivery date.

Wrong. The T20 series discrete cards have a single DVI video out and the cards can drive up to 2560x1600 with a dual link DVI monitor, or 2048 x 1536 with a single link DVI monitor.

Wrong again. Only 448 cores in both the C2050 and C2070. And the C2050, which is the same GPU with less memory, is already in customers hands.

And finally, also wrong. Prices were publicly announced 8 months ago when the products were announced - $3999 RRP.

Not a great batting average in that post, eh?


In my defense:

  1. The single video out on T20 cards is new, and a minor deal, really, they are for crunching numbers.

  2. Tech sites have speculated that the C2070 cards will be designed to the full spec of the GTX100 design at 512 cores. Whether this happens or not is unknown, but nVidia aimed low with their prerelease announcement.

  3. I stand by the 3rd statement - an announced price is a good guideline, but fab issues (or successes) may change the final price. And it did hit my (admittedly broad) range.


Thanks for the feedback. I will put C2050/C2070 display spec on our website.

The max res we support is 2560x1600 at 60Hz.

Pls take a look at this link: http://us.download.nvidia.com/Windows/Quad…lease_Notes.pdf

The C2050s are shipping but not the C2070 yet. We already announced that C2070 MSRP is $3999.