• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Quadro vs. Geforce for video editing?

TheInternal

Senior member
I'm having a lot of trouble finding good articles about why a Quadro should be purchased for video editing over a Geforce... especially considering that NVIDIA's Quadro vs. Geforce document is almost entirely about 3d and CAD optimizations, not video.

Has anyone seen any good write ups by folks other than fanboys or marketing nerds for why (or if) the Quadro has a significant performance gain in video rendering that can justify it's price point?

I'm looking to make a dedicated video (and possibly audio) editing workstation that will be primarily doing Premiere and some After Effects. I don't care about 3D performance for this build. Thus far, I've really only heard parroting of what sounded like marketing hype with things like "more precise" and "professional" when it comes to Quadro.

I could really use help cutting through the B.S. Any help appreciated.
 
"More precise" isn't marketing. It means better double precision performance compared to whatever they were comparing the card to. Basically 32 bit (float) vs 64 bit (double) in the graphics segment. Double precision is only needed for scientific or accurate pre-rendering as it's very costly for real-time rendering with very negligible gains (pretty much nil since it just aligns pixels twice as accurately as a 32 bit renderer).

If anything I think Quadro is the way to go because of OpenCL and generally better double and single precision performance compared to Nvidia's closed CUDA and 'its OK, not bad' performance
 
It depends on the specific software you use.

In fact, today, an unbroken support network is the ONLY good generic reason to buy a Quadro. Every other reason depends on your software, displays, and workflow.

Adobe's software will run a little faster on a Quadro, but not as much as the added cost, and the end results will always be the same between a Geforce and Quadro. What performance you can utilize really depends on how much you will be applying effects to the videos. Blending, adding watermarks, color correction, scaling to a different resolution, and so on and so forth, can be much faster with a GPU; while splicing content will be all CPU, as will a high quality encoding job.

High-end Quadros have ECC, but even if you're doing final renders with the GPU, and care for that, you also need an ECC-supporting platform for your CPU, and to actually utilize it, with ECC RAM.

If some minor software bugs could cause you to lose more money than a good Quadro costs, and you'd rather not take that risk, buy a Quadro. Then you will have an easier time getting to people that can actually try to help, instead of being SOL. OTOH, if doing that, you might as well buy a workstation that is certified in its totality.

"More precise" isn't marketing. It means better double precision performance compared to whatever they were comparing the card to. Basically 32 bit (float) vs 64 bit (double) in the graphics segment. Double precision is only needed for scientific or accurate pre-rendering as it's very costly for real-time rendering with very negligible gains (pretty much nil since it just aligns pixels twice as accurately as a 32 bit renderer).
No, it very much i marketing, for DP. Compare the Quadro K5200 with the Quadro K5000, FI. Many Quadros are no different than their respective Geforces, when it comes to DP. The results you get will be exactly the same, just faster completion times on some Quadros, and price point is not even the best way to figure that out.

Where it isn't marketing is for CAD and modeling problems that use old school OpenGL for display, where the desktop card drivers will be slower with very complex models, have lower quality blending, and lack some high quality AA options.
 
Last edited:
Back
Top