Hi all,
Has someone recently compared a Quadro vs a Geforce card with similar specs to see if VTK performance improves? If so, any feedback would be appreciated.
Thanks!
Results would vary wildly depending on what you are doing, and of the chosen GPUs.
- Rendering heavy code would be faster on RTX card
- VTK-m cuda code would give the Quadro an edge
- Pure OSPRay rendering would not depend on your GPU at all
What are you trying to do ?
I want to display 6 volumes of 512x512x512 in different render windows using vtkGPUVolumeRayCastMapper with some objects in Wavefront OBJ format.
RTX cards should be better than quadro for this.
Thanks Mathieu!
Quadro cards are more expensive mostly because they are intended for “professionals”, who are willing to pay more for basically the same features. There are some small (but for some people, important) benefits, such as that Quadros are tested with a selected set of applications (mostly CAD and graphics design), they have longer shelf life than a GeForce, and they don’t have some limitations that are introduced to GeForce cards to impede high-end usage (e.g., maximum texture size).
However, Quadro cards are used by much less people and less tested with various applications (games and consumer and research applications), so it is more likely that you will run into compatibility issues.
For researchers it usually makes more sense to get a GeForce, as you may get about 2x better performance for the same price.
For some very specific use cases, such as run a certified application, put it in a shared server, build into a medical device, have access to maximum possible amount of memory, a Quadro may be the better choice, despite the extra cost.
3 Likes
Thanks a lot Andras for your detailed answer, it is very enlightening.
Another difference is that Quadro cards (I think) use ECC memory, which may matter in some applications.
1 Like
Indeed, but in the context of visualization and rendering, ECC memory does not really matter.
1 Like