Compute model Tweet

[newfriend]

Hi, is there any topic where is described the difference between different compute models? Is there any difference with 3, 4, 5 or else models for Octane? How it's combines? If I have one card with V3 compute model and another with V5, how render is going? What do I loose? Speed, functionality or else?.. Where I can read about it? Thanks.

[abstrax]

Hi, is there any topic where is described the difference between different compute models? Is there any difference with 3, 4, 5 or else models for Octane? How it's combines? If I have one card with V3 compute model and another with V5, how render is going? What do I loose? Speed, functionality or else?.. Where I can read about it? Thanks.

Sorry, I didn't see your post. The compute model (or "compute capability") defines the hardware generation and architecture of a GPU. Each generation usually has a different architecture and newer generations usually have more CUDA features than older ones. We support all compute models down to 2.0 and mixing of different compute models is possible.