Using Cycles render with either my CPU (intel i5-4690k) and my GPU (GTX 750ti) yeilds the same result. When I go into user preferences and select my GPU under CUDA, it displays it as: "GeForce GTX 750 Ti (Display)" Does the (Display) mean that only my CPU is rendering, and it lets my GPU then process the other UI in Blender?


1 Answer 1


The "(Display)" in system preferences means that GPU will work to process all elements in viewport, not on rendering. You just make it available. If you really wanna have GPU rendering acceleration, you have to select in Render section

enter image description here

  • $\begingroup$ I've been using that option too. Was more wondering if there was a standard way to select the GPU under CUDA. That being said, maybe my CPU is just better than my GPU at rendering in cycles? Does that make sense given which respective models they are? $\endgroup$
    – O ciN
    Jun 17, 2017 at 4:33
  • $\begingroup$ It's a possibility... It's a entry-level grapphics card. And the architecture is old too. But I'm not sure... Just having some tests... I'll tell some results I've got here. My CPU It's a Core i5 7400 and my GPU its a GTX 1060 3GB Test 1: I created I simple scene with a cube, and applied 4 subdividions surface on it, making a total of 1536 faces. I applied a glass shader on it. The CPU rendered on an average 10.6 seconds. The GPU rendered in 6.6 seconds (I ran around three times) it's a 40% speed up $\endgroup$ Jun 17, 2017 at 7:23
  • $\begingroup$ Test 2: The same object, but now I applied 5x subdivision surface making 6144 faces. The CPU now has took about 12.1 seconds to render and the GPU took around 6.9 seconds... Almost 50% speed up now Test 3: Now I put 6x subdvision and duplicated the the sphere making a total of 49,156 faces on scene. The CPU now took around 17.3 seconds. The GPU took about 10.05 seconds. A 40% speed up again Test 4: One single cube, with standard material... The CPU took about 5.3 seconds and the GPU 2.6 seconds, about 50% of speed up again $\endgroup$ Jun 17, 2017 at 7:44
  • $\begingroup$ Depends a lot how many objects there's on scene, polygons to calculate, textures, lightning, particles effects, pyshics... The Intel 4th generation is sitll a good processor but the 750ti I have my doubts... If the processing power difference between these chips is small, maybe you only see this little difference on specifics tasks that GPU was designed for (like physics, effects, a huge amount of polygons to calculate) $\endgroup$ Jun 17, 2017 at 7:50
  • $\begingroup$ My hardware got about 40-50% percent of speed up, yours maybe can get only 5-15% (just picking a random range), but a high end grapphics card like GTX 1080 can get even more than you and me, like 70-80% (depending of the scene, of course) of speed up... It's relative... There's no way to predict this. Only testing and benchmarking... And depends a lot of the complexity of scene... Try put hundreds of thousands of polygons on scene, each one with a differente texture, UV map, lightning and shadows and maybe you'll see better results $\endgroup$ Jun 17, 2017 at 7:54

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .