I keep getting a CUDA error when I try to render cycles with GPU Compute selected. In this answer (the site won't allow me to respond there) https://blender.stackexchange.com/a/18408/20471, @BlendingJake says that SSS on the GPU eats up so much memory that it is recommended to have more than 1 GB of memory on for your GPU. I'm having the same problem as the OP in that string; however, I have 2 GB of RAM on my GPU. Why do I still get this error? My specs are below:

1.System: OSX 10.11.3

2.Blender version: 2.76

3.Processor: 2.3 GHz Intel Core i7

4.Memory: 16 GB 1600 MHz DDR3

5.Graphics: NVIDIA GeForce GT 750M 2048 MB

6.CUDA Driver Version: 7.5.25

  • 1
    $\begingroup$ Depending on the size of your scene overall, it's quite likely that 2GB isn't even enough. $\endgroup$
    – PGmath
    Commented Mar 4, 2016 at 19:50
  • $\begingroup$ Also depends if you're using your card for video display as well as rendering, this leaves less resources for Blender. $\endgroup$ Commented Mar 8, 2016 at 1:21

1 Answer 1


I have the same issue on a Mac Pro 5,1 using an Nvidia 980GTX with 4gb of ram. It'll CUDA error on a simple scene sometimes.

• What I can find help is to shut all other apps down reopen Blender and try again.

• Reboot, only open blender and try again

• I use a 4k monitor, sometimes, halving the resolution can make it work.

I think it's a known error, that I really hope it's something that's addressed in the future.

Is this only mac users or do windows / linux users suffer the same problem? I might make a bootcamp or linux partition if there's no issues there.

  • $\begingroup$ Thanks, @cshelswell. When I only have Blender open, I am able to render a moderately complex shot. But a shot that is probably twice as complex results in the same error. I guess I know what my limitations are. I don't know if this is a Mac problem or not. Please post if you find that the bootcamp or linux partition solves it. $\endgroup$
    – Fletch
    Commented Mar 5, 2016 at 17:12

You must log in to answer this question.

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