You are not logged in.
Pages: 1
Hi all,
I'm currently testing the bumblebee and nvidia-xrun approach on my laptop (Lenovo Legion Y740, RTX 2070).
Does the performance penalty of bumblebee only applies to graphical applications?
I tested both approaches at rendering from Blender on the command line and using tensorflow-cuda and no significant differences appeared. For example one case rendered on 3:34 minutes on BB and 3:35 on nvidia-xrun (using openbox on a separate TTY).
If that's correct, is bumblebee the simplest way for scientific calculations (using optirun when needed)?
Thanks in advance
Offline
Pages: 1