Hello,
I use CUDA on both Windows and Linux Ubuntu 12.10.
When running the particles example the Windows version runs at 360 frames per second on my GTX560M and on the same hardware under the Linux version it runs at only 60 frames per second. All the examples are slower too if run under the Linux version. Someone can tell me why?
What should I do to achieve the same performances undex Linux?
Thankyou very much.
Just a wild guess, but do you have the latest drivers?
Also, are you sure you’re not running something like Compiz at the same time?
And, as a wild guess as well, have you checked if you are running with debug symbols? I’ve seen a very slowdown when running with debugging enabled.
60 frames per second sounds like you are running with VSYNC on under Linux, limiting the frame rate to the screen refresh rate. Try turning VSYNC off.
Yes that’s true, I turned it off via nvidia-settings and now performance is acceptable. Thankyou very much for your help.