I have a GTX570 in a win7 machine that I use for CUDA GPGPU work. I am getting some device timeouts; the solution to this is to perform computations on a second card with no displays extended to it.

Is it possible to stop extending the display to a card and in doing so circumvent the timeout restriction? (i would then remote desktop in)

Note: I have tried unplugging the monitor and disabling the card in device manager. If I disable the card in device manager I can no longer do any GPU computations.


You can just edit your TDR settings. We share information about that in the ArrayFire/Jacket Documentation. No need to disable your GPU in Device Manager.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.