watch dog


I have a GTX8800 card and recently test a couple CUDA programs. I found that whenever the run time exceeds 10-11 secs, the program gets aborted.

Is there any watch dog machanism monitoring the run time?

If so, how can I avoid this? I have two GTX8800 cards installed and one is the main video card.

Many thanks in advance!!!

You didn’t state which OS you’re using. However, please see the CUDA release notes for a discussion of this problem, and potential work-arounds.


I run Fedora Core 6.