According to my previous experience, Nvidia prefers to tell nothing rather than something in case of such questions (subj).
In case of my project, watchdog limitation becomes crucial. It must have been fixed in CUDA 2.0, however, this issue is still in the case.
I’d kindly ask Nvidia guys to answer … What is the very approximate time of next CUDA release that would fix the watchdog problem ? Just an estimation: days, weeks, months ?
This issue wastes the whole work and stalls the whole project, I can not plan the future work.
Thanks in advance!