cuMemcpyDtoH error error problem

Hi all,
new here and mostly to whole GPU stuff, been working on some easier codes by now
and I’ve came across this error and I can’t seem to find any documented solution:

“call to cuMemcpyDtoH returned error 702: Launch timeout”

and was wondering if anyone came across it and perhaps have an insight he’d like
to share with me about it :)

Thanks in advance and best regards,
Goran

It probably means that the kernel launch prior to the memory copy call. Looking at the documentation for the call it says this:

The launch timeout probably arises because you kernel execution time is longer than the display watchdog timer limit on your platform and you are running code on a display GPU. The limit is something like 5 seconds on most platforms.

Thanks for info. After some testing, seems like linux platform allows up to 12 sec limit before shuting operation down. (That was tested with CUDA fortran code)