memory access fault when linking dynamic lib that uses hipcub about rocprim HOT 7 CLOSED

I'll look into it.

from rocprim.

Does the same thing happen with rocPRIM?

from rocprim.

No, does not happen with rocPRIM.

from rocprim.

We identified the problem: for some reasons in said configuration (linking to a dynamic library that uses hipcub if I use hipcub also in the main program) every kernel is run twice by HIP framework. Some kernels are resistant to that, some are not (access errors). We are not sure why it happens only for hipCUB.

Right now we are thinking that probably in here should be a break after hipModuleLaunchKernel().

from rocprim.

Here are HIP API traces for correct and incorrect (with bug) execution, where we can clearly the difference.

bug_hiptrace.txt
correct_hiptrace.txt

from rocprim.

I guess we should open issue in HIP repo.

from rocprim.

Should be fixed by ROCm/HIP#575. I don't know which ROCm release will include that.

from rocprim.