I have an issue on premier (6000162760) titled "OpenMP Tasking crash" relating to nested tasks causing program crash. A reproducer was submitted. Issue opened 9/7/2016 and there appears to be no progress or information (e.g. ... we can get the reproducer to produce crash). This is a critical issue. It is not show stopper since I can code without using nested task levels.... but at the cost of not being able to make use of all the threading capacity of the system. The work is asymmetrical and is not suitable for partitioning using other means than tasks, and not fully efficient without use of nested tasks. My suspicion (from crash postmortem) is a race condition in the TBB (yes TBB) code merged into the OpenMP library. If you could get a comment on this from engineering (other than investigating...) I would appreciate it.
I have asked the support engineer handling that issue to provide an update. If I don't hear back this week I'll see if I can find someone else to take a look.