It appears that 'libomp :: ompt/tasks/dependences.c' has a race. I don't know how my patches would have caused the failure below, and it passed at the next build.
Joel
A lot of these tests seem to have races or at least timing dependent behavior.
We recently had to disable some already. Our test infrastructure needs two improve here.
We need to run them multiple times if they are necessarily timing dependent.
We also should check if the ones that are need to be, e.g., if they all pass when we force
single threaded execution (or if failing is what we want them to do).
Long story short, I doubt you caused the failure but if you are interested you could look
into improvements in this area 
Thanks for the info! I’ll keep this in mind.
Joel