Comment (by Erik Schnetter):

Ah yes, silly me. We should be using posix_memalign.

--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2832/possible-race-condition-in-loopcontrol