Utilities for low overhead threading in Julia.
Please see the documentation.
If you're using Windows, please note that Windows often allocates memory when neither Mac or Linux do. I do not know why. If you can help diagnose/fix the problem, please take a look at count_allocated() in /test/staticarrays.jl.
If you are trying to mix code relying on ThreadingUtitlies.jl (e.g. LoopVectorization.jl's threaded code, Polyester.jl, or Octavian.jl) with other threaded code, you can use ThreadingUtilities.sleep_all_tasks() to put ThreadingUtilities's tasks to sleep before switching from ThreadingUtilities to the other threaded code. Manually putting ThreadingUtilities's tasks to sleep in this way should stop them from interfearing.