This mutex serialized computed events; this is not neccesary when processing
is only happening in the main thread. It seems like OpenCL might become the
properly encouraged way to use more cores with GEGL.