Hello everyone, I am trying to understand how the worm code runs in parallel, is it still a single run or multiple independent runs? I am asking because when I run, say task.in.xml, with mpirun and 4 processors, apart from task.out.xml, there are also task.run1.out.h5,..,task.run4.out.h5 and task.out.h5. What does each run do and how are the results of the different runs "merged" to obtain the results in task.out.h5 and task.out.xml? Thanks in advance, Francisco Cordobes