I managed to run this using 2 cores instead of all available cores (took only 30 mins), since Lorenzo I think had suggested earlier in the forum that it is a pain to run open_grid.x in parallel. Still would be interested to know if there are other ways of reducing memory requirement. Zeeshan