OK, we can support this. (You could do this already through MPI if there is an MPI parallel make.) werner.krebs@yale.edu