Utility subroutines and functions

Current implementations include an assortment of utility routines:

• Subroutines to convert process type and process indices to process number, and vice versa. (These routines identify each process with a unique single number, which is sometimes useful.)

          integer function iprocnum(iptype, ipx, ipy, ipz)
          subroutine pxpypz(iproc, iptype, ipx, ipy, ipz)

  In no-host-process implementations, the argument iptype is omitted.
• Subroutines to convert from global indices to local indices (plus process-grid indices), and vice versa.

          subroutine ilocal(ii, ipx, iiloc)
          subroutine jlocal(jj, ipy, jjloc)
          subroutine klocal(kk, ipz, kkloc)
          subroutine iglobal(ipx, iiloc, ii)
          subroutine jglobal(ipy, jjloc, jj)
          subroutine kglobal(ipz, kkloc, kk)

• Subroutines to compute the intersection of a range of global indices with the local intersection. (This is useful in determining, e.g., whether the local section contains part of a boundary and if so what the local indices are.)

          subroutine xintersect(iglob1, iglob2, iloc1, iloc2,
       -           lempty)
          subroutine yintersect(jglob1, jglob2, jloc1, jloc2,
       -           lempty)
          subroutine zintersect(jglob1, jglob2, kloc1, kloc2,
       -           lempty)

• A subroutine to append a unique-to-each-process suffix to a filename. (This allows each process to access a different file.)

          subroutine fname(iptype, ipx, ipy, ipz, inname,
       -          outname)

  In no-host-process implementations, the argument iptype is omitted.