PROGRESS STATUS.
------------------------

Thurs July 16 11:57:06 CDT 1998

Changes made by Paul Plassmann

Got rid of interface to tools.n

main is now "start.f" where all the MPI initialization is done.

I converted all old "PI" calls to MPI calls and linked to the new MPI 
version of blkcm.

grep on "PI" to see the commented out PI calls and the new MPI calls.

------------------------

Sat May 20 17:08:05 CDT 1995

-) Finished and tested find_decom.f routine to compute a maximally
   cubic (or squared) decomposition.

-) Made some changes to worker.f. NOT tested YET !.

NEXT: write mg3d find_level.

------------------------

Wed May 31 18:37:56 CDT 1995

-) Created the utility libraries: utils_array.f utils_mg.f
   utils_scalar.f

   See library.txt for a description of the routines included.
   See meshf.h for a description of the descriptors.

-) Basically, the worker() routine is complete now, and it seems
   to work fine.

NEXT: write c_driver.c
      mg_find_buffers  should be modified, really, but we will 
		       consider this later.

------------------------

Wed Jun 14 06:26:24 CDT 1995

-) Extensive progress has been achieved. Basically now the
   code structure is present, and all the parameter arrays
   are defined appropriately. 

-) The grid is set up correctly.

NEXT: write init_cond.f
      This is the routine that defines the initial condition.

------------------------

Wed Jun 14 21:25:30 CDT 1995

-) Even better now. The code works in parallel, it seems.

-) Parallel IO is also there for the big arrays.

NEXT: write the hydro solver: fluid_3d.f

------------------------

Fri Jun 16 23:07:16 CDT 1995

-) The 3D fluid driver is in place. 

-) The PPM interface seems to be in place as well.

BUGS: -) the Makefile is not smart about recompiling meshf.h

      -) when changing from scalar to parallel one needs to
	 do a  'make clean' fisrt, or trouble will happen.
	 This is related to switching from single to double
	 precision.

NEXT: Write Boundary condition routine.

      Verify that the PPM interface works.

      Modify the mg structure to include the number
      of iterations at a given level.

------------------------

Sat Jun 17 19:04:25 CDT 1995

-) Implemented the Boundary Conditions driver. Only periodic
   boundary conditions are actually in place.

BUGS: -) Still all of those as of "Fri Jun 16 23:07:16 CDT 1995"


NEXT: Poisson Solver.

NOTE: The fluid solver has potential for inclusion of an artificial
      viscosity term and for source terms, such as those that arise
      in cosmological situations.

-------------------------

Thu Jun 22 14:43:29 CDT 1995

-) Basically, all the code is there and works in parallel,
   with all the IO pieces (including the dumping of the
   volume rendering arrays). The only thing missing is the
   multigrid.

-) The restart structure may have to be rethought. I need to
   take into account that nbegin is not always 1 any more.

NEXT: The parallel multigrid.
      Also, one may want to save the gravitational potential
      in order to save time on the initial poisson iteration.
      This is less important once the multigrid works.

BUGS: -) the Makefile is not smart about recompiling meshf.h

      -) when changing from scalar to parallel one needs to
	 do a  'make clean' fisrt, or trouble will happen.
	 This is related to switching from single to double
	 precision.

      -) On the Sp1 the code hangs if it trieds to read the
	 restart file back in from the unitree.

-------------------------
