The MCHF Atomic Structure Package
Fortran C Objective-J Perl
Switch branches/tags
Nothing to show
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
examples
src
README
make_atsp

README

      The MCHF Atomic Structure Package
      ==================================

	Charlotte Froese Fischer

	and collaborators:

	 Alan Hibbert and Ronnie Glass
	 Michel R. Godefroid
	 Bin Liu
	 Tomas Brage
	 Gregory Miecznik
	 Nathalie Vaeck
	 Per J\"onsson
	 Claes Goran Wahlstrom
         Jinhua Xi
         Gediminas Gaigalas

	 December 20, 1996
	 -----------------

Installation:
-------------

This directory contains:
. README (the present file)
. make_atsp, a script file which on most Unix systems
  will compile and link the atomic structure package; i.e.
  the command
       make_atsp
  issued from the current directory will
  . compile and make the libraries
  . compile and link the programs
  . compile and link the utility routines
 
. three subdirectories
  . bin -- a directory that will contain the library and
	executables oncce the package is installed
  . examples -  a directory which contains a subdirectory
	for examples of each major program.  The latter contains
	. one or more <name>.ex file which is a script file
	  for a complete calculation
	. output files and output (.out) for the examples.
  . src - directory with source files and a makefile for
	generating the package. This makefile has been prepared
	in a simple manner to compile on many Unix workstations.
	Even so, it is advisable to review the compiler flags
	set in the makefile

I/O related matters.
--------------------
 
  The applications have been written for easy interactive use
where the standard output can also be redirected to a file
rather than the screen. In order that the prompts for input
not also be redirected, "standard error" should be set to be
the screen.

  Most unix systems have predefined units:

  standard input  -- unit 5
  standard output -- unit 6
  standard error  -- unit 0 ( used here) 

However, Hewlett Packard uses unit=7 for standard error.  For
files which used to be directed to the printer (long lines)
it was customary to use unit=3.  The early sections of
applications define the unit numbers as indicated above.
Systems deviating from this convention will be easier to use
if code is modified to the environment.

  In the transition codes (MLTPOL, LSTR, LSJTR) it may happen
that the files defining the inital state and final state may
be the same for transitions where parity does not change as
for some forbidden transitions.  Most systems allow two unit
numbers to be associated with the same file through extensions
to standard F77, but not all. The present code does not use the 
extensions, but various possibilities are included in the code and 
may be used to replace the present, more simplistic implementation.
See the subroutine, CFGIN2, for possible extensions.

 For the LSTR and LSJTR, it is useful to collect results in one
file by appending to an existing file.  Again, the append option
is not standard, but can be implemented by choosing an
appropriate statement.  The OPEN statements that position the 
file pointer, again, are near the beginning of the main program.

Application Installation
------------------------

The file make_atsp should be an executable. This can be achieved
with the  command "chmod +x make_atsp".  In theory, the command
"make_atsp" should then bring up the system, but since compilers
differ, it may be desirable to enter each Unix command in the
script file separately, so as to be able to make sure one step
has completed successfully before attempting the next.

For the script files (names ending in ".ex") in the examples
directory, it is helpful to "set clobber"; otherwise, if the
file exists, the step may be aborted.  This may happen when, for
some reason or other, the example is run more than once, though
an effort has been made to avoid this possibility.

Post Installation Considerations
---------------------------------

The dimensions in the current version have been set for moderate
size calculations, appropriate for simple cases and for
learning.  All important parameters are defined by KEYWORDS
and can easily be changed with the use of an editor.

However, it is important to remember some relationships.

COM.f -- library file, independent of parameters of COMMON
ANG.f -- library for angular integrations
         Name  Value   Purpose
	 ----  -----   -------
	 NWD   30      Maximum number of radial wave functions
	 NCD   100     Maximum number of configuration states
	 NWD2  2*NWD   Appropriate for transitions with two wfn files
	 NCD2  2*NCD   Appropriate for transitions with two CSF files
	 NCD4  4*NCD   Appropriate for Breit Pauli calculation
NJGRAF.f -- Parameters of the form KFLxx (Change with care) Each
	    must be consistent across all applications
RAD3.f -- library for radial properties (single set)
         Name  Value   Purpose
	 ----  -----   -------
         NWD   30      Maximum number of radial wave functions
	 NOD   220     Maximum number of points in a radial function
	 IWRITE  6     Unit number for output
RAD6.f -- library for radial properties (two sets as in transitions)
	  same as RAD3.f except that NWD2 = 2*NWD is used as 
	  dimension parameter.

Application programs  -- in each application, PARAMETERS may be
set to have values that are consistent with the libraries that are
used in their installation (see makefile for details).  Thus, in
MCHF for example, NCDIM may be freely increased (or decreased)
since this parameter does not appear in any library.  Of course,
it should be modified in the entire MCHF.f file.  But this
parameter also appears in NONH.f, so it would be more consistent
to change it in all applications.