      --------------------------------------------------------
      README FOR USPP 7.3.5 PSEUDOPOTENTIAL GENERATION PROGRAM
      --------------------------------------------------------

Version uspp-733 is exactly identical to uspp-734 except for
insertion of some comment lines in runatom.f and scgsubs.f.
This was done in order to facilitate interoperability with the
"USpp2Abinit" module of the ABINIT software project, which allows
"PAW potentials" suitable for use in ABINIT to be generated from
the USPP package.  See
  http://www.abinit.org/ABINIT
  http://www.abinit.org/ABINIT/USpp2Abinit-Manual-html/USppPAW.html
  http://www.abinit.org/ABINIT/Psps/PAW/paw.html

      --------------------------------------------------------
      README FOR USPP 7.3.4 PSEUDOPOTENTIAL GENERATION PROGRAM
      --------------------------------------------------------

Implemented several minor bug fixes (none appear to be serious) and
cosmetic improvements.  See header part of runatom.f for deatils.

      --------------------------------------------------------
      README FOR USPP 7.3.3 PSEUDOPOTENTIAL GENERATION PROGRAM
      --------------------------------------------------------

Version uspp-733 should be functionally identical to the previous
distribution known as a7.3.2.

Regarding the fortran programs, the only changes are very minor
(some small bug fixes and changes to ensure problem-free
compilation on a wider variety of platforms).

The change from the 'aX.X.X' series to the 'uspp-XXX' series
involves a significant reworking of the "packaging," i.e., the
directory structure, the Makefiles, the sample files, etc.  Two
important modifications are that:

-- The GNU GPL License is now explicitly included as a header
   of each source file.  See Source/LICENSE.

-- The Work subdirectory contains a library of input files that
   can be used to generate a library of pseudopotentials.
   Hopefully this will grow to cover many atoms of the periodic
   table.

These changes were made with the intention that uspp-733 and
subsequent versions will be posted publically on the Web at
www.physics.rutgers.edu/~dhv/uspp/ .

       -------------------------------------------------------
       README FOR A7.3.2 PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

Cosmetic changes: Fixed minor problem that made reported "target
and achieved values of dqfunc" appear to differ more than they
should, and overhauled 'ifprt' print control.

*** June 4 2001: A minor bug fix has been incorporated, without ***
*** generating a new version of the program with a new version  ***
*** number.  See file 'REFORM_BUG' for details.                 ***

       -------------------------------------------------------
       README FOR A7.3.1 PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

Mainly minor changes: add makefile support for IBM; add ability
to plot beta and chi functions; add example directory for
generating potentials needed for CUSP tutorials.

       -------------------------------------------------------
       README FOR A7.3 PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

This version can handle f-states (modifications provided by Chris
Pickard).

In addition, changes have been made to improved the stability of
the convergence of the eigenvalue ecur in subroutines schsol,
dirsol, koesol.  Also, the initialization of the inhomogenous
solution 'eta' of the radial schroedinger equation has been changed
in subroutine schgef, again to improve numerical stability of the
solution.  (As a result of the latter change, some wavefunction
solutions may change sign relative to previous versions; this
should have no significant consequences.)

A few minor cosmetic changes were made as well.

       -------------------------------------------------------
       README FOR A7.2.1 PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

This version fixes a benign bug in the pseudopotential report (that
entered somehow in going from a5.0 to a5.1); fixes a bug in
aesubs.f that affects the xc energy in the pbe case; has some
cosmetic improvements; and has a sample makefile for Si showing how
to do pseudopotential transferability tests on excited
configurations.

***  Finally, it also fixes a bug in pswf1 that has been      ***
***  present since a3.0.  Egad!  This bug affects the pseudo  ***
***  wavefunctions when iptype=1 (exponential pseudization).  ***
***  Since this style of pseudization is also used in         ***
***  generating the local potential for a certain angular     ***
***  momentum channel, the bug also affects the local         ***
***  potential, except when lloc=-1.  So, this also destroys  ***
***  backward compatibility:  the new version, when run on    ***
***  an old input file, can give a different pseudopotential. ***

       -------------------------------------------------------
       README FOR A7.2.0 PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

This version introduces variables nqf,ptryc that are distinct
from nqf,qtryc (these govern the smooth pseudization of the
pseudo wavefunctions and q-functions, respectively).

***  This destroys compatibility with input files of previous  ***
***  versions, but the program will issue a warning and        ***
***  terminate if it thinks it is reading an old-style         ***
***  input file.                                               ***

       -------------------------------------------------------
       README FOR A7.1.0 PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

Ah, finally, this one is successfully backwards compatible with
a5.1.4.  In this version, ifqopt=2 generates the old-style
pseudization of the q-functions, while ifqopt=3 does it
Stokbro's way.

       -------------------------------------------------------
       README FOR A7.0.3 PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

This is intended to be a 'cleaned-up' version of Stokbro's a7.0.
Relative to A7.0.2_NG, it has a few bug fixes and cosmetic fixes,
but it still generates pseudopotentials the way Stokbro intended.

However, it is *** NOT BACKWARDS COMPATIBLE *** with a5.1.3
versions, since it pseudizes the q-functions differently in the
case ifqopt=2.  Thus, when ifqopt=2, it generates DIFFERENT
pseudopotentials than a5.1.3 when run on IDENTICAL input files.

       -------------------------------------------------------
       README FOR A7.0.2_NG PSEUDOPOTENTIAL GENERATION PROGRAM
       -------------------------------------------------------

Additions by Keith Ng:

   included PBE-GGA in the call, by choosing exfact=5.
   in the input file .   ( by Keith Ng 2/10/97)
   
   bessel.f,  initsubs.f, trou.f from a7.0.2.dv/ version

Apparently does not run for all cases.

       ----------------------------------------------------
       README FOR A7.0.2 PSEUDOPOTENTIAL GENERATION PROGRAM
       ----------------------------------------------------

This distribution contains version a7.0.2 of the ultrasoft
pseudopotential generation codes from the group of D. Vanderbilt,
Rutgers University.  Please do not pass it on to third parties
without permission.

This will be 7.0.1 plus some clean-up operations:
(a) get rid of nag routines in favor of eispack routines
(b) make sure it compiles without error messages, at least on Sun

                                        D. Vanderbilt 21 June 1996
                                        dhv@physics.rutgers.edu

       ----------------------------------------------------
       README FOR A7.0.1 PSEUDOPOTENTIAL GENERATION PROGRAM
       ----------------------------------------------------

This distribution contains version a7.0.1 of the ultrasoft
pseudopotential generation codes from the group of D. Vanderbilt,
Rutgers University.  Please do not pass it on to third parties
without permission.

This is a temporary version.  It was sent by Kurt Stokbro 20 June
1996.  Basically all I did was to change the README file, the
comments at the top of runatom.f, and the version setting (`iver')
in runatom.f.  The program has not yet been tested by anybody in my
immediate group.  Note that some minor improvements incorporated in
going from a5.1.1 to a5.1.3 are not yet incorporated in this
version.  These will soon be incorporated in a new version a7.1.

                                        D. Vanderbilt 20 June 1996
                                        dhv@physics.rutgers.edu

       ----------------------------------------------------
       README FOR A7.0 PSEUDOPOTENTIAL GENERATION PROGRAM
       ----------------------------------------------------

Several features have been added to the 5.0 program, especially the
generation of GGA pseudopotentials and mixed normconserved/ultrasoft 
pseudopotentials. For a list of features see the runatom.f file. A
more extended list of input_examples is given in the input_examples
subdirectory. See also the file INPUTs which gives a description
of the input file format.

                                        Kurt Stokbro,
	                            Email: stokbro@mic.dtu.dk
                                        16 February 1996

       ----------------------------------------------------
       README FOR A5.1 PSEUDOPOTENTIAL GENERATION PROGRAM
       ----------------------------------------------------

This distribution contains version a5.1 of the ultrasoft
pseudopotential generation codes from the group of D. Vanderbilt,
Rutgers University.  Please do not pass it on to third parties
without permission.

The distribution is not very well documented at the moment.  Here
is a quick summary.

The main directory where this README file resides contains several
fortran files which are needed to compile the executable runatom.x,
and one file needed to compile the auxiliary executable reform.x.
The latter is only used to convert the pseudopotential data files
from unformatted (default) to formatted form, or vice versa, for
moving pseudopotentials between different computer systems.
Issuing the command "make" should automatically compile both
executables (see the Makefile).

Example input files are provided in the subdirectory
input_examples.  Change your working directory to this
subdirectory, make sure you have execute permission on the shell
script 'run_allatom', and issue 'make'.  If all goes well it will
run the all-electron calculation and then the pseudopotential
generation for each each of the three elements O, Ti, and Pb.  (You
can see that for Pb we had to 'fiddle' in order to avoid negative
charge densities in the core region when pseudizing the charge
augmentation function.)

The program is also capable of running "transferability tests" in
which a pseudopotential generated in one configuration is used to
calculate eigenvalues etc. in a second configuration, and the
result is compared with an all-electron calculation in that
second configuration.  However, an example of this type is not
yet provided in the input_examples subdirectory.

Please direct questions, comments, or (especially) concrete
suggestions for improvement of the documentation, to David
Vanderbilt at dhv@physics.rutgers.edu.

					David Vanderbilt
					11 March 1994


c     ------------------------------------------------------
c     note: in this program:
c     potentials, e.g. rucore, are really r*v(r)
c     wave funcs, e.g. snl, are really proportional to r*psi(r)
c       and are normalized so int dr (snl**2) = 1
c     thus psi(r-vec)=(1/r)*snl(r)*y_lm(theta,phi)
c     conventions carry over to beta, etc
c     charge dens, e.g. rscore, really 4*pi*r**2*rho
c     ------------------------------------------------------
c
c     rydberg units are used throughout this program
c
c     ------------------------------------------------------
c
