HISTORY

This is the atomic program originally written (circa 1982) by Sverre
Froyen at the University of California at Berkeley, modified starting
in 1990 by Norman Troullier and Jose Luis Martins at the University of
Minnesota, and currently maintained by Alberto Garcia, who added some
features (such as the calculation of logarithmic derivatives) and made
substantial structural changes to the April 1990 (5.0) Minnesota
version while at Berkeley and elsewhere.

Jose Luis Martins (JLM) is maintaining his own version of the
code. 

This version can be run in two 'compatibility modes' (see Samples
directory for examples).

* The default mode (also called 'mons' for other reasons), which uses
the same internal parameters as the current (5.63) JLM version and a new
exchange-correlation module (including GGA) written by Carlos Balbas
and Jose Soler. It gives numerical results which are, to a very high
degree of approximation, identical to those of the JLM version.

* The 'ucb' mode uses a different set of internal parameters (for grid
generation and pseudopotential tail treatment) which give numerical
results identical to those obtained at Berkeley (and physically very
similar to those of the JLM version, as the parameter differences are
minimal). 


INSTALLATION

To compile, edit the makefile to suit your compiler and 
architecture.

USAGE

See the new manual in Docs/manual.tex.

TIPS AND CAVEATS

The input format is quite rigid at the moment, and any resulting
errors tend to be a bit cryptic.

To have logarithmic derivatives computed, the user needs
to specify the radius at which to do it (variable logder_radius).
It is input on the "flavor" line (see subroutine input) (a.u.).

* Note that the logarithmic derivatives are computed
non-relativistically. Deviations between the pseudo and all-electron
results are to be expected for relativistic calculations, even if
the pseudopotential is "good". 


Please send bug reports and comments to wdpgaara@lg.ehu.es.






