ADF version 2005.01 subreleases

2005.01b (October 11, 2005)

New platforms:

ADF-GUI fixes:

.adf files written by old ADFinput versions could not be read
import coordinates problem: lookup of atom names and charges did not work
when saving a .adf (and .run) file outside SCM_RESULTDIR, the logfile would not display properly
when opening a .adf file, ADFinput did not check for updated .t21 files in SCM_RESULTDIR
on some platforms (at least macosx and windows) excitation spectra did not work in ADFspectra
the status field (lower left field) in ADFinput and ADFmovie now remains visible also for small window sizes
updating the coordinates in ADFinput from .t21 (after Open or Revert) now also works after an optimization using internal coordinates
ADFlevels did not handle interactions with core orbitals properly, leading to crash when showing interactions
ADFlevels crashed when showing interactions of old .t21 files
adfspectra now can read frequency data from .t21 produced by SD
after importing from t21, the undo stack is updated properly
problem with Help menu fixed (Macintosh only)

ADF-GUI new features:

delocalized optimization added
integration grid smoothing option added
the symmetric displacements and IR-visible modes only options are added for frequencies runs
list of available Hessian update schemes updated
SD added
spin-flip excitations added
adfinput only updates coordinates from t21 when the matching .adf file is older
saving while old job results available now always activates save as
by default hide bonds to hidden atoms (existing .scm_guirc files will override this)
transition state search template now specifies BOFILL hessian update scheme
by default, do not run densf any more (default changed in templates)
when adjusting distances, angles or dihedrals the first selected atom determines always the group to be moved
when running an optimization, adfinput now detects finished run and suggest to update coordinates

atomicdata:

moved actinide with 6p frozen core to atomicdata/OLD directory,
such large cores can give unreasonable results

All versions:

work-around for XLF compiler bug on IBM AiX - memory leak in a KF routine leading to ADF crashing
fixed the issue with segmentation faults during the 1-st SCF cycle on some, mainly Linux, machines
SCM_NODOMAINCHECK environment variable added to disable domain checking
bug fix ADF did not run when the adfhome directory was nested too deep (path too long)
added mPBE functional, fixed some references in list of meta-GGA XC functionals
error handling BLYP if LDA VWN is specified
BP86 shortcut
bug fix for exchange-only REVPBEx, RPBEx, MPBEx
allow DRF input keys with spaces at the beginning
bug fix DRF which could cause segmentation faults
bug fix CINEB if numer of stored geometries too large
bug fix accuracy gradients closer to adf2004
bug fix symmetric frequencies for linear molecules
bug fixes parallel COSMO (was crashing)
bug fix reading restart file TAPE21 (frequencies) generated by adf2004.01
when using the BASIS key, all ALLOW options are now also passed to the create runs
bug fix passing around unallocated (and unused) array caused segment violation on at least compaq
bug fixes dividing by zero of unused variables which caused segment violations on at least compaq
license code could cause segment violation

MPI versions:

fixed issue with pdb2adf when SCM_USETMPDIR variable is set
I/O buffer size was set too small due to wrong handling of the SCM_IOBUFFERSIZE variable
SCMLICENSE variable was not passed properly from main task to child tasks, sometimes leading to license errors