when nonbonded and nonbonded14 flags would be different in an input
file. There are no tests either that have such combination of
input flags. Moreover: The name in the source code and the input are made consistent. It
was also confusing to have a nonbonded flag and a nonbonded section
in the FORCEFIELD section.
svn-origin-rev: 10558
contribution is simply removed from fist_intra_force, and changes a few
other things to activate the onfo evaluation on fist_nonbond_force.F.
Many preparative patches are now effectively used. There is still a lot
of stuff to be removed, which will be done in later patches (Toon).
svn-origin-rev: 10554
* Reset of 204 regtests due to numerical noise
* Please, check new code (especially basic functions) always for a missing "_dp" and use the constants from the module mathconstants preferibly to avoid such annoying incidents
svn-origin-rev: 10155
* Old weights are still available using the compile flag -D__USE_LEGACY_WEIGHTS (source of these weights is not specified)
* Ground state electronic configuration of Ce and Tc corrected
* Unused component mass removed from ptable (weight of most abundant isotope)
* Elements up to 118 added
* Reset of more than 300 regtests
svn-origin-rev: 9847
from CELL and defining instead units for A,B,C and ABC keywords. Added two more keywords
in COORD sections to have an additional and disentangled control of the units for
coordinates (UNIT and SCALEDi, in &COORD). Updating the whole bunch of regtests and
resetting 9 for numerics (due to the new conversion tools). Several cosmetics.
Moreover, this commit introduces the possibility to have units also for output quantities.
At the moment this keyword is supported only for a minor set of print_keys (essentially
the ones that were depending on the unit_of_length of cell). More work to do on this
topic in the future.
svn-origin-rev: 7967
specifying a more easy to use way for the iteration levels. The section
EACH constains several keywords each attached to a specific each iteration
level (for example the keyword MD that maps the steps of the MD, and so on).
This should make more user-friendly how to decide the frequency of a print
for each print_key..
fixed a bug in the routine deciding if the present iteration level was the
right one to print (i.e. in the old version very weird behavior with the
keyword EACH was possible). Now it should show the right behavior.
Regtests have been properly adapted.
Moreover the iteration_info has been fully made consistent with retain/release.
This makes impossible (gives an error at run time) to create iteration levels and
destroy them randomly or even keep them allocated. They must be destroyed in the
inverse order they are created. When adding a new iteration level
remember to update also the create_print_key_section properly (if not so a
runtime error will pop up..).
RESTART keywords have been also specified with the add_last_numeric specification
(this was essentially the behavior we had so far, though induced by the bug
present in the code, much more clean now).
svn-origin-rev: 6878
you need to specify the keyword "STRESS_TENSOR" in the MD section. PV_AVA and PV_NUM and
others have been deleted. STRESS_TENSOR takes care of them. This is the biggest user
change regarding keywords controlling stress tensor.
Next commit will make FIST possible skipping the computation of the stress tensor in order
to save some cpu time if stress tensor is not needed.
svn-origin-rev: 6716
BAROSTAT and NOSE non-optional.. this means that in NP* and NVT runs the header of the
BAROSTAT or NOSE must always be present (default values can always be used)..
this avoids possible bugs when using in cascade restarts produced with MD with
different ensembles..
added one more regtest for restart mixed_env NPT..
svn-origin-rev: 6169
