Tips for ABINIT users and developers

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Summary ======
This page is destined to be a memo, to list ongoing work or propositions to improve the naming and default values for abinit input variables
and other variables or other namings in documentation ...





===== General philosophy =====

We should move towards more useful defaults, and cull some which are counter productive or too specific

===== Propositions =====

default for ionmov should now be 22 (XGonze will do it)

default for ntime should be 20

default for shiftq should be 0 0 0 (Mverstraete will do it)

default for shiftk should be 
automatically set depending on space group.  (XGonze will do it)

nnsclo should be switched to 1 if MD/RELAX beyond the first step (MTorrent will do it)

[temporary] Use useria to decrease the number of times the SCF cycle must be converged from 2 to 1.   (MTorrent will do it)
Only for MD.

boxcutmin: Document usage for DFPT code. Make it clear that 2 is only needed to prepare DFPT calculations (check to be done by HMiranda)
Check the value of ngfft in hdr% when starting the DFPT code.

Meta-variable to define behaviour if unexpected values or weird configurations are detected by the code
e.g. chkprim, nkpt % nprocs != 0  

symsigma defaults to 1. Except for SC-GW

default for accuracy should be 4 

introduce new input variable "expertise", to trigger default values for the "chk" variables ...

==== To be clarified ===

(From ALherbier 20180614 - observed during the beautification ...)

cpuh, cpum, cpus => should'nt we use INPUT_ONLY and NO_MULTI both for all three variables?

diecut/diegap => should we incorporate only relevant if iprcel >=21 as in dielam?

kptbounds => personal remark : a big part of the description should be moved in another place of the website (kind of tutorial somehow).

occ => not sure if the dimension given in the description is correct (as it does not seem to correspond to the text).

postoldff => should we give the units (I suppose in hartree/Bohr) in the text description?

Should we remove the "specified" word in the following case ? *The use of this variable forbids the use of: specified(ngqpt) 

ratsph => the description text contains this "(in version 4.2, this procedure is NOT implemented, unfortunately)". Should it be removed?

scphon_supercell => "TO BE IMPROVED: should contain a tutorial on how to do self-consistent phonon calculations, David Waroquiers 090831" What should I do with this sentence ?

slabzbeg/slabzend => should be a scalar according to what I understood but the default value is an array =[0.0, 0.0] ?

so_psp => Is usepaw==0 still a condition? Or can we now use spin-orbit with paw in recent Abinit version?

spinat => In the description text, I found "In case of non-collinear magnetism (nsppol=1, nspinor=1, nspden=4)". Shouldn't it be nspinor=2?

symafm => In the description text, I found "non-collinear magnetism ([[nsppol]]=1, [[nspinor]]=1, [[nspden]]=4)". Shouldn't it be nspinor=2?

timopt => "If 4  -->  close to [[timopt]]=1, except that the different parts of the lobpcg routine are timed in detail. A different splitting of lobpcg than for [[timopt]]=-3 is provided." Should'nt it be [[timopt]]=3 at the end of this sentence ?

tl_nprccg => "TO BE IMPROVED: all tl_* and wvl_* variables should contain a link to a tutorial, David Waroquiers 090831." What should I do with this sentence?

wvl_hgrid => Not in the list of gstate ! should'nt it be the case ? (as for other wvl_* variables)