It will be easier to discover the
present file with the help of the tutorial
(Sorry, not yet available for mrgddb).
It is worthwhile to print this help file, for ease of reading.
The mrgddb code has the purpose to merge "transfer" DDBs
(that were generated from the ABINIT code) to make a complete DDB
that can be exploited by the Anaddb code.
Previously, it was also needed for the initialisation step of complete DDBs.
This is no longer required.
The input is very simple, and could be given directly at the screen,
or more conveniently, piped from a file.
The user should provide first the name of the new (output) DDB.
He/she should then give a short description (one line) of this new DDB
that will be created. This line will be printed at the beginning
of the DDB.
The user should then give the number of DDBs that will be merged, then
the whole set of filenames for the DDBs to be merged, one on each line.
--------------------------------------------------------------------
Is a merging code really useful ?
The ABINIT code in its present
version is only able to produce results for one q wavevector
for each dataset. A database for more than one q point can thus be created
using MRGDDB. Also, it is useful to be able to merge different DDBs
if they are produced independently on different machines.
-------------------------------------------------------------------
How does the merging code work ?
The DDBs are made of two parts. The first
part is a list of the parameters that were used to make the DDB,
and the second part lists the 2DTEs and 3DTEs.
The merging code will check if the following variables are exactly
the same in the different input DDB : natom,
ntypat, nband, acell, amu, ecut, ixc, lloc, ngfft, occ, rprim,
typat, xred, zion.
For nband and occ, the value of occopt is taken into account
(see abinis_help). If possible, MRGDDB will produce
a DDB with occopt=0.
In case of two different data sets, the code will print an error message and stop.
The code cannot merge two DDB that have been generated
using two different geometries
or convergence (ecut, ...) parameters.
The only exception is connected to the possibility to use Time-reversal
symmetry to decrease the number of special k points when the
wavevector of the perturbation is Gamma. In that case, the code
will merge the DDBs and put the largest set of k-points inside
the new DDB.
MRGDDB will copy the latest date of the transfer or current
DDB and copy it in the new DDB.
It will also take the less accurate tolwfr and copy it in the new DDB.
This ends the first part of the action of MRGDDB, namely
to compare the information of the two different DDB.
When the checking is done, MRGDDB will check the content
of the different data blocks and constitute the new DDB by
copying sequentially the non-identical blocks and merging the
identical blocks. In case two elements are identical, MRGDDB
copies the value of the transfer DDB.
(This latter property makes it easy to get rid of old, erroneous data
and put new, correct data in its place)
Finally, the summary of the blok content of the DDB is provided
at the end of the DDB file.