The ABINIT Users Mailing List ( CLOSED )

[Rob <spamrefuse@yahoo.com>]

Josef Zwanziger wrote:
>
> The "extra" bytes are due to the Fortran write statements.
> The format of the file is just what the documentation says,
> but it's easier to understand in terms of the Fortran read/write
> statements than byte counts.

I don't agree with this. I have analysed binary files before, either
from Fortran or from C. Moreover, the documentation explains
the header structure in terms of the Fortran write statement !
For example:

write(unit=header) codvsn,headform,fform
write(unit=header) bantot,date,intxc,ixc,natom,ngfft(1:3),&
& nkpt,nspden,nspinor,nsppol,nsym,npsp,ntypat,occopt,pertcase,usepaw,&
& ecut,ecutdg,ecutsm,ecut_eff,qptn(1:3),rprimd(1:3,1:3),stmbias,tphysel,tsmear

The first one, "codvsn", is a 6 byte long string. The rest are
integers; all the integer variables, of course, have the same
byte length (they are all of the same type).

I assume an integer is the usual 4 bytes.
This assumption is kind of confirmed by analysing the
density files from Tutorial 1.3 for the hydrogen molecule.
See my analysis of a "hexdump -C density-file":

  http://skku.homeip.net/~lahaye/abinit/DensityFile.jpg

In this analysis, I clearly recognize the 'date' integer, and
therefore before the 'date' variable, there are more integers
than is documented:
where the documentation specifies only two integers
(headform and fform), I see 4 integers (57, 52, 14, and 228).

Hence I think the documentation is incomplete here!

> If you want something more device-independent, you can
> use the netcdf option.

I want to access the binary files directly, without cut3d.

> All the output variables (usepaw etc) are fully documented
> in the "input variables" documentation pages.

I can't find:
  headform, fform, bantot, pertcase, ecutdg, ecut_eff


Rob.