[gpaw-users] Error in transport calculation

[jingzhe Chen]

Hi Brandon,

          In transport calculation , the lead and the scattering region
should
keep the same k sampling in x, y directions. For the leads, the code does
 not judge the k point one by one to determine if it is a different k point
in
x-y plane, so the number of k points in x-y plane is simply got by total k
number divided by the k point along z axis.  Therefore so far, the k number
in x, y should better be even,  or at least avoid two odd numbers in
the pl_kpts. Because after the k point reduction due to time reversal
symmetry,
the simple way above does not work.

       usesymm=None is always safe for any case.

       Best.
       Jingzhe




2012/2/22 Brandon Cook <brandon.g.cook at vanderbilt.edu>

> Hi all,
>
> I am doing some transport calculations with GPAW and I am running into
> errors with certain combinations of k-points and number of cores used.
> For example if I use kpts=(7,1,1) and pl_kpts=(7,1,13) (z is the
> transport axis) the calculation works fine with 4 cores, but fails
> with 2. This doesn't seem to depend on what system I calculate. This
> is with: GPAW 0.9.0.8913 and ASE 3.5.2.2505
>
> Are there certain numbers of k-points that should be avoided?
>
> This is error message I get:
>
> Traceback (most recent call last):
>  File "au.py", line 39, in <module>
>    t.calculate_iv()
>  File "/home/brandon/src/gpaw/gpaw/transport/calculator.py", line
> 2348, in calculate_iv
>    self.negf_prepare()
>  File "/home/brandon/src/gpaw/gpaw/transport/calculator.py", line
> 1185, in negf_prepare
>    self.initialize_transport()
>  File "/home/brandon/src/gpaw/gpaw/transport/calculator.py", line
> 399, in initialize_transport
>    self.calculate_leads()
>  File "/home/brandon/src/gpaw/gpaw/transport/calculator.py", line
> 339, in calculate_leads
>    self.collect_leads_matrices(atoms.calc, i)
>  File "/home/brandon/src/gpaw/gpaw/transport/calculator.py", line
> 1070, in collect_leads_matrices
>    self.lead_hsd[l].reset(0, pk, sl_pkmm[pk], 'S', init=True)
> IndexError: index out of bounds
> GPAW CLEANUP (node 0): <type 'exceptions.IndexError'> occurred.
> Calling MPI_Abort!
>
>
> Thanks,
> Brandon Cook
> _______________________________________________
> gpaw-users mailing list
> gpaw-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://listserv.fysik.dtu.dk/pipermail/gpaw-users/attachments/20120222/31727080/attachment.html