[Pw_forum] QE 5.1: "error in DGETRF" restarting vc-relax calculation

[Lorenzo Paulatto]

Dear V,
there is indeed a bug, I think I have fixed it in the SVN version. If 
you just want a quick work-around, replace PW/src/input.f90 with the 
file attached and recompile.

I have only done very basic testing on this fix, so please come back to 
us if you find any strange behaviour.

best regards and thank you for reporting


On 09/06/2014 08:04 PM, V.Sokolov wrote:
> Dear Quantum Espresso developers and users,
>
> There is a problem with QE 5.1.
> If I try to restart vc-relax cell and atomic positions optimization,
> pw.x (QE 5.1) terminates with the following message:
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>        Error in routine invmat (1):
>        error in DGETRF
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>
> The same task is restarted without problems in QE 5.0.3.
>
> As far as I can understand, the error occurs in
>          CALL invmat (3, rot, overlap, value)
> (sym_base.f90, line 240) since all elements of the rot(3,3) array are zero.
>
>
> Help please...
> Thank you in advance.
>

-- 
Dr. Lorenzo Paulatto
IdR @ IMPMC -- CNRS & Université Paris 6
+33 (0)1 44 275 084 / skype: paulatz
http://www.impmc.upmc.fr/~paulatto/
23-24/4é16 Boîte courrier 115, 4 place Jussieu 75252 Paris Cédex 05

-------------- next part --------------
A non-text attachment was scrubbed...
Name: input.f90
Type: text/x-fortran
Size: 50151 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20140908/2874e998/attachment.bin>