LAPACK: Routine ZPOTRF failed! 1 1 1

[maria_dresser]

Hello,

I am trying to simulate a Cl2 molecule under pressure, and performing a relaxation calculation with PSTRESS = 10. I also performed the calculation with same input files except INCAR did not include PSTRESS, and the calculation completed without any issues. Now when I add PSTRESS, it completes one step and then says "LAPACK: Routine ZPOTRF failed! 1 1 1"

The POSCAR is
Cl2
1.00000000000000
33.0000000000000000 0.0000000000000000 0.0000000000000000
0.0000000000000000 33.0000000000000000 0.0000000000000000
0.0000000000000000 0.0000000000000000 33.0000000000000000
Cl
2
Direct
0.3885519999999971 0.0000000000000000 0.6040440000000018
0.6114480000000029 0.0000000000000000 0.3959559999999982

The INCAR is
ALGO = Normal
EDIFF = 0.0001
ENCUT = 520
IBRION = 2
ISIF = 3
ISMEAR = 0
ISPIN = 2
LASPH = True
LORBIT = 11
LREAL = .FALSE.
LWAVE = False
MAGMOM = 2*0.6
NELM = 100
NSW = 99
PREC = Accurate
SIGMA = 0.05
NCORE = 8
PSTRESS = 10

(I have tried both ALGO = Normal and ALGO = Fast, both give same error).

The KPOINTS:
K-Points
0
Gamma
1 1 1

Any help is greatly appreciated! Thank you.

[jonathan_lahnsteiner2]

Dear Maria Dresser,

I was checking your simulation.
First thing I recognized is that the Cl-Cl bond length of your Cl2 molecule is
10Angstroem which is very long. But this is not the reason for the error you get.

VASP has convergence issues during the structure optimization which is switched on
with IBRION=2. You can solve this issue by setting POTIM=0.01. This will reduce
the step width during the conjugate gradient algorithm.
wiki/index.php/POTIM
wiki/index.php/IBRION#IBRION.3D2:_ionic ... gorithm.29.
I hope this solves your issue.

All the best Jonathan

[maria_dresser]

Thanks so much for your response! I didn't intend for my bond lengths to be 10 angstroms, so I'm glad you caught that. I adjusted POTIM as you mentioned and this fixed the error for my lower pressure cases, but when I use PSTRESS=100 I get the same error (tried with both ALGO Fast and Normal). Any other ideas are greatly appreciated!

[jonathan_lahnsteiner2]

Dear Maria Dresser,

I checked your calculation with PSTRESS=100. I get the same error message as you
already mentioned.

Code: Select all

 -----------------------------------------------------------------------------
|                                                                             |
|     EEEEEEE  RRRRRR   RRRRRR   OOOOOOO  RRRRRR      ###     ###     ###     |
|     E        R     R  R     R  O     O  R     R     ###     ###     ###     |
|     E        R     R  R     R  O     O  R     R     ###     ###     ###     |
|     EEEEE    RRRRRR   RRRRRR   O     O  RRRRRR       #       #       #      |
|     E        R   R    R   R    O     O  R   R                               |
|     E        R    R   R    R   O     O  R    R      ###     ###     ###     |
|     EEEEEEE  R     R  R     R  OOOOOOO  R     R     ###     ###     ###     |
|                                                                             |
|     Orbital orthonormalization failed in the inversion of matrix            |
|     LAPACK: Routine ZPOTRF failed! kpoint: 1 spin: 1                        |
|                                                                             |
|     Possible solutions for this issue are:                                  |
|     - Please check whether the atoms are too close to each other.           |
|     - Decreasing POTIM might help during relaxations or MDs.                |
|     - Deleting CHGCAR or WAVECAR if incompatible with the structure.        |
|     - Try calculating without symmetrizing (ISYM = 0 or ISYM = -1).         |
|     - As last resort: using a different PAW for some of the atoms.          |
|                                                                             |
|       ---->  I REFUSE TO CONTINUE WITH THIS SICK JOB ... BYE!!! <----       |
|                                                                             |
 -----------------------------------------------------------------------------

Vasp also tries to give you some suggestions what you could try to fix the error. You see that vasp is suggesting to try to reduce/decrease POTIM during relaxation or MD's.
I reduced POTIM to 0.001 and the calculation converges. This works for the simulation box with the long Cl2 bond but also for a simulation with smaller Cl2 bond. Maybe also a slightly higher POTIM value might work.
I hope this helps and if you have further questions don't hesitate to ask.

All the best Jonathan