Error: "LAPACK: Routine ZPOTRF failed" "call to ZHEGV failed

Vasp transition state theory tools

Moderator: moderators

Post Reply
Danpho
Posts: 1
Joined: Fri May 08, 2015 9:09 pm

Error: "LAPACK: Routine ZPOTRF failed" "call to ZHEGV failed

Post by Danpho »

Is there any solution for "LAPACK: Routine ZPOTRF failed" and "call to ZHEGV failed"?

My simulation died and received many following failure or Warning message. Any suggestion will be highly appreciated.


I noticed the error message show something periodically. (Pay attention to the last digit number. it is from 1 to 99, and following 5 **, and then count from 1 again,."

********************Error message or Warning message ***********************
.....

WARNING in EDDRMM: call to ZHEGV failed, returncode = 3 2 1
....
WARNING in EDDRMM: call to ZHEGV failed, returncode = 3 2 99
WARNING in EDDRMM: call to ZHEGV failed, returncode = 3 2 **
WARNING in EDDRMM: call to ZHEGV failed, returncode = 3 2 **
WARNING in EDDRMM: call to ZHEGV failed, returncode = 3 2 **
WARNING in EDDRMM: call to ZHEGV failed, returncode = 3 2 **
WARNING in EDDRMM: call to ZHEGV failed, returncode = 3 2 **

LAPACK: Routine ZPOTRF failed! 1 1 1
LAPACK: Routine ZPOTRF failed!
LAPACK: Routine ZPOTRF failed! 1 1 1
LAPACK: Routine ZPOTRF failed!
LAPACK: Routine ZPOTRF failed! 1 1 1
LAPACK: Routine ZPOTRF failed! 1 1 1
LAPACK: Routine ZPOTRF failed! 1 1 1
LAPACK: Routine ZPOTRF failed!
LAPACK: Routine ZPOTRF failed! 1 1 1
LAPACK: Routine ZPOTRF failed!
LAPACK: Routine ZPOTRF failed! 1 1 1

*********************************** INCAR **************************************

System = Al47CrO72

!Basic setting
ISTART = 0 ! How to start. 0: NEW; 1: CONT; 2 SAMECUT
ISMEAR = 0 ! -4:tet, -1:Fermi, 0: Gaussian.
SIGMA = 0.05 ! Broadening in eV of smearing method.dft=0.1
PREC = Accurate ! low, medium, normal are other options. Use suitable one.
ENCUT = 520 eV ! Kinetic cutoff of plane wave. ENCUT=max(ENMAX*(1+30%)).
ISPIN =2
MAXMIX = 40

!Extended setting
EDIFF = 1.0E-5 ! Enegy difference covergence limit for electronic optimization.org=1.E-5
EDIFFG = -1.0E-2 ! Enegy difference covergence limit for ionic optimization.org=-1E-3

!ion opt. practicle setting SMASS=0.5-1 small POTIM=0.05-0.1
!Increase POTIM by 20% until relaxation runs diverge;
!fix POTIM to the largest value;
!try a set of different SMASS until convergence is fastest.
IBRION = 2 ! 0 for MD,1 org best(DISS)[POTIM,NFREE],2(CG)[POTIM] for diff rlx problems.
!IBRION=3(damped MD)[POTIM,SMASS].-1 use RMM-DIIS algorithm for ions.

NSW = 500 ! Total number of ionic steps.
ISIF = 3 ! 2 and 4 ionic, 7 volume and 3 both.

LREAL = Auto ! Projection on real space. use FALSE (default) for reciprocal space.
NPAR = 4 ! Best sqrt of NCPUs used. should be >= NCPUs/32.
ALGO = Fast ! default is Normal.

LCHARG = F ! No writting in CHG and CHGCAR files
LWAVE = F ! No writting in WAVECAR file

ankittk
Posts: 1
Joined: Thu Aug 06, 2020 5:22 am

Re: Error: "LAPACK: Routine ZPOTRF failed" "call to ZHEGV failed

Post by ankittk »

Hello, I was getting the same error with correct structure for a charged graphene system. I tried increasing the POTIM value and changing ALGO to Normal. However, those did not work at least in my case. Reducing the POTIM value from 0.5 to 0.001 resolved the error, but it distorted my structure. I was using ISIF=3 to relax ions and cell shape and volume. So, my guess is that changing POTIM to 0.001 would make the steps too small which will cause the system to crash in some cases. I appreciate any help on this.
https://www.techikids.com

Post Reply