This thread is locked.Only browsing is available.
Top Page > Browsing
ERROR: Lapack routine DSTEQR failed
Date: 2017/03/23 20:05
Name: Pui-Wai (Leo) Ma   <>

Dear Prof. Ozaki,

I encountered another issue. There is an error listed below. When I was using "mpirun -np 60 openmx", there was no error. When I was using "mpirun -np 100 openmx", following error appears at the beginning of second iteration step of geometric relaxation using DIIS. I am suspecting it is an error from ELPA.

SCF calculation at MD = 2

<MD= 2> Calculation of the overlap matrix
<MD= 2> Calculation of the nonlocal matrix
<MD= 2> Calculation of the VNA projector matrix

******************* MD= 2 SCF= 1 *******************
<Restart> Found restart files
<Poisson> Poisson's equation using FFT...
<Band> Solving the eigenvalue problem...
KGrids1: -0.42857 -0.28571 -0.14286 0.00000 0.14286 0.28572 0.42857
KGrids2: -0.42857 -0.28572 -0.14286 -0.00000 0.14286 0.28571 0.42857
KGrids3: -0.42857 -0.28571 -0.14286 0.00000 0.14286 0.28572 0.42857
ERROR: Lapack routine DSTEQR failed, info= 503, Aborting!

Best regards,
Leo Ma

Dr. Pui-Wai (Leo) Ma
Senior Scientist
Culham Centre for Fusion Energy (CCFE)
Abingdon, Oxfordshire OX14 3DB, UK
Tel. +44 (0)1235 465194
Fax. +44 (0)1235 466435
Page: [1]

Re: ERROR: Lapack routine DSTEQR failed ( No.1 )
Date: 2017/03/24 01:56
Name: Pui-Wai (Leo) Ma  <>

Dear Prof. Ozaki,

It seems the error disappear after recompiling the program using -O2 instead of -O3.

Best regards,
Leo Ma
Re: ERROR: Lapack routine DSTEQR failed ( No.2 )
Date: 2017/03/26 18:27
Name: T. Ozaki


I wonder that the issue is caused by an eigenvalue solver, and that
another library such as ACML will resolve the issue as well as setting
the compiler option to -O2.



Page: [1]