Page 1 of 1

WARNING in EDDRMM: call to ZHEGV failed

Posted: Thu May 28, 2020 7:18 pm
by solankikiran
I am doing a total energy calculation for FeHe system in Vasp 5.4.1 which i encounter the following error in the second step onwards of ionic relaxation.
WARNING in EDDRMM: call to ZHEGV failed, returncode = 6 3 197
I looked into the thread forum/viewtopic.php?f=3&t=214&p=215 however I am not able to figure out what would be the cause of error in my case.
Earlier in the same calculation I encountered the error DAV Sub space matrix is not Hermitian which I resolved by adding the tags POTIM PREC and IALGO
Please have a look at the input and output files attached.
Any suggestions as to how I can resolve this error?

Re: WARNING in EDDRMM: call to ZHEGV failed

Posted: Tue Jun 09, 2020 6:40 am
by ferenc_karsai
We cannot really offer support for such an old version of VASP.

I've tried the calculations with VASP 6.1.0 and there is no problem.
I suggest an upgrade of your VASP version.

Re: WARNING in EDDRMM: call to ZHEGV failed

Posted: Mon Jul 06, 2020 1:46 pm
by juergen.furthmueller
It's just a warning (no error). It's actually present in any version of VASP, even in vasp.6
(the only difference is maybe that different compiler optimizations / semantics may sometimes
trigger or sometimes prevent it). Just ignore ... . It occurs if you are very close to machine
accuracy in the RMM-DIIS scheme and single band iteration only results in "random" progress
anymore (due to machine accuracy limits). It says then indirectly to you "could not optimize
this band further" -- and the internal action taken is then simply NOT to update the wave
function further and to proceed with the next band(s). It confused already many people over
many generations and please, could somebody simply comment out these WRITE statements
in rmm-diis.F ... ? (somewhere around lines 510 and 630; "WRITE(IU?,219) IFAIL,ITER,N)")

Re: WARNING in EDDRMM: call to ZHEGV failed

Posted: Tue Jul 07, 2020 7:23 am
by martin.schlipf
I thought I had already done this. Thanks for pointing it out, I need to check if it didn't get merged into the main branch so that it will be included in the next version.