[CP2K-user] Bad condition number error

[Brian Day]

Hi everyone,

I've recently started to get the following error while running single point 
energy calculations on some systems. 

 SCF WAVEFUNCTION OPTIMIZATION

  Step     Update method      Time    Convergence         Total energy    
Change
  
------------------------------------------------------------------------------
     1 NoMix/Diag. 0.10E+00    3.8     0.75631948      -243.4503623748 
-2.43E+02
     2 Broy./Diag. 0.10E+00    4.7     0.36183064      -245.7284413793 
-2.28E+00

 *******************************************************************************
 *   ___                                                                    
   *
 *  /   \                                                                  
    *
 * [ABORT]                                                                  
   *
 *  \___/    Bad condition number R_COND = 0.000E+00 (smaller than the 
machine *
 *    |                             working precision)                      
   *
 *  O/|                                                                    
    *
 * /| |                                                                    
    *
 * / \                                                    
common/mathlib.F:710 *
 *******************************************************************************

Unfortunately, I do not seem to get this error in any consistent fashion, 
and I have yet to find a common thread as to when/why I get it. Moreover, 
there are typically two successful SCF calculations before the error is 
printed, as shown above. These results correspond to the file 
/NeutralGeo/Cu3(HIB)2-ReorgEnergy-wMetal-eagn-run.out, which I have 
attached. The general procedure for these calculations is as follows: 
Submit the "build-submit-reorg-single.bash" file, which creates one 
directory per charge state (i.e. Anionic/Cationic/NeutralGeo) and fills in 
the cp2k-reorg-*geo-template.slurm file. This file is submitted to our 
cluster, along with the template cp2k file, here reorg-BLYP-wD3.inp. The 
calculations it then runs are a geometry optimization for the corresponding 
charge state (i.e. GeoOpt for neutral state in NeutralGeo directory), a 
single point energy calculation using the same charge state and optimized 
geometry (this was for printing cube files, but I have that turned off now 
for the sake of figuring out this error), and then two more single point 
energy calculations for the anionic/cationic state, using the optimized 
neutral geometry. I had run many fo these before without issue, but after 
making some small changes to my reorg-BLYP-wD3.inp, this issue appeared. 
However, commenting out the changes does not seem to solve the issue. I am 
still working to debug this, but if anybody has some insight, that would be 
greatly appreciated. Thanks!

Best,
     Brian Day

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190204/83e41b61/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: NeutralGeo.zip
Type: application/zip
Size: 4005466 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190204/83e41b61/attachment.zip>