[CP2K-user] convergence problem

[ma...@gmail.com]

Hi,

I think 1E-06 for "MAX_DR MAX_FORCE   RMS_DR    RMS_FORCE are too tight".  
1/3 of the default value of cp2k should be tight enough.

Regards,
Hongyang
在2021年5月31日星期一 UTC+10 上午7:09:07<alberto> 写道：

> Dear cp2k users,
> I am doing the geometry optimization of a configuration (supercell 
> containing 4 molecules of Fe(CO)5 with symmetry C2/c and Z=4) but the 
> energy is locked on a value and is not decreasing. 
> How can I solve this problem? 
> Should I use a higher CUTOFF value?
>
>
> *******************************************************************************
>  ***                 BRENT   - NUMBER OF ENERGY EVALUATIONS :       1      
>   ***
>
>  *******************************************************************************
>
>  --------  Informations at step =   186 ------------
>   Optimization Method        =                   SD
>   Total Energy               =      -927.5280017548
>   Real energy change         =         0.0000000000
>   Decrease in energy         =                   NO
>   Used time                  =             2410.831
>
>   Convergence check :
>   Max. step size             =         0.0000000000
>   Conv. limit for step size  =         0.0000010000
>   Convergence in step size   =                  YES
>   RMS step size              =         0.0000000000
>   Conv. limit for RMS step   =         0.0000010000
>   Convergence in RMS step    =                  YES
>   Max. gradient              =         0.0089297310
>   Conv. limit for gradients  =         0.0000010000
>   Conv. for gradients        =                   NO
>   RMS gradient               =         0.0016280805
>   Conv. limit for RMS grad.  =         0.0000010000
>   Conv. for gradients        =                   NO
>  ---------------------------------------------------
> I attach the input
> Thanks
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210530/e7af2120/attachment.htm>