DFT Energy failure during calculation
Clicked A Few Times
1:03:39 AM PST - Sun, Mar 12th 2017
Hi.
I am a novice in this field and having trouble with negatively charged(-1) system...
Here is a input file command line

Contents

===================================================

ECHO
START
TITLE "amino -1"
BASIS
 * library 6-31+G
END
CHARGE -1
DFT
 XC b3lyp

 ODFT

 MULT 2

 NOIO
END
PROPERTY
 MULLIKEN

 DIPOLE
END
GEOMETRY
...
END
TASK DFT energy
TASK DFT property
===================================================

The geometry is made of fluorinated carbon molecule with amino bond.
  • FYI: I run a calculation with same geometry but without charge (CHARGE 0, XC b3lyp, ODFT, NOIO) and it went fine.
Below is error comment, It seems like issues related to MPI...
I would be really appreciated if you help me how to solve this issue...


Below is error comment
===================================================
Parallel integral file used  567253 records with       0 large values

------------------------------------------------------------------------

dft energy failed                                                                                  0

------------------------------------------------------------------------

------------------------------------------------------------------------

 current input line :

  306: TASK DFT energy

------------------------------------------------------------------------

------------------------------------------------------------------------

This type of error is most commonly associated with calculations not reaching convergence criteria

------------------------------------------------------------------------

For more information see the NWChem manual at http://nwchemgit.github.io/index.php/NWChem_Documentation



For further details see manual section:                                                                               $
0:dft energy failed:Received an Error in Communication
application called MPI_Abort(comm=0x84000000, -1) - process 0
===================================================



Also, here are conditions I used when I first build nwchem.
===================================================
  • OS :SUSE Linux Enterprise Server 11 SP3 (x86_64)
  • intel-mpi/5.1.3.210
  • intel-mkl/17.0.0.098
  • gnu/4.8.5
  • intel/16.0.4.258

export NWCHEM_TOP=$HOME/nwchem
export NWCHEM_TARGET=LINUX64
export NWCHEM_MODULES=all
export USE_MPI=y
export USE_MPIF=y
export USE_MPIF4=y
export MPI_LOC=/home/app/intel/impi/5.1.3.210
export MPI_LIB=/home/app/intel/impi/5.1.3.210/lib64
export MPI_INCLUDE=/home/app/intel/impi/5.1.3.210/include64
export LIBMPI="-I/home/app/intel/impi/5.1.3.210/intel64/include -L/home/app/intel/impi/5.1.3.210/intel64/lib/release_mt\
-L/home/app/intel/impi/5.1.3.210/intel64/lib -Xlinker --enable-new-dtags -Xlinker -rpath -Xlinker /home/app/intel/impi\
/5.1.3.210/intel64/lib/release_mt -Xlinker -rpath -Xlinker /home/app/intel/impi/5.1.3.210/intel64/lib -Xlinker -rpath -\
Xlinker /opt/intel/mpi-rt/5.0/intel64/lib/release_mt -Xlinker -rpath -Xlinker /opt/intel/mpi-rt/5.0/intel64/lib -lmpifo\
rt -lmpi -lmpigi -ldl -lrt -lpthread"
export BLASOPT="-L/home/app/intel/compliers_and_libraries/mkl -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -liomp5 -l\pthread"

===================================================

Forum Regular
5:52:10 AM PDT - Mon, Mar 13th 2017
Your error message means that the SCF procedure ran out of cycles before convergence was reached. Try added the keyword

maxiter n

to your DFT input block, where n is some integer larger than 30 (the default value)

Clicked A Few Times
5:18:48 PM PDT - Fri, Mar 17th 2017
Quote:Sean Mar 13th 5:52 am
Your error message means that the SCF procedure ran out of cycles before convergence was reached. Try added the keyword

maxiter n

to your DFT input block, where n is some integer larger than 30 (the default value)


Thanks! I'll try.

Clicked A Few Times
10:15:13 AM PST - Thu, Dec 27th 2018
Quote:Sean Mar 13th 12:52 pm
Your error message means that the SCF procedure ran out of cycles before convergence was reached. Try added the keyword

maxiter n

to your DFT input block, where n is some integer larger than 30 (the default value)


Dear sean,

I have the same problem, while (iteration 1200) has been used in this case!
and after I see your advice I tried (iteration 2200) but still these appears  :

dft energy failed 0
------------------------------------------------------------------------

------------------------------------------------------------------------

 current input line : 

   73: task dft

------------------------------------------------------------------------

------------------------------------------------------------------------

This type of error is most commonly associated with calculations not reaching convergence criteria

------------------------------------------------------------------------

would U plz advice me about

Best Regards,

Forum Regular
9:27:13 AM PST - Fri, Dec 28th 2018
Simply increasing the number of iterations is only a crude way to try to achieve convergence; if your calculation hasn't converged after a couple hundred iterations, it is not likely to converge with the current settings. Achieving convergence in SCF calculations is a bit of an art and very often system dependent. Assuming that you do not have an error in your input (e.g. a bad geometry, unrealistic charge, etc.), some useful approaches could be to converge the calculation with a smaller basis set and then project that solution on to the basis set that is desired (https://github.com/nwchemgit/nwchem/wiki/Hartree-Fock-Theory-for-Molecules#vectors----inpu...) or play with the convergence controls to modify when damping, DIIS, and level-shifting are used (https://github.com/nwchemgit/nwchem/wiki/Density-Functional-Theory-for-Molecules#convergen...). Which specific combinations of modifications will work (if any will work) will depend on the system being studied and the behavior of the SCF cycles.


Forum >> NWChem's corner >> Running NWChem