Exceeding memory of input-upper-limit for a heavy noble gas

[Scott Muller]

Hello,

I'm trying to compare values for the ground state energy for the noble gas xenon using different basis sets and functionals. I was able to perform calculations using a variety of functionals using the cc-PVDZ, cc-PVTZ, and cc-PVQZ basis sets without issue. I'm currently working with the cc-PV5Z basis set and have run into a wall. Only the coupled cluster functional, CCSD(t), was able to successfully perform calculations for Xe using the cc-PV5Z basis set. Other functionals such as Hartree-Fock, LDA, and B3LYP, had no issue with the previous basis sets, but all have the same issue here. After successfully performing orbital analysis, I get the following error:

 A block with only 1 ij-pair, 1 kl-pair and 1 quartet

 requires more memory than the input-upper-limit

 input limit =               6000000  requested =              14273791

      Change the value of LIMXMEM

      by adding this input line

      set int:txs:limxmem               14273791

 texas:blksizer                   0

If I make the recommended change and add "set int:txs:limxmem 14273791" to my input script, I get a more generic error:

  texas integral default override: limxmem =              14273791

  texas integral default override: limxmem =              14273791

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

=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES

=   PID 27651 RUNNING AT node009

=   EXIT CODE: 11

=   CLEANING UP REMAINING PROCESSES

=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES

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

I don't know why the functional CCSD(t) works when the others don't. I assume it's because this functional is more tailored to using advanced correlation consistent basis sets than other functionals. Is there a way to modify an input script to allow functionals other than CCSD(t) to use the cc-PV5Z basis sets? Or is there another way forward in computing these calculations?

Thanks,

Scott

I kept getting an error in the forum when I tried to attach my input/output files, so I'm copying my input file as text below. I'm happy to answer any questions about my output file.

start xeCalc

echo

memory stack 500 mb heap 100 mb global 1300 mb

geometry units angstroms

Xe     0.0000000000    0.0000000000    0.0000000000

end

basis spherical

    Xe library cc-pv5z-pp

end

ecp

    Xe library cc-pv5z-pp

end

scf

    thresh 1.0e-10

    tol2e 1.0e-10

    singlet

    rhf

end

driver

    xyz pos

    tight

end

task scf optimize

[Edoardo Aprà]

If you just replace "task scf optimize" with "task scf"  the run will complete.

Since it does not make sense to perform the geometry optimization of a single atom, "task scf" is all you need

[Scott Muller]

Thanks! That worked great!

I'm now trying to find ground state energy for molecular iodine (two iodine atoms) using the cc-PV5Z basis set for different functionals. I would like to be able to use optimize so that I can get ground state energy with an optimized bond length, but then I run into the issue discussed above. Do you know if there is a way to make this work?

Thanks again,

Scott

[Edoardo Aprà]

You cant try to edit the input file by having the following two last lines

set int:txs:limxmem               15000000

task scf optimize

[Scott Muller]

When I try that, I get the following error in my output file (with surrounding lines included for context):

~~~~~~~~~

 Parallel integral file used     692 records with       0 large values

                              NWChem Gradients Module

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

  wavefunction    =   RHF

  Using symmetry

  texas integral default override: limxmem =              15000000

  texas integral default override: limxmem =              15000000

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

=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES

=   PID 22303 RUNNING AT node185

=   EXIT CODE: 11

=   CLEANING UP REMAINING PROCESSES

=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES

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

   Intel(R) MPI Library troubleshooting guide:

      https://software.intel.com/node/561764

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

~~~~~~~~~~~

It seems to be recognizing the command

set int:txs:limxmem               15000000

but for whatever reason, it can't work with it. Any other recommendations?

Thanks again for your help,

-Scott

[Edoardo Aprà]

"EXIT CODE: 11" means that your run has stopped for a Segmentation fault

This might be caused by either a variety of factors: incorrect NWChem installation, bug in the NWChem code or bug in the third party software libraries used (e.g. Intel MPI)

However, it is hard to tell what is the root cause of your problem without having a look at the complete input file and by knowing the full details of your NWChem installation