EDA analysis with psi4


#1

Hi!
I am very new to psi4. I tried to do the EDA (energy decomposition analysis) with psi4 for my water cluster system. Unfortunately, my calculation terminated with the following error. I want to know is it due to memory problem or some other issues?
Although, it worked well for a water dimer system.
Thanks in advance

Ind20,r (A<-B) = -0.000023373068 [Eh]
Ind20,r (B<-A) = -0.000125298200 [Eh]
Ind20,r = -0.000148671268 [Eh]
Exch-Ind20,r (A<-B) = 0.000013710825 [Eh]
Exch-Ind20,r (B<-A) = 0.000129146750 [Eh]
Exch-Ind20,r = 0.000142857575 [Eh]
An error has occurred python-side. Traceback (most recent call last):

File “”, line 167, in

File “/home/miniconda/share/psi4/python/driver.py”, line 446, in energy
wfn = procedures[‘energy’][lowername](lowername, molecule=molecule, **kwargs)

File “/homeI/miniconda/share/psi4/python/procedures/proc.py”, line 2920, in run_sapt
e_sapt = psi4.sapt(dimer_wfn, monomerA_wfn, monomerB_wfn)

RuntimeError:
Fatal Error: PSIO Error
Error occurred in file: /scratch/cdsgroup/conda-builds/work/src/lib/libpsio/error.cc on line: 128
The most recent 5 function calls were:

psi::PsiException::PsiException(std::string, char const*, int)
psi::psio_error(unsigned int, unsigned int)
psi::PSIO::rw(unsigned int, char*, psi::psio_address, unsigned long, int)
psi::sapt::SAPT::zero_disk(int, char const*, int, int)
psi::sapt::SAPT0::q12()


#2

Could you have run out of scratch space? I can try running the input if you post/send it.


#3

Hi lori,
I asked for the input file from her and it runs fine with my psi4 version. As you pointed out, this must be
a scratch issue. Also, the user is using psi4 version 1.0.54 but I think that should not be a problem.


#4

Thanks, @ashutosh. Agreed, there were some real changes to SAPT printing and psivars just before 1.0 but no real SAPT changes between 1.0.54 and current development head.


#5

Thanks for your suggestions. I have changed my scratch directory and now it is running well.

Thanks :slight_smile:


#6

Hi everyone!!
Greetings of the day!!
I am trying to perform SAPT0 analysis for Metal organic Frameworks (MOF). Unfortunately, I am getting the following error. This calculation I performed at STO3G taking 12GB memory.

==> Integral Setup <==

Traceback (most recent call last):
File “/home/name/miniconda/bin/psi4”, line 269, in
exec(content)
File “”, line 188, in
File “/home/name/miniconda/lib//python2.7/site-packages/psi4/driver/driver.py”, line 492, in energy
wfn = procedures[‘energy’][lowername](lowername, molecule=molecule, **kwargs)
File “/home/name/miniconda/lib//python2.7/site-packages/psi4/driver/procrouting/proc.py”, line 3343, in run_sapt
dimer_wfn = scf_helper(‘RHF’, molecule=sapt_dimer, **kwargs)
File “/home/name/miniconda/lib//python2.7/site-packages/psi4/driver/procrouting/proc.py”, line 1363, in scf_helper
e_scf = scf_wfn.compute_energy()

RuntimeError:
Fatal Error: SCF::DF: Disk based algorithm requires 2 (A|B) fitting metrics and an (A|mn) chunk on core.
This is 2Q^2 + QNP doubles, where Q is the auxiliary basis size, N is the
primary basis size, and P is the maximum number of functions in a primary shell.
For this problem, that is 1723993440 bytes before taxes,2462847771 bytes after taxes.

Error occurred in file: /scratch/psilocaluser/conda-builds/psi4-multiout_1532493090788/work/psi4/src/psi4/libfock/DiskDFJK.cc on line: 727
The most recent 5 function calls were:

psi::DiskDFJK::preiterations()

       I tried it by varying memory and basis sets also. 

Please suggest me the reasons of the termination of my calculations.
Thanks in advance.


#7

What you’re reporting is strongly reminiscent of another report of SAPT giving bad memory estimates (https://github.com/psi4/psi4/issues/1361).

We fixed a memory bug in SAPT (https://github.com/psi4/psi4/pull/1398) just two weeks ago, and that might have solved this problem. If it hasn’t, we have another bug.

The next step is to test how your computation runs on the developer version of Psi. Do you have Psi conda installed? If so, conda update psi4 -c psi4/label/dev should update you to the developer version of Psi.


#8

Thank you very much for your reply.

Yes, conda is installed in my machine.
SAPT0 calculation ran successfully for smaller molecules, however not for the larger system.

Thanks in advance.


#9

And, just to be clear, this is on the developer version of Psi4?

The output file should start like:

    -----------------------------------------------------------------------
          Psi4: An Open-Source Ab Initio Electronic Structure Package
                               Psi4 1.3a2.dev229 

                         Git: Rev {master} 16b8532 


    R. M. Parrish, L. A. Burns, D. G. A. Smith, A. C. Simmonett,
    A. E. DePrince III, E. G. Hohenstein, U. Bozkaya, A. Yu. Sokolov,
    R. Di Remigio, R. M. Richard, J. F. Gonthier, A. M. James,
    H. R. McAlexander, A. Kumar, M. Saitow, X. Wang, B. P. Pritchard,
    P. Verma, H. F. Schaefer III, K. Patkowski, R. A. King, E. F. Valeev,
    F. A. Evangelista, J. M. Turney, T. D. Crawford, and C. D. Sherrill,
    J. Chem. Theory Comput. 13(7) pp 3185--3197 (2017).
    (doi: 10.1021/acs.jctc.7b00174)


                         Additional Contributions by
    P. Kraus, H. Kruse, M. H. Lechner, M. C. Schieber, and R. A. Shaw 

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

What do you have at the start of your output file? Let’s make sure that you really are using the developer version.


#10

Hi!!
At the start of my output file, I am getting this:


      Psi4: An Open-Source Ab Initio Electronic Structure Package
                           Psi4 1.2.1 release

                     Git: Rev {HEAD} 406f4de


R. M. Parrish, L. A. Burns, D. G. A. Smith, A. C. Simmonett,
A. E. DePrince III, E. G. Hohenstein, U. Bozkaya, A. Yu. Sokolov,
R. Di Remigio, R. M. Richard, J. F. Gonthier, A. M. James,
H. R. McAlexander, A. Kumar, M. Saitow, X. Wang, B. P. Pritchard,
P. Verma, H. F. Schaefer III, K. Patkowski, R. A. King, E. F. Valeev,
F. A. Evangelista, J. M. Turney, T. D. Crawford, and C. D. Sherrill,
J. Chem. Theory Comput. 13(7) pp 3185--3197 (2017).
(doi: 10.1021/acs.jctc.7b00174)


                     Additional Contributions by
P. Kraus, H. Kruse, M. H. Lechner, M. C. Schieber, and R. A. Shaw

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

Thanks in advance.


#11

You are not using the developer version of Psi. The changes we made that might fix your input file are very recent and not in the 1.2.1 release.

You need to do one of the following:

  • Conda install the developer version of Psi using the command from my previous post. It is not enough just to have conda installed on your machine.
  • Give us your input file, so devs can run it ourselves
  • Wait until we have the next 1.3 alpha build out, so you can use that. It should be released in a week or two.

#12

Thank you very much for your suggestions. I will also try it with new versions.
The input of the complex is attached herewith.MC-8H2.dat (9.3 KB)

Thank you again.


#13

My cluster is down for maintenance at the moment, but I’ll try your input on an up-to-date version of Psi when it comes back up. Ping me if I don’t post again within a week.


#14

Sure and thank you very much for kind responses. :slight_smile:


#15

Hi!!
Have you tried SAPT0 calculations for my input files?
Thank you in advance.


#16

Using a more recent version of Psi, your input file fails with “not enough memory” during the Exch-Disp20 computation. In your report, the HF doesn’t even finish, so this is progress.

I’m not familiar enough with the memory limits of SAPT to definitively say whether this indicates an inefficiency with the code or whether a system of this size (952 electrons, 716 basis functions) just need more memory.