Error in SAPT calculation

SAPT
1

Dear PSI4 community,
I encountered this error in SAPT calculation.

Printing out the relevant lines from the Psithon → Python processed input file:
mol.set_basis_by_symbol(“Cl”, “jun-cc-pvtz”, role=role)
mol.set_basis_by_label(“H1”, “jun-cc-pvtz”, role=role)
return basstrings
qcdb.libmintsbasisset.basishorde[‘ANONYMOUSC62A7CA2’] = basisspec_psi4_yo__anonymousc62a7ca2
core.set_global_option(“BASIS”, “anonymousc62a7ca2”)
→ energy(‘sapt0’)

Psi4 stopped on: Monday, 08 June 2020 03:46PM
Psi4 wall time for execution: 0:58:33.56

*** Psi4 encountered an error. Buy a developer more coffee!

It seems to be something related to basis sets. My calculation used mixed basis sets of 6-31G(d) and jun-cc-pvtz. The input file is shown below.

memory 200 GB

molecule dimer {
0 1
N 0.00900000 -0.00800000 -3.91400000
(other atoms)

-1 1
Cl 0.01000000 -0.00400000 0.50100000
(other atoms)
}

basis {
assign DZ
assign H 6-31G(d)
assign C 6-31G(d)
assign N 6-31G(d)
assign O 6-31G(d)
assign Cl jun-cc-pvtz
assign H1 jun-cc-pvtz
}

energy(‘sapt0’)

Is there any wrong? Thank you very much for your help!

2

First, please use backticks (```) rather than blocks (>) when posting input files.

Second, can you show us the actual error message is? There should have been more information, before the lines you showed us, showing the issue. All I can tell from the “error” you gave is that it occurred during a SAPT0 computation.

3

Thank your very much for your reply.
I’m sorry for the disturbance made in the topic. More information has been provided in files attached to the E-mail.

发件人：Jonathon Misiewicz via Psi4 – Open-Source Quantum Chemistry noreply@forum.psicode.org
收件人：jwxu_njnu@sina.com
主题：[Psi4_–Open-Source_Quantum_Chemistry][SAPT]_Error_in_SAPT_calculation
日期：2020年06月13日 02点09分

| jmisiewicz
June 12 |

  • | - |

First, please use backticks (```) rather than blocks (>) when posting input files.

Second, can you show us the actual error message is? There should have been more information, before the lines you showed us, showing the issue. All I can tell from the “error” you gave is that it occurred during a SAPT0 computation.

(Attachment chloridecage.inp is missing)

(Attachment chloridecage.out is missing)

4

The input and output files didn’t appear for some reason. Can you please try to upload those again?

5

inp
inp365×665 9.3 KB

6

out
out1085×560 20.9 KB

7

I guess .inp and .out files are not supported.
Figures have been uploaded instead.

8

Well, knowing that this is a PSIO error is a start… @hokru, does this look like a scratch issue to you?

As for .inp and .out files, just change the file extensions to .txt, and you should be able to upload them.

9

Possibly, yes.

@jwxu Do you specify a scratch directory or use the the default (/tmp)?
Also, is it a local directory or part of the parallel file system?

If you upload the output, it could help us identify the issue (use .txt ending)

10

inp.txt (9.0 KB) out.txt (214.5 KB)

11

I run PSI4 on my school’s public cluster. The manager specify a local scratch directory for me.

12

This rather seems an issue with RAM memory and not disk.
These large systems can have sometimes unexpected grows in matrices that are not always caught.
We’ll have to look at the code.

From experience 40GB may not be enough for this calculation. (Also best to use GiB keyword).
How much memory do you have? It seems to work for me specifying 100 GiB.

13

I don’t know the maximum memory but it seems enough for 100GB.
I’ll try 100GiB and later tell you the result.
Thank you for your reply.

发件人：Holger Kruse via Psi4 – Open-Source Quantum Chemistry noreply@forum.psicode.org
收件人：jwxu_njnu@sina.com
主题：[Psi4_–Open-Source_Quantum_Chemistry][SAPT]_Error_in_SAPT_calculation
日期：2020年06月16日 15点15分

| hokru
June 16 |

  • | - |

This rather seems an issue with RAM memory and not disk.
These large systems can have sometimes unexpected grows in matrices that are not always caught.
We’ll have to look at the code.

From experience 40GB may not be enough for this calculation. (Also best to use GiB keyword).
How much memory do you have? It seems to work for me specifying 100 GiB.

14

I’ve used 100 GiB in the input file, yet the same error still happened.

发件人：Holger Kruse via Psi4 – Open-Source Quantum Chemistry noreply@forum.psicode.org
收件人：jwxu_njnu@sina.com
主题：[Psi4_–Open-Source_Quantum_Chemistry][SAPT]_Error_in_SAPT_calculation
日期：2020年06月16日 15点15分

| hokru
June 16 |

  • | - |

This rather seems an issue with RAM memory and not disk.
These large systems can have sometimes unexpected grows in matrices that are not always caught.
We’ll have to look at the code.

From experience 40GB may not be enough for this calculation. (Also best to use GiB keyword).
How much memory do you have? It seems to work for me specifying 100 GiB.