CCSDT calling EOM?

More
1 year 5 months ago #512 by bkwx97
CCSDT calling EOM? was created by bkwx97
Dear developers,

I'm attempting to run a CCSDT/cc-pvqz computation on n-pentane. I am using the mp2 natural orbitals approach to truncate the virtual space and make the computation feasible. I've noticed when I attempt to throw more memory at the problem the code calls the EOM code, and then crashes. What might be causing this? Attached are my input and output. I am using the Feb 4,2018 binary.
Attachments:

Please Log in or Create an account to join the conversation.

More
1 year 5 months ago #533 by rolik
Replied by rolik on topic CCSDT calling EOM?
Dear User,

I have tested your input file on system with 8Gbyte memory (Intel 14.0.3 compiler).

Let me to mention that the dfbasis_scf=auto keyword of your input file is missing from the output.

I tested the code with the dfbasis_set=auto in the previous weeks and found that it does not work. Latter this problem will be fixed.

As it turned out from your output file, the EOMCC issue is not related to the dfbasis_scf problem. We assumed that the problem is in the integral
transformation routine which prints the input of the mrcc.

To reproduce the problem I have tested the code without the dfbasis_scf=auto keyword. To fit into the memory CCSD calculations were performed instead of CCSDT. Using various basis sets (including cc-pvqz basis) and memory values I found no problem.

To check whether the problem is in the input of mrcc, could you post the first five lines of the fort.55 file?

Best wishes, Zoltán Rolik

Please Log in or Create an account to join the conversation.

More
1 year 5 months ago #547 by bkwx97
Replied by bkwx97 on topic CCSDT calling EOM?
Here's the first five lines of fort.55. I dropped the df_basis=auto command

224 32
1 3 1
3 1 2
3 4 1
1 2 3

Please Log in or Create an account to join the conversation.

More
1 year 4 months ago #559 by rolik
Replied by rolik on topic CCSDT calling EOM?
Thanks!

I have checked the number of basis functions (224). Both the value and the format look fine...

If the fort.55 and fort.56 are still in your working directory, you can try to
reproduce the error without starting from beginning:

1, set the PATH
2, run "goldstone > goldstone_log" (takes a second)
3, run "mrcc > mrcc_output" (NOT dmrcc!!!)

If the error is still there, could you post the content of the fort.56 file?

Best wishes, Zoltan Rolik

Please Log in or Create an account to join the conversation.

More
1 year 2 months ago - 1 year 2 months ago #585 by bkwx97
Replied by bkwx97 on topic CCSDT calling EOM?
It's been awhile since I looked at this.

I've ran into the same problem with ortho-xylene. I've updated to the 05-23-18 binary. I've attached my input and output as well as the results of fort.55,fort.56, as well as my goldstone_log and mrcc_log

I copied the fort.55 and fort.56 over. Here's the first line of my fort.55

154 42

I want to ask the ignorant question first. What do you mean by set the path? I have the path for where I have my mrcc binary set in my .bashrc if that's what you mean.

Here's the output of goldstone_log

Generation of CC equations in terms of H and T...
Generation of antisymmetrized Goldstone diagrams...
Number of diagrams in T^1 equations: 15
Number of diagrams in T^2 equations: 37
Number of diagrams in T^3 equations: 47
Translation of diagrams to factorized equations...
Optimizing intermediate calculation...
Optimizing right-hand EOMCC calculation...
Optimizing left-hand EOMCC calculation...
Optimizing density-matrix calculation...
Optimizing perturbed lambda equations...
Optimizing perturbed density calculation...
Number of floating-point operations per iteration step: 3.0481E+14
Probable CPU time per iteration step (hours): 3048.07
Required memory (Mbytes): 69097.6
Number of intermediates: 593
Number of intermediates to be stored: 183
Length of intermediate file (Mbytes): 160938.3

Here's the mrcc_output

**********************************************************************
CCSDT analytic second derivative calculation


OpenMP parallel version is running.
Number of CPUs: 32
Allocation of 1.0 Mbytes of memory...
Number of spinorbitals: 308
Number of alpha electrons: 21
Number of beta electrons: 21
Spin multiplicity: 1
z-component of spin: 0.0
Spatial symmetry: 1
Convergence criterion: 1.0E-06
Construction of occupation graphs...



Here's the output of my fort.56

3 1 0 0 1 0 0 1 0 1 1 1 0 0 0 6 0 0 0.000 0132000 1 0.97500000

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
Attachments:
Last edit: 1 year 2 months ago by bkwx97.

Please Log in or Create an account to join the conversation.

More
1 year 2 months ago #586 by kallay
Replied by kallay on topic CCSDT calling EOM?
Please edit ovirt.f, go to line 18484, change it to
write(56,'(18I6,F7.3,3I9,F12.8)')segedi(1:7),ize,
and recompile the code.

Best regards,
Mihaly Kallay

Please Log in or Create an account to join the conversation.

Time to create page: 0.023 seconds
Powered by Kunena Forum