× If you have problems during the execution of MRCC, please attach the output with an adequate description of your case as well as the followings:
  • the way mrcc was invoked
  • the way build.mrcc was invoked
  • the output of build.mrcc
  • compiler version (for example: ifort -V, gfortran -v)
  • blas/lapack versions
  • as well as gcc and glibc versions

This information really helps us during troubleshooting :)

FCI does not work in some cases

More
1 year 2 months ago - 1 year 2 months ago #526 by leos
Dear Mihály,

I've noted that in some cases it is not possible to use the FCI method with the dirac interface.
In the attached file one can see calculation of the Boron atom (5 electrons).
The CISDTQ calculation passes without problems, but both FCI and CISDTQP crash.

Also, for some reason mrcc.2018-02-04 does not work with the dirac interface at all, it crashes with the following message:
Formula tape is too old!
Execute goldstone before this program!
(if I use the same input with mrcc.2017-04-12 - there are no problems).

In the attachement:
1. directory "ints" - dirac integrals (can be downloaded from qchem.pnpi.spb.ru/data/ints.tar.gz )
2. directory vdz_C2V_CISDTQ_2017 - CISDTQ calculation within mrcc.2017-04-12 - OK
3. file mrcc.out_CISDTQP_2017 - CISDTQP(=FCI for the case of the Boron atom) calculation within mrcc.2017-04-12 - crashed
4. file mrcc.out_FCI_2017 - FCI calculation within mrcc.2017-04-12 - crashed
5. file mrcc.out_CISD_2018 - CISD output of mrcc.2018-02-04 (crashed)

Could you help, please?

Best,
Leonid.



File Attachment:

File Name: vdz_C2V_CI...7.tar.gz
File Size:4 KB


File Attachment:

File Name: outputs.zip
File Size:9 KB
Attachments:
Last edit: 1 year 2 months ago by leos.

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

More
1 year 2 months ago #529 by kallay
Dear Leonid,
You will find two files, mrcc.f and xmrcc.f, in the download area. Please recompile the program with those files. This will resolve the issue.

Best regards,
Mihaly Kallay

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

More
1 year 2 months ago #537 by leos
Dear Mihály,

Thank you, for the files.
When I tried to compile mrcc.2018-02-04 with these files the compiler showed error because subroutine omplim defined in two files
"combin.f" and new "mrcc.f".
So I renamed omplim->omplim1 in the "combin.f" file.
After that the compilation was successful.
And now program works for the example given above.


However, if I use larger basis set (60 spinors) then program crashes after words:
"Calculation of coupling coefficients...

Fatal error in xmrcc.
Program will stop."

Please find attached
-file mrcc.out_CISD_2018.03.17
it shows that CISD calculation is OK

-file mrcc.out_CISDTQ_2018.03.17
it shows that the CISDTQ calculation is also OK (I killed the process after the first iteration).

-file mrcc.out_FCI_2018.03.17
for some reason the job crashes.

Integrals can be downloaded from here:
qchem.pnpi.spb.ru/data/MDCINT
qchem.pnpi.spb.ru/data/MRCONEE

What do you think about it?

Best,
Leonid.

File Attachment:

File Name: outputs.tar.gz
File Size:4 KB
Attachments:

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

More
1 year 2 months ago #539 by kallay
Dear Leonid,
Edit file goldstone.f, go to line 3066. Replace
$dble(memory)/2.d0),ihf
with
$dble(memory)),ihf

Best regards,
Mihaly Kallay

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

More
1 year 2 months ago #540 by leos

kallay wrote: Dear Leonid,
Edit file goldstone.f, go to line 3066. Replace
$dble(memory)/2.d0),ihf
with
$dble(memory)),ihf


Dear Mihály,
I did this and recompiled all files,
However, the problem still exists.

Please, find attached the output file.

P.S.
I use x86_64 Linux and the following compiler options: Intel -i64 -pOMP -d.

File Attachment:

File Name: mrcc.out_F...8.03.zip
File Size:2 KB
Attachments:

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

More
1 year 2 months ago #541 by kallay
Dear Leonid,
You should increase the memory available to the program and use symmetry. I have no better idea at the moment. The program is not designed for large basis set full CI calculations, and that part consumes too much memory.

Best regards,
Mihaly Kallay

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

Time to create page: 0.022 seconds
Powered by Kunena Forum