Symmetry in CCSD calculations

Find answers or ask questions regarding Dalton calculations.
Please upload an output file showing the problem, if applicable.
(It is not necessary to upload input files, they can be found in the output file.)

Post Reply
mcastro1
Posts: 4
Joined: 16 Mar 2015, 14:19
First name(s): Marcos
Middle name(s): Antônio de
Last name(s): Castro
Affiliation: Instituto de Física, Universidade Federal de Goiás
Country: Brazil

Symmetry in CCSD calculations

Post by mcastro1 » 23 Mar 2015, 14:10

Dalton Users/Developers

Could someone explain to me why CCSD calculations using symmetry give different results from those with disabled symmetry. Here are two CCSD calculations for methanol molecule with Cs and C1 symmetries. The HF and MP2 energies are equal but the CCSD enegies are different.

With best regards,
Marcos A. Castro
Attachments
ccsd_methS.out
Calculation using symmetry
(41.58 KiB) Downloaded 233 times
ccsd_methA.out
Calculation with disabled symmetry
(38.58 KiB) Downloaded 215 times

sonco
Posts: 87
Joined: 06 Sep 2013, 13:49
First name(s): Sonia
Last name(s): Coriani
Affiliation: DTU Chemistry
Country: Denmark

Re: Symmetry in CCSD calculations

Post by sonco » 23 Mar 2015, 19:59

Hello
It looks like a problem with your version of the code. The two calculations should of course give the same result
to the chosen number of significant digits. I just rerun your calculation with three different versions of the code (2011, 2013, and 2015)
on my mac and the results are correct.
Could you try and recompile your code with 32bit integers and see if it makes any difference?
My executables were 32bit. I will recompile one of my versions with 64 to countercheck the results as well
Sonia

mcastro1
Posts: 4
Joined: 16 Mar 2015, 14:19
First name(s): Marcos
Middle name(s): Antônio de
Last name(s): Castro
Affiliation: Instituto de Física, Universidade Federal de Goiás
Country: Brazil

Re: Symmetry in CCSD calculations

Post by mcastro1 » 24 Mar 2015, 13:53

Thank you, Sonia. I will recompile the code.

Marcos A. Castro

sonco
Posts: 87
Joined: 06 Sep 2013, 13:49
First name(s): Sonia
Last name(s): Coriani
Affiliation: DTU Chemistry
Country: Denmark

Re: Symmetry in CCSD calculations

Post by sonco » 24 Mar 2015, 14:38

I have now tried 64int on my mac (using gfortran) and the results are still ok.
What I notice is that your first CCSD iteration energy is already wrong in the case with symmetry,
and it iterates 16 times (versus 9 when you don't use symmetry).
Have you seen this problem in any other case?
Sonia

mcastro1
Posts: 4
Joined: 16 Mar 2015, 14:19
First name(s): Marcos
Middle name(s): Antônio de
Last name(s): Castro
Affiliation: Instituto de Física, Universidade Federal de Goiás
Country: Brazil

Re: Symmetry in CCSD calculations

Post by mcastro1 » 24 Mar 2015, 18:23

Yes, I did similar calculations for other systems and noticed the same problem.

Marcos

sonco
Posts: 87
Joined: 06 Sep 2013, 13:49
First name(s): Sonia
Last name(s): Coriani
Affiliation: DTU Chemistry
Country: Denmark

Re: Symmetry in CCSD calculations

Post by sonco » 24 Mar 2015, 19:53

Then you have a buggy code, either as such or as a result of the compilation.
Which version are you using? Have you tried installing the latest released version?
Did you run the test suite when you installed the code? Did the CC tests pass?
Sonia

mcastro1
Posts: 4
Joined: 16 Mar 2015, 14:19
First name(s): Marcos
Middle name(s): Antônio de
Last name(s): Castro
Affiliation: Instituto de Física, Universidade Federal de Goiás
Country: Brazil

Re: Symmetry in CCSD calculations

Post by mcastro1 » 25 Mar 2015, 19:06

Dear Sonia
Thank you for your attention. I am using the 2013 version. On installation all tests had been made and I had not noticed any mistakes. I am installing the 2015 version now.

Marcos

Post Reply

Who is online

Users browsing this forum: No registered users and 7 guests