KR-CI problem

Bingbing Suo

unread,

Jan 14, 2013, 2:25:17 AM1/14/13

to dirac...@googlegroups.com

Dear Dirac users,

I would like to perform an 4-components MR-CISD calculation by using KR-CI for B atom. I have some problems about such calculation.

My B.inp file looks like this, C2v point group is used in calculation

**DIRAC

.TITLE

Li. DHF test calculations.

.WAVE F

.PROPERTIES

#.4INDEX

.ANALYZE

**HAMILTONIAN

.DOSSSS

**INTEGRALS

*READIN

.UNCONTRACT

**WAVE FUNCTIONS

.SCF

.KR CI

*SCF

.EVCCNV

1.0E-10 1.0E-9

.CNVINT

0.01 0.0001

.CLOSED SHELL

4

.OPEN SHELL

1

1/6

##### Start CI input####

*KRCICALC

.CI PROGRAM

LUCIAREL

.GAS SHELLS

3

0 2 / 1

5 5 / 4

5 5 / 25

.CIROOTS

3 1

.MAX CI

20

*END OF

The first question is about CIROOT. We know C2v double group has 4 boson irreps and 1 femi irrep. Since the B has odd number of electrons, I expect the group state of B has femi irrep. So I write input as

.CIROOTS

1 1

but it does't work.

I found following lines in output.

Number of determinants per double group irrep

=============================================

0.00000000000 0.00000000000 126.000000000 126.000000000

Why there are four irreps? If I input

.CIROOTS

3 1

It seems the calculation can go further but I am fully confuse about the symmetry. Could you please recommend me some books about double group symmetry?

Further, I still could not run KR-CI. I encountered such error with my previous input and exist

==============================================================

==> allocation of two CI vectors and one resolution vector <==

current available free memory in double words: 2047957120

allocate two CI vectors each of length: 126

allocate resolution vector of length: 768

==============================================================

No integral found in geth2dia_sigden.

Type of integral = COUL

Index pair 1 6

Indices 5 -5 -5

5

--- SEVERE ERROR, PROGRAM WILL BE ABORTED ---

Date and time (Linux) : Mon Jan 14 12:37:18 2013

*** error in geth2dia_sigden: no 2e- diagonal integral found. ***

It seems some MO integrals could not been found?

Dear Dirace expert, what is the matter of my calculation?

Bingbing Suo

Institute of Modern Physics, Northwest University

229 Taibaibei Road

Xi'an, Shaanxi 710069

Email: bingbi...@gmail.com

bs...@nwu.edu.cn

Bingbing Suo

unread,

Jan 16, 2013, 2:25:57 AM1/16/13

to dirac...@googlegroups.com

Hi Dirac Users, I found my previous question about KR-CI has some error in my input.

I would like calculate B atom at C2v symmetry. The full relativistic calculation is preferred.

My previous input is not correct. The Gas space looks like

*KRCICALC
.CI PROGRAM
LUCIAREL
.GAS SHELLS
3
0 2 / 1
5 5 / 4
5 5 / 25
.CIROOTS
3 1
.MAX CI
20

It should be

.GAS SHELLS
3
0 2 / 1

3 5 / 4
5 5 / 25

The KR-CI runs smoothly. However, I still confuse about CIroots in my input.

.CIROOTS
3 1

3 means irrep 3, but I only have one fermion irrep in C2v double point group.

Bingbing Suo

Institute of Modern Physics, Northwest University

229 Taibaibei Road

Xi'an, Shaanxi 710069

Email: bingbi...@gmail.com

bs...@nwu.edu.cn

下面是被转发的邮件：

发件人: Bingbing Suo <bingbi...@gmail.com>
主题: KR-CI problem
日期: January 14, 2013 3:25:17 PM GMT+08:00
收件人: "dirac...@googlegroups.com" <dirac...@googlegroups.com>

Stefan Knecht

unread,

Jan 16, 2013, 3:21:23 AM1/16/13

to dirac...@googlegroups.com

Dear Bingbing,

sorry for the slow response...
what you really want with the B atom (or any atom/linear molecule for that matter) is to run Dirac/KR-CI in linear symmetry where you can
directly specify the Mj\OMEGA value for the target electronic state. this also saves computational time and memory.
see for example here:

http://diracprogram.org/doc/release-12/manual/krci.html#ciroots

in your case, the input for 1 root with Mj = 1/2u would look like:

.CIROOTS
1u 1

Linear symmetry is ensured by symmetry detection in Dirac, just leave out any symmetry generator in the .mol file, e.g. for B it could look like:

INTGRL
B atom

C   1                  .10D-15
        5.    1
B             .00000000           .00000000           .00000000
LARGE BASIS 6-31G
FINISH

Now to your symmetry question:
if you ask for C2v in Dirac, KRCI will actually run with the Abelian subgroup C2 where
the first two irreps (gamma_1 and gamma_2, see attached pdf) are active for systems with an even number of e- and
gamma_3 and gamma_4 for systems with an odd number of e- (as B).

hope, this helps. as recommended above, whenever possible one should exploit (linear) symmetry...

with best regards,

stefan

锟斤拷锟斤拷锟角憋拷转锟斤拷锟斤拷锟绞硷拷锟斤拷

锟斤拷锟斤拷锟斤拷: Bingbing Suo <bingbi...@gmail.com>

锟斤拷锟斤拷: KR-CI problem

锟斤拷锟斤拷: January 14, 2013 3:25:17 PM GMT+08:00

锟秸硷拷锟斤拷: "dirac...@googlegroups.com" <dirac...@googlegroups.com>

--

-- 

Stefan Knecht
Institut for Fysik, Kemi og Farmaci
Syddansk Universitet Odense
Campusvej 55
DK-5230 Odense M
Denmark

phone: +45 6550 3512   fax: +45 6615 8760
email: kne...@sdu.dk
web: http://www.theochem.uni-duesseldorf.de/users/stefan/index.htm

c2.PDF

Bingbing Suo

unread,

Jan 16, 2013, 3:38:24 AM1/16/13

to dirac...@googlegroups.com

Dear Stefan,

Thank you for replying!

I am still confuse with symmetry. My .mol file for B is

INTGRL

B DHF

Basis from dalton

C 1 2 Y X

5. 1

B .0000000000 0.0000000000 0.0000000000

LARGE BASIS cc-pCVDZ

FINISH

That means I chose C2v symmetry in calculation. It seems DHF use C2v.

So, do you means KR-CI will automatically use C2 group in calculation? Why? Does KR-CI use low order group to avoid some difficulties arise by degenerate irreps?

Furthermore, If C2 symmetry is used, Dirac use 1 and 2 label to Boson irreps, 3 and 4 label to Fermion irreps. Is it correct?

Bingbing Suo

Institute of Modern Physics, Northwest University

229 Taibaibei Road

Xi'an, Shaanxi 710069

Email: bingbi...@gmail.com

bs...@nwu.edu.cn

下面是被转发的邮件：

发件人: Bingbing Suo <bingbi...@gmail.com>

主题: KR-CI problem

日期: January 14, 2013 3:25:17 PM GMT+08:00

收件人: "dirac...@googlegroups.com" <dirac...@googlegroups.com>

--

--

<c2.PDF>

Stefan Knecht

unread,

Jan 16, 2013, 3:47:07 AM1/16/13

to dirac...@googlegroups.com

Dear Bingbing,

On 16/01/13 09.38, Bingbing Suo wrote:

Dear Stefan,

Thank you for replying!

I am still confuse with symmetry. My .mol file for B is

INTGRL

B DHF

Basis from dalton

C 1 2 Y X

5. 1

B .0000000000 0.0000000000 0.0000000000

LARGE BASIS cc-pCVDZ

FINISH

That means I chose C2v symmetry in calculation. It seems DHF use C2v.

So, do you means KR-CI will automatically use C2 group in calculation? Why? Does KR-CI use low order group to avoid some difficulties arise by degenerate irreps?

indeed, the SCF uses C2v but KR-CI will use C2. i can't remember why exactly though but in general it will use Abelian sub-groups with non-degenerate irreps.

Furthermore, If C2 symmetry is used, Dirac use 1 and 2 label to Boson irreps, 3 and 4 label to Fermion irreps. Is it correct?

yes. :)

to summarize, unless you have a very good reason for it, i strongly recommend to use the highest possible symmetry of your system in order to save computational time and memory.

with best regards,

stefan

Bingbing Suo

Institute of Modern Physics, Northwest University

229 Taibaibei Road

Xi'an, Shaanxi 710069

Email: bingbi...@gmail.com

bs...@nwu.edu.cn

锟斤拷 Jan 16, 2013锟斤拷4:21 PM锟斤拷Stefan Knecht <kne...@sdu.dk> 写锟斤拷锟斤拷

锟斤拷锟斤拷锟角憋拷转锟斤拷锟斤拷锟绞硷拷锟斤拷

锟斤拷锟斤拷锟斤拷: Bingbing Suo <bingbi...@gmail.com>

锟斤拷锟斤拷: KR-CI problem

锟斤拷锟斤拷: January 14, 2013 3:25:17 PM GMT+08:00

锟秸硷拷锟斤拷: "dirac...@googlegroups.com" <dirac...@googlegroups.com>