ERROR **** CHOLSK **** BASIS SET LINEARLY DEPENDENT

job314

I am running a rather simple crystal Na2Si2O5 with B973C and mTZVP . When calculation begins, I immediately get

ERROR **** CHOLSK **** BASIS SET LINEARLY DEPENDENT

This is strange since I already used B973C and mTZVP in other situations and it worked. This one seems random. Plus this basis set is built in and mandated to be used with B973C. Why is it linearly dependent and how to circumvent that?

thank you

GiacomoAmbrogio

Hi job314,
The linear dependence of the basis set can be due by several reasons. Most likely, it is the presence of diffuse orbitals in the basis set. The reason you are experiencing this error in this paricular system is likely the geometry: some orbitals may be closer together compared to other systems, leading to the issue.

To address the linear dependencies, you have a couple of options:

Manually removing diffuse functions A common approach is to eliminate basis functions with exponents below 0.1, as these are often the cause of linear dependence.
Using the LDREMO keyword You can add the following keyword in the third section of your input file (below the SHRINK keyword should work):

LDREMO
<integer>

This will systematically remove linearly dependent functions by diagonalizing the overlap matrix in reciprocal space before the SCF step. Basis functions corresponding to eigenvalues below <integer> * 10^-5 will be excluded. You could start with 4 and increase it if needed.

You should see printed information about the excluded basis functions in the output file, but note that this is only available in serial mode (i.e., running with a single process).

Could you kindly share your input file? This would allow us to run some tests on the specific basis set and structure. And let me know if you need further help!

job314

Giacomo, but here is the kick... This is a built-in optimized mTZVP basis set. I do not want to mess with built in basis sets since that becomes random on my part. But thank you for responding, see the input. And yes, to see the error I had to run on serial mode on my windows machine since Linux parallel output did not tell any errors

CAPPA:IS1  6;IS2  2;IS3  8; K PTS MONK NET  52; SYMMOPS:K SPACE   4;G SPACE   4

Abort(1) on node 19 (rank 19 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 19

input.d12

GiacomoAmbrogio

Thanks for sharing the input.
Even though the mTZVP basis set is built into the code, it is still a molecular basis and is therefore prone to linear dependence due to small exponents.

job314

Workign on it. when I specify LDREMO 4 I get

TTTTTTTTTTTTTTTTTTTTTTTTTTTTTT MOQGAD TELAPSE 1197.77 TCPU 1185.41
ERROR **** CLASSS **** ILA DIMENSION EXCEEDED - INCREASE ILASIZE 6000
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTT ERR TELAPSE 1230.13 TCPU 1215.62

GiacomoAmbrogio

This is a completely unrelated error due to the size of the system.
Check ILASIZE keyword on the CRYSTAL user manual, page 117.

job314

OK, I only ran into it since I could not diagnose why my parallel job with LDREMO 4 aborted with no errors:

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

 CHARGE NORMALIZATION FACTOR   1.00000000
 TOTAL ATOMIC CHARGES:
  11.0000000  11.0000000  11.0000000  11.0000000  11.0000000  11.0000000
  11.0000000  11.0000000  11.0000000  11.0000000  11.0000000  11.0000000
  11.0000000  11.0000000  11.0000000  11.0000000  14.0000000  14.0000000
  14.0000000  14.0000000  14.0000000  14.0000000  14.0000000  14.0000000
  14.0000000  14.0000000  14.0000000  14.0000000  14.0000000  14.0000000
  14.0000000  14.0000000   8.0000000   8.0000000   8.0000000   8.0000000
   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000
   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000
   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000
   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000
   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000
   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000   8.0000000
 TTTTTTTTTTTTTTTTTTTTTTTTTTTTTT MOQGAD      TELAPSE       60.12 TCPU       58.96
 WARNING **** GENBUD **** COUL. BIPO BUFFER TOO SMALL - TO AVOID I/O SET BIPOSIZE =    11868000
Abort(

GiacomoAmbrogio

Hi job314,
After looking into your issue further, I’m following up with more information.

The B973C functional is a composite method with built-in corrections specifically designed for the mTZVP basis set. Modifying the basis set can introduce errors and is not the right approach. This method and basis set were primarily developed for molecular systems and, at most, molecular crystals, not bulk materials like yours.

Explicit warnings about this functional can be found in the user manual on page 161.

Given this, I recommend choosing a different functional and basis set better suited for your system.

job314

thank you