EOMCCSDT Calculation Does Not Reach Convergence Criteria


Forum Vet
The input is like this

START tce_eomccsd_t

echo
memory total 3500 stack 1000 heap 500 global 2000 mb

geometry units angstroms print xyz autosym
  more than twenty atoms
end


BASIS SPHERICAL
  • library cc-pvdz
end

scf
thresh 1.0e-10
tol2e 1.0e-10
singlet
rhf
end

cosmo
solvent water
end

tce
freeze atomic
ccsdt
thresh 1.0d-5
nroots 1
2emet 13
maxiter 500
end

task tce energy.

The log file exhibits
...

Global array virtual files algorithm will be used

Parallel file system coherency ......... OK

Integral file          = ./tce_eomccsd_t.aoints.0
Record size in doubles = 65536 No. of integs per rec = 43688
Max. records in memory = 1140 Max. records in file = 128663
No. of bits per label = 8 No. of bits per value = 64


#quartets = 4.599D+06 #integrals = 8.722D+07 #direct =  0.0% #cached =100.0%


File balance: exchanges= 2 moved= 152 time= 0.0


Fock matrix recomputed
1-e file size = 26244
1-e file name = ./tce_eomccsd_t.f1
Cpu & wall time / sec 8.1 8.2
4-index algorithm nr. 13 is used
------------------------------------------------------------------------
tce_energy: invalid 2emet: 13
4-index algorithm nr. 13 is used
tce_energy: invalid 2emet: 13
------------------------------------------------------------------------
------------------------------------------------------------------------
current input line :
0:
------------------------------------------------------------------------
------------------------------------------------------------------------
This type of error is most commonly associated with calculations
not reaching convergence criteria
4-index algorithm nr. 13 is used
tce_energy: invalid 2emet: 13
------------------------------------------------------------------------
------------------------------------------------------------------------
current input line :
0:
------------------------------------------------------------------------
------------------------------------------------------------------------
This type of error is most commonly associated with calculations
not reaching convergence criteria.
In fact, there are no iterations here.
Only "io ga" can be used in this situation?
Thanks for your advice.
By the way, test tce_active_ccsdt is successful with the following result
...
Iterations converged
largest EOMCCSDt amplitudes: R1 and R2

Singles contributions
8a1 (alpha) --- 5b2 (alpha) 0.6803669699
10a1 (alpha) --- 5b2 (alpha) 0.2149012535
18a1 (alpha) --- 5b2 (alpha) -0.1006544488
9b2 (alpha) --- 6a1 (alpha) -0.2149069605

Doubles contributions
8a1 (alpha) 8a1 (beta ) --- 6a1 (alpha) 5b2 (beta ) -0.2388813098
8a1 (alpha) 10a1 (beta ) --- 6a1 (alpha) 5b2 (beta ) -0.1087535267
8a1 (alpha) 8a1 (beta ) --- 5b2 (alpha) 6a1 (beta ) -0.2388813098
10a1 (alpha) 8a1 (beta ) --- 5b2 (alpha) 6a1 (beta ) -0.1087535267
8a1 (alpha) 9b2 (beta ) --- 5b2 (alpha) 5b2 (beta ) 0.1119473525
9b2 (alpha) 8a1 (beta ) --- 5b2 (alpha) 5b2 (beta ) 0.1119473525

Parallel integral file used       7 records with       0 large values.

Forum Vet
The use of io ga here is really ga size demanding, therefore 2reob and other alternative two-electron
integral storage methods are urgently expected to be integrated into this method.


Forum >> NWChem's corner >> Running NWChem