# Overlay 11

IOp(11/5)

IFWRT: derivative integral write option.

0 | Do not produce a D2E file. |

1 | Produce a D2E file. |

IOp(11/6)

IFHFFX: Whether or not to contract integral derivatives with Hartree-Fock density matrix terms to produce Hartree-Fock two-electron contribution to the forces.

0 | No. |

1 | Yes. |

2 | Yes, also contracted electric field density matrix derivatives to form thetwo-electron integral derivative contribution to the pol. derivatives, but ignore frequency-dependent density derivs. |

3 | Yes, do polarizability derivatives using frequency-dependent density derivatives if the FD density derivatives are available. |

IOp(11/7)

IFTPDM: whether or not to contract integral derivatives with a ‘read-in’ two-particle density-matrix.

0 | No. |

1 | Yes. |

2 | Yes, but generate and write out the HF 2PDM here for debugging purposes. |

-N | Generate and use the 2PDM for CIS state N. |

IOp(11/8)

IFF1: whether or not to compute F1 over AO’s.

0 | No. |

1 | Yes. |

2 | Yes, then compress to active atoms. |

3 | Generate active list. |

IOp(11/9)

IDOUT: First-derivative output option.

Contains I2*100+I1*10+I0.

I0 | Whether or not to use the contents of IRWFX. |

0 | No. |

1 | Yes, if not there, merely set the array to zeroes. |

I1 | Processing of two-electron Hartree-Fock contributions. |

0 | None. |

1 | Take HF contributions from FX1 (A LA IFHFFX). |

2 | Take HF contributions from F1 (A LA IFF1). (forms the 1/2(F-H) term in link 1110). |

3 | Form 1/2(F+H) term in link 1110. |

I2 | Processing of TPDM contributions. |

0 | None. |

1 | Add in contents of FX2. |

IOp(11/10)

L1110: Whether to compute Fock matrices, Lagrangian, and SCF energy.

0 | No. |

1 | Yes. |

2 | Yes, and in addition, compute other pieces necessary for second-order simultaneous optimization. |

IOp(11/11)

Control of integral derivative algorithm.

0 | Default: use IsAlg to decide. |

2 | Scalar Rys SPDF. |

3 | Illegal here. |

4 | Illegal here. |

5 | Illegal here. |

6 | Illegal here. |

7 | Illegal here. |

8 | Illegal here. |

9 | Illegal here. |

10 | Illegal here. |

11 | Illegal here. |

12 | FoFJK: Prism spdf. |

13 | Illegal here. |

IOp(11/12)

L1102, L1110: Selection of 1PDM.

0 | Usual SCF density. |

N | Use generalized density number N for both the one-electron integral derivatives and the corresponding 2PDM terms. |

-N | Contract with HF density, CI density for state N, and CIS 1PDM for state N. |

IOp(11/13)

L1112: Flags.

0 | Default for IxÞSx (same as 1). |

1 | Use Ix. |

2 | Use L(x) and Ux*I. |

00 | Formation of Ux*I*T terms, default, same as 1. |

10 | N^{4} I/O algorithm. |

20 | Old gOV3 I/O algorithm. |

000 | Formation of Fx*T*T terms: default is to choose based on available memory. |

100 | Force O2V2 method. |

200 | Use(2g+O)V2 memory algorithm even if O2V2 memory is available. |

300 | Force old N^{5} I/O algorithm. |

0000 | Default Ix*T algorithm (1) |

1000 | Force new algorithm. |

2000 | Force old algorithm. |

00000 | Default availability of MO basis Ix — use if avail. |

10000 | Ix file is not available (omit OO/VV blocks of Px,Wx). |

20000 | Ix file is available (i.e. do OO/VV blocks of Px,Wx). |

000000 | Default non-zero Delta(ij+ab) processing: if Full and some Delta’s are non-zero. |

100000 | Force addition of these terms. |

200000 | Suppress addition of Delta terms. |

IOp(11/14)

The nature of the perturbation(s).

0 | Default (1st order nuclear and electric field). |

IJK | Nuclear Kth order. Electric field Jth order. Magnetic field Ith order. |

IOp(11/15)

Controls output of derivatives to rw-files.

1 | Load fxyz from rw-files if it exists. |

10 | Calculate nuclear contribution. |

100 | Calculate one-electroon contribution. |

1000 | Output of ‘old’ format. |

10000 | Forces out-of-core algorithm. |

IOp(11/16)

L1102: Mode of operation.

0 | Default: compute dipole derivative matrices only. |

1 | Also compute dipole derivative integral contribution to the HF dipole derivatives. |

10 | Also compute HF contribution to the dipole moment. |

IOp(11/17)

L1111: Frozen-core.

0 | Default (use AO 2PDM for Lagrangian only if orbitals are frozen in /Orb/). |

1 | Do C1, C2, S1, and S2 off the AO 2PDM. |

2 | Convert /Orb/ to full, for debugging frozen-core with integrals over the full window. |

3 | Save as 2, but leave the full version of /Orb/ on the disk. |

10 | Form the derivative integral contribution to the Lagrangian as well. This is stored on disk as RL(NBasis,NBasis,NAt3,IOpCl+1) in RWF 1001. |

IOp(11/18)

L1111: Save AO 2PDM?

0 | No. |

N | Save the AO 2PDM on RWF N. It is (NTT,NTT) and includes factors (2-Delta(ij))(2-Delta(kl)). It doesn’t include any normalization factor. |

IOp(11/19)

L1112: Whether to delete MO integrals after.

0 | Default (Yes). |

1 | Yes. |

2 | No. |

IOp(11/20)

L1112: How to handle 2e integral contributions.

0 | Default (same as 1). |

1 | Read the 2e integral files, MO if possible. |

2 | Compute the 2e integrals when needed. |

3 | Force use of AO integrals, even if MO ones are available. |

MNx | Use option MN in control of 2e integral calculation. |

IOp(11/21)

Size of buffers for integral derivative file.

0 | Default (Machine dependent; see DSet2E). |

N | N integer words. |

IOp(11/22)

L1112: In-core option for W(Tilde) term.

-6 | Force in-core storage. |

-3 | Suppress in-core storage. |

0 | Default: in-core if possible. |

IOp(11/23)

L1112: Use of Raffenetti integrals during direct term.

-N | All integrals done as Raffenetti if there are N or more matrices; all as regular if there are less than N. |

0 | Default: let FoFJK decide. |

1 | All integrals are done as regular integrals. |

N | Integrals with degree of contraction greater than or equal to N are done at regular integrals. |

IOp(11/24)

L1102: Output.

00 | Default (01). |

1 | Contract with density matrix to form dipole derivative contributions. |

10 | Store dipole derivative matrices on disk. |

IOp(11/26)

Program accuracy option.

0 | Do integrals economically to 10^{-10} accuracy. |

1 | ‘Test’ option bypass cutoffs. |

IOp(11/27)

L1110: Integral retention parameter if writing d2e file.

0 | Retain integrals GE 10^{-10} in the D2E file (if selected) and/or 10^{-10} in the integral heap if IFF1=1 and MODE=2. |

N | Retain integrals GE 10^{-N}. |

L1111: Save unsymmetrized S1 and S2.

0 | No. |

1 | Yes. |

IOp(11/28)

L1111: Location or generation of MO 1 and 2 PDMs.

-16 | Compute EOM-CCSD 2PDM. |

-15 | Compute Direct SAC-CI 2PDM. |

-14 | Compute SAC-CI General-R 2PDM. |

-13 | Compute SAC-CI 2PDM. |

-12 | Compute MP4SDQ 2PDM. |

-11 | Compute MP4DQ 2PDM. |

-10 | Compute MP3 2PDM. |

-9 | Compute BD 2PDM. |

-8 | Compute CCSD 2PDM. |

-7 | Compute QCISD 2PDM. |

-6 | Compute CCD 2PDM. |

-5 | Compute CIS 2PDM. |

-4 | Compute CISD 2PDM. |

-3 | Compute CID 2PDM. |

-2 | Compute MP2 2PDM. |

-1 | Compute HF DMs. |

0 | Default (RWFs 626, 627, and 628). |

N | RWFS N (1PDM), N+1 (W), and N+2 (2PDM). |

IOp(11/29)

What to do:

0 | Nothing. |

1 | Transform 1PDM and Lagrangian from MO to AO. |

10 | Transform 2PDM from MO to AO. |

100 | Sort AO 2PDM into shell order. If back transformation has not been requested, the double-length AO 2PDM is expected in file 1001. The sorted 2PDM is left in file 602. |

200 | Form the contribution of the 2PDM to the forces right here. Note that if the 2PDM is also to be left behind, it will be over 6d/10f and have the HGP d and f scale factors in it. |

1000 | Suppress writing alpha, beta, and spin density RWFs. |

10000 | Form and sort the 2PDM derivatives rather than the 2PDM. |

20000 | Generate replicated 2PDM copies for testing. |

IOp(11/30)

What to compute using integrals or D2E file.

0 | Nothing. |

1 | Energy. |

10 | Gradient. |

IOp(11/31)

L1110: Whether to use symmetry in Rys integral derivatives.

0 | Yes. |

1 | No. |

2 | Yes. |

3 | Yes, skip check of density symmetry in L1110. |

IOp(11/32)

L1111: Whether to do 2PDM or just Lagrangian.

0 | Compute full gradient |

1 | Compute full gradient (same as default). |

2 | Compute density only. |

3 | Compute density and W only. |

4 | Compute 2PDM only, no density or W. |

5 | Compute non-separable terms only. |

6 | Testing (no lag currently). |

7 | T-relaxed MO-unrelaxed 1PDM for (EOM-)CCSD (no need for 2PDM. 1PDM stored in IODens). |

IOp(11/33)

IPRINT print option.

0 | No printing. |

1 | Print computed first-derivatives. |

2 | Print F1 matrices. |

IOp(11/39)

Compression of derivative matrices.

0 | Default (2 if expanded matrices, otherwise 4 or 5). |

1 | Compute over active atoms only. |

2 | Compute over the full list of atoms. |

3 | Compute over the full list of atoms, but blank contributions for inactive atoms. |

4 | Compute over active atoms only, and store second deriv. contributions over only active atoms. |

5 | Store only matrices for QM atoms, but include the contribution of EE centers in the matrices. |

IOp(11/42)

Compressed file formats.

0 | Default: compressed. |

1 | Force expanded form. |

2 | Force compressed form. |

3 | Compressed Sx but separate H1 and F1. |

IOp(11/43)

Batching in overlay 11.

0 | Default, smallest possible number of passes. |

1 | Do at least one pass, but using the out-of-core algorithms. |

N | Do at least N passes. |

For | Rys in L1110, N is 0/1/2 for default/in-core/out-of-core. |

IOp(11/45)

Force NAt3 instead of NAt3+3 storage of matrices (for debugging).

0 | No. |

1 | Yes. |

IOp(11/46)

Whether to include orbital rotation gradient terms for SAC-CI.

0 | No. |

1 | Convert 1PDM to canonical representation. |

2 | Save gradients to disk, needed for non-canonical methods. |

IOp(11/53)

Convert forces over shells to field-dependent dipole and forces over atoms (for debugging).

0 | No. |

1 | Yes. |

IOp(11/60)

Override standard values of IRadAn.

IOp(11/61)

Override standard values of IRanWt.

IOp(11/62)

Override standard values of IRanGd.

IOp(11/63)

Whether to do FMM.

0 | Use global default. |

1 | Turn off FMM here regardless. |

IOp(11/75)

Print during NMR.

0 | Default (1). |

1 | Print tensors and eigenvalues. |

2 | Print eigenvectors as well. |

IOp(11/76)

L1102: Force DoH1 logic for debugging.

0 | Default (No). |

1 | Yes. |

2 | No. |

IOp(11/77)

Debugging option for DBF derivatives.

0 | Normal processing. |

1 | Ignore fitting density and just process real density in L1110. |

4 | Skip increment of Fx with J(Z^{-1}*Jx(P-Pfit)). |

6 | Compute only Pfit and not P terms involving 2e integral derivatives. |

7 | Clear both Pfit and P before FoFJK. |

1x | Do polarizability derivative contribution separately; only works with density fitting. |

11x | Do polarizability derivatives for density fitting separately and keep only dbf-ao terms. |

21x | Do polarizability derivatives for density fitting separately and keep only dbf-dbf terms. |

31x | Do polarizability derivatives for density fitting separately via 2PDM in one call to FoFCou. |

IOp(11/81)

Control of number of passes in GXX.

0 | Default: at most 96 matrices at a time if doing FMM, otherwise no limit. |

-1 | As few passes (as many matrices) as possible. |

N>0 | Do at most N densities per pass. |

N<-1 | Do at least -N passes. |

IOp(11/87)

L1110: Accuracy of 2e integrals.

0 | Default. |

N | 10^{-N}. |

IOp(11/101)

Raffenetti in DD1Dir.

IOp(11/102)

Control of FMM for nuclear repulsion.

0 | Default: Use for 5K or more atoms. |

N | Use for N or more atoms. |

-1 | Always use FMM. |

2 | Never use FMM. |

IOp(11/103)

Flag for PTED with CCSD and BD.

0 | Normal solvation. |

1 | PTED. |

2 | PTE-S. |

3 | PTES. |

IOp(11/126)

Maximum number of matrices to handle at a time in DD1Dir.

0 | Default (-1). |

-1 | No limit. |

N>0 | At most N matrices at a time. |

Last updated on: 21 October 2016. [G16 Rev. C.01]