Technical Support Information
Last update: 26 September 2006

Overlay 8

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 27 28 29 30 31 32 35 36 38 39 40 41 42 43 44 45 46 47

IOp(8/5)

Whether to pseudocanonicalize ROHF orbitals.

-1        Yes.

0        No.

IOp(8/6)

Bucket selection.

0        BUCKETS FOR MP2: (IA/JB).

1        BUCKETS FOR STABILITY: (IA/JB),(IJ/AB).

2        BUCKETS FOR CID OR MP3: (IJ/AB),(IA/JB),(IJ/KL).

3        BUCKETS FOR SEMI-DIRECT MP4DQ, CISD, QCISD, BD: (IJ/AB),(IA/JB),(IK/KL),(IJ/KA).

4        CISD or MP4SDQ or MP4SDTQ, BUT INCLUDES (IA/BC).

5        THE COMPLETE SET OF TRANSFORMED INTEGRALS.

6        Full transformation if this is consistent with MaxDisk, otherwise same as 3.

7        Full transformation if this is consistent with MaxDisk, otherwise same as 4.

IOp(8/7)

SCF convergence test.

0        Test that SCF has convergd.

1        Do not test SCF convergence (mainly used for testing).

IOp(8/8)

Whether to delete MO integrals in L811.

0        Default (No).

1        Yes.

2        No.

IOp(8/9)

Debug control (L802):

0        Operate normally.

-N      Force N orbitals per pass.

Direct Transformation Control (L804, L811):

0        Operate normally.

1        Generate and test RInt3 array (L804).

2        Accumulate MP2 force constant terms in direct fashion

3        Write the MO basis first derivative ERI's to disk

10       Force fully in-Core algorithm (L804 only).

20       Force transformed integrals in Core algorithm.

30       Force semi-direct transformation.

100     Force output bucket in Core antisymmetrization.

200      Force sorting for output bucks.

1000    Force semi-direct mode 1.

2000    Force semi-direct mode 2.

3000    Force semi-direct mode 3 if IOp(6)=3.

4000    Force semi-direct mode 4 if IOp(6)=3.

00000  Default (10000)

10000  Do not symmetry compress transformed integrals.

20000  Do symmetry compress transformed integrals (buckets) (This will cause windowed MOs,
            reordered in the order of representations like occ-rep1,occ-rep2,... virt-rep1,virt-rep2,...
            Eigenvalues and symmetry assignment vectors will be put in correspondence with vectors.)

30000    Symmetry compress transformed integrals only if RHF. (Upper triangle of symmetry
              compressed integrals for IOp(6)=5 or 4 only!)

100000  Reorder MOs, eigenvalues and symmetry assignment vectors according to ther representations

IOp(8/10)

Window is selected as follows:

-N      Use the top N occupieds and lowest N virtuals.

0        Default, same as 4.

N       1 <= N <= 89 selects frozen-core type N:

1        The largest noble gas core is frozen.

2        G2 frozen-core: the largest noble gas core and main group d orbitals are frozen,
          except that the outer sp electrons of 3rd row and later alkalai and alkalai earth elements are retained.

3        The next to the largest noble gas core is frozen.

4        The largest noble gas core and main group d's are frozen. For basis sets with double-zeta
          cores, core virtuals are also frozen.

90        Use all MOs.

91        The window is specified by IOp(37-38). If IOp(37) is 0, a card is read in indicating
            the start and the end. A negative value for the end deletes the top virtuals.

92        The window is recovered from rwf 569.

93        The window is recovered from file 569 on the checkpoint file.

94        Read a list of orbitals to freeze.

000      Default (200).

10x      Use orbital energies to choose core orbitals.

20x      Use overlap with atomic core orbitals from Harris to choose core orbitals.

30x      Use overlap with atomic core orbitals from Core Ham to choose core orbitals.

IOp(8/11)

MO coefficient, orbital energy, and number of electrons test.

0        Default, same as 2

1        Just print a warning message.

2        Kill the job if any mo coefficients are greater than 1000.0 or the smallest difference between
          occupied and virtual orbital energies is less than 0.001. Also, kill a frozen-core job if there
          there is significant core-valence mixing in the canonical orbitals

00        Default, same as 10.

10        Suppress such a test (CPHF may still be done for such a case).

20        Kill the job if there is no correlation energy; e.g., if there is only 1 electron or 1 virtual spin-orbital.

IOp(8/12)

Calculation of frozen-Core contributions.

0        No.

1        Calculate 2 J - K over deleted orbitals and add to Core-Hamiltonian. This is done when IOpCl=0
          or when IOpCl=1 and the calculation is rohf or gvb.

IOp(8/13)

Control of output. Used to select output mode.

0        Output to Gaussian system buckets.

1         Output transformed integrals for DRT-CI calculation.

IOp(8/14)

Control of drt input.

0        Take necessary input from Gaussian data structures.

1        Read 'old-style' drt input cards.

IOp(8/15)

Control of DRT output.

0        Write DRT output to RW-files.

1        Write DRT output to Fortran unit 'drttap'.

2        Do both.

IOp(8/16)

Maximum number of orbitals per pass in L811. (only if integral derivative file is being written)

Excitation level for SDGUGA-CI.

0        Default excitation level = 2.

N       Excitation level = N.

IOp(8/17)

Specification of integral block size for GUGA CI programs.

0        Default let program decide.

N        Integral Block Size = N.

IOp(8/18)

Which type of derivative transformation to do in L811:

0        Default, same as 3.

1        Non-canonical, Uij,x = -1/2 Sij,x.

2        Canonical, Uij,x = (Fij,x - EjSij,x) / (Ei-Ej) Note that this blows up for degenerate orbitals
          and is intended primarily for debugging.

3        Non-canonical, Uij,x = -1/2 Sij,x, except canonical in frozen-active blocks.

4        Non-canonical, Uij,x = -Sij,x Uji,x = 0.

5        Canonical occupieds, Uab,x = -Sab,x/2

6        Canonical virtuals, Uij,x = -Sij,x/2

IOp(8/19)

The nature of the perturbation(s) in L811:

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

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

IOp(8/20)

Which terms to include in L811:

0        Default (same as 11).

1        MO derivative times integral term.

10      MO times integral derivative term.

IOp(8/22)

These options control the in-Core post-SCF link, L805. Look there for more information.

IOp(8/27)

Maximum amount of disk to use in L804 and L811:

0        Unlimited.

N       N words.

IOp(8/28)

Hack number of occupieds for full ci using links 921 or 922:

-1        Transform all orbitals (after freezing Core) as occupieds (i.e., set NOA=NOB=NROrb in transformation).

0         No.

N        Transform N orbitals (after frozen Core) as occupieds (i.e., set NOA=NOB=N for purposes of transformation).

IOp(8/29)

Maximum number of perturbations per batch in L811: (only applies if integral deriv file is written)

-3        Do not use batching logic.

-2        Do as many in a batch as can be overlapped with sorting space for half transformed integrals.

-1        Do one batch, but use multi-batch logic.

0         Default (same as 1).

1         Do a single atom at a time (minimum disk usage).

N        N triplets.

Requested disk usage. This will determine the number of times AO integrals and derivatives are evaluated unless overridden by IOp(31). This only applies if the integral derivatives are not stored.

-3        Use as much as desired, independant of MAXDISK.

-2        Use an amount which is similar to the maximum disk usage in other parts of the MP2 frequency code.

-1        Use as much as needed for maximum efficiency, subject to the limit imposed by MAXDISK (IOp(27)).

0          Default (-1)

N         N evaluations and hence N coarse tiled batches (1...6 are the currently implemented options)

IOp(8/30)

Type of window.

0        Default. Set up /Orb/ as indicated by IOp(10).

1        Test window. Set up for full but zero Core MOs.

-1       Set up /Orb/ for a full window but then blank the wavefunction coefficients in L804.

IOp(8/31)

PERFORM PRIMITIVE POST-SCF OPERATIONS (NOT CURRENTLY FUNCTIONAL).

0        NO

1        YES

IOp(8/32)

Whether to do CI in the interacting space only.

0        Default (all spin-eigenfunctions).

1        All.

2        Interacting only.

IOp(8/35)

Output format for closed-shell and debugging control: (only for when integral derivative file is written)

0        Default (consistent with integrals for open-shell, i<=jab alpha-beta only for closed-shell).

1        Store only the unique AB integral derivatives (gO2V2/4, order g i,=j a<=b) for closed-shell
          2        AA and AB consistent with integrals.

10      Do extra debugging computations.

Explicit control of the fine tile batch size (largely a debugging option, only for no-Ix routines).

0         Let the program choose (make it as large as possible)

N        maximum fine tile batch size, up to 9.

IOp(8/36)

Whether to update force constants with the MP2 product of MP2 integral derivatives term (only applies if integral derivative file is not written).

0        Default (Yes).

1        Yes.

2        No.

00      Default on whether to make Poo and Pvv for MP2. (Yes if Ix is not stored, no otherwise).

10      Yes.

20      No.

IOp(8/38)

Integers specifying the window to use.

IOp(8/39)

Localized orbital method adopted in SAC/SAC-CI.

0        Default. No localization.

1        Boys method.

2        Population method.

3        Boys + population method.

IOp(8/40)

Handling of ROHF window:

0        Default (2).

1        Use ROMP2 approach, forming pseudo-canonical alpha and beta orbitals and doing UHF transformation.

2        Treat as RHF, transforming only alpha orbitals.

IOp(8/41)

Transformation of spin-orbitals (alpha only) within occupied and unoccupied orbital subspaces by minimum orbital-deformation (MOD) method.

0        Default. No.

1        No, but save MOs.

2        Yes. Take reference MOs from disk if available.

3        No for the 1st geometry of opt, yes otherwise.

IOp(8/42)

Whether to reorder MOs during potential surface exploration.

0        No

1        Yes.

2        Yes. (for SAC-CI single point calculation)

00       Use orbital energies in ordering

10       Don't use orbital energies in ordering

000     Use second moments in ordering

100     Don't use second moments in ordering

0000   Use dipole moments in ordering

1000   Don't use dipole moments in ordering

IOp(8/43)

Number of Laplace points to use.

N        Use N points for MP2.

-N       Use N points and set up for band gap correction.

IOp(8/44)

K-point specification for MP2 band correction.

0        Use the k-point for which the hoco is highest and luco is lowest.

N        K-point at which to apply correction.

IOp(8/45)

Type of quasiparticle job:

0        Band gap.

-1      Ionization potential.

1        Electron affinity.

-N      N-th occupied band at the k-point for which the hoco is highest (by default) or at k-point specified by IOp(44)

N       N-th virtual band at the k-point for which the luco is lowest (by default) or at k-point specified by IOp(44)

IOp(8/46)

Indicates special case of non-HF calculation.

0        Default - MOs are canonical HF orbitals.

1        Input orbitals are not canonical HF and pseudocanonical orbitals must be generated here for the post-SCF.

IOp(8/47)

Whether 804/811 should generate results compressed over active atoms:

0        Default (2).

1        Active atoms.

2        Full list.

3        Full list, but blank contributions from inactive atoms. No difference from 2 for overlay 8.