Sample Inputs and Outputs
AUX
-
Basic set-up facilities, features associated with
establishment of Normal ModesDownload some sample input-outputs, here:
READ THIS FIRST
|
|
|
|
|
Sample Inputs and Outputs
Download some sample input-outputs, here:
1..AUX: This category is mainly concerned with basic set-up facilities, in particular, the features associated with the establishment of the Normal Modes.
1.1 H2CS TEST....TRANSFORM TO PRINCIPAL AXIS SYSTEM
Test AUX.1
Tests the transformation of the input equilibrium geometry to the principal axis system, via the centre of mass coordinates.
Key input parameters
MCHECK > 0 forces `multimode' along the required route. ISCFCI < 0 terminates `multimode' on completion of the axis transformation.
Remarks
For many molecules, the equilibrium cartesian coordinates are easiest to determine if, for example, the origin is on an atom and one axis is along a bond. The parameter MCHECK can then be used to transform these coordinates into the principal axis coordinates required by the Normal Coordinate procedure within `multimode'. In keeping with all AUX examples, ISCFCI < 0 is used to terminate `multimode' after completion of the required procedure. To continue with the correct coordinates, set ISCFCI = 0 (for SCF) or > 0 (for CI), EITHER with the original input geometry and MCHECK > 0, OR with the principle axis geometry given in the output from AUX.1 and with MCHECK = 0. If symmetry exists, it is important that coordinates whose absolute values should be equal are, in fact input such that this requirement is satisfied. In most cases, this will entail modifying the last few digits in those given by `multimode' in the transformed coordinates to cancel rounding errors. In this case, it will be necessary to continue the calculations with MCHECK = 0.
1.2 H2CS TEST....TRANSFORM TO PRINCIPAL AXIS SYSTEM
Test AUX.2
Tests the internal Normal Mode analysis facility.
Key input parameters
INORM > 0 forces `multimode' along the required route.
IWHICH > 0 indicates user-supplied potential in internal coordinates.
NPOT = 0 (suggested value...see below...parameter ignored)
ISCFCI < 0 terminates `multimode' on completion of the Normal Mode analysis.
Remarks
In the majority of cases, the user will require `multimode' to carry out the Normal Mode analysis that is required in order to facilitate the SCF/CI procedures associated with the Watson hamiltonian. In such cases, note that the potential function MUST be given in internal coordinates (IWHICH > 0) and supplied by the user (see Manual for definitions of user-supplied routines USERIN and GETPOT). For potentials in internal coordinates, the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes), and is ignored on input, although an input value of NPOT = 0 is suggested...see below. When using the Normal Mode facility (INORM > 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are omitted from the input parameters. Also, when using the Normal Mode facility (INORM > 0), the NORMAL COORDINATE DISPLACEMENTS are omitted from the input parameters. In keeping with all AUX examples, ISCFCI < 0 is used to terminate `multimode' after completion of the required procedure. To continue with the correct Normal Coordinates, set ISCFCI = 0 (for SCF) or > 0 (for CI). If symmetry exists, it may be necessary to modify the Normal Mode vectors for those whose absolute values should be equal. This will usually entail modifying the last few digits of the appropriate vectors in those given by `multimode' to cancel rounding errors. In this case, it will be necessary to continue the calculations with INORM = 0.
1.3 H2CS TEST....SEARCH FOR MINIMUM OF POTENTIAL
Test AUX.3
Tests the facility to search for the minimum of the potential.
Key input parameters
MCHECK < 0 forces `multimode' along the required route.
ISCFCI < 0 terminates `multimode' on completion of the minimum search.
Remarks
For many molecules, the precise equilibrium geometry (potential minimum) may not be known exactly. The closer the input equilibrium geometry is to the true minimum of the potential, the more accurate will be the results produced by `multimode'. The parameter MCHECK can be used to determine the required coordinates.
AUX.3 is a test in 3 parts.
Part (a) is the starting position. For the initial (trial) geometry, the Normal Modes are produced (INORM > 0), and the Gauss quadrature points and optimised HEG quadrature points are determined (see the Manual for NBF, MBF, NVF). The potential along each Normal Mode is inspected, in turn, and the potential is fit to a quadratic polynomial, using the three central quadrature points, from which the position of the minimum is found. The minimum is shifted to Q = 0 (the minimum is defined such that all Q = 0). The process is repeated with a cubic, quartic, pentic,... polynomial until all of the integration points are exhausted. The principal axes associated with this search are output by `multimode'.
Part (b) continues with the principal axes produced in part (a), as the exact minimum was not quite found with the choice of integration points.
Part (c) does an SCF run (ISCFCI = 0) using the exact minimum of the potential found in part (b), when it is now no longer necessary to search for the potential minimum (MCHECK = 0).
2.1 H2CS TEST....SCF FOR 10 LOWEST J=0 STATES WITH GENERATED NORMAL MODES
Test SCF.1
Tests the calculation of the ten lowest-energy J=0 SCF states, using Normal Modes determined by `multimode'.
Key input parameters
ISCFCI = 0 forces `multimode' along the required SCF route.
INORM > 0 indicates Normal Modes to be constructed by `multimode'.
NPOT = 0 (suggested `book-keeping' value...see below...parameter ignored)
NSTAT < 0 produces SCF calculations for |NSTAT| states.
Remarks
The lowest 10 (NSTAT = -10) SCF hartree product functions are determined from a knowledge of the values of `omega' resulting from the Normal Mode analysis (INORM > 0). These 10 SCF states are then determined to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes), and is ignored on input, although an input value of NPOT = 0 is suggested...see below. When using the Normal Mode facility (INORM > 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are omitted from the input parameters. Also, when using the Normal Mode facility (INORM > 0), the NORMAL COORDINATE DISPLACEMENTS are omitted from the input parameters.
2.2 H2CS TEST....SCF FOR 10 LOWEST J=0 STATES WITH USER-SUPPLIED NORMAL MODES
Test SCF.2
Tests the calculation of the ten lowest-energy J=0 SCF states, using Normal Modes supplied by the user.
Key input parameters
ISCFCI = 0 forces `multimode' along the required SCF route.
INORM = 0 indicates Normal Modes to be supplied by the user.
NPOT = NMODE (suggested `book-keeping' value...see below...parameter ignored)
NSTAT < 0 produces SCF calculations for |NSTAT| states.
Remarks
The lowest 10 (NSTAT = -10) SCF hartree product functions are determined from a knowledge of the values of `omega' supplied by the user (INORM = 0). These 10 SCF states are then determined to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes (NMODE)), and is ignored on input. When the Normal Modes are supplied by the user (INORM = 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are included in the input parameters. Also, when the Normal Modes are supplied by the user (INORM = 0), the NORMAL COORDINATE DISPLACEMENTS are included in the input parameters.
2.3 H2CS TEST....SCF FOR 10 SPECIFIC INPUT J=0 STATES WITH INPUT NORMAL MODES
Test SCF.3
Tests the calculation of ten specific J=0 SCF states, using Normal Modes supplied by the user.
Key input parameters
ISCFCI = 0 forces `multimode' along the required SCF route.
INORM = 0 indicates Normal Modes to be supplied by the user.
NPOT = NMODE (suggested `book-keeping' value...see below...parameter ignored)
NSTAT > 0 produces SCF calculations for NSTAT states (see also:
SCF STATE DEFINITIONS below).
Remarks
The 10 specific (NSTAT = 10) SCF hartree product functions are determined from a knowledge of the input parameters: SCF STATE DEFINITIONS. The initial parameter (< 0) indicates that NSTAT = 10 specific Normal Mode representations are to be supplied. These 10 SCF states are then determined to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes (NMODE)), and is ignored on input. When the Normal Modes are supplied by the user (INORM = 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are included in the input parameters. Also, when the Normal Modes are supplied by the user (INORM = 0), the NORMAL COORDINATE DISPLACEMENTS are included in the input parameters.
2.4 H2CS TEST....SCF FOR 10 DEFAULT INPUT J=0 STATES WITH INPUT NORMAL MODES
Test SCF.4
Tests the calculation of ten default J=0 SCF states, using Normal Modes supplied by the user.
Key input parameters
ISCFCI = 0 forces `multimode' along the required SCF route.
INORM = 0 indicates Normal Modes to be supplied by the user.
NPOT = NMODE (suggested `book-keeping' value...see below...parameter ignored)
NSTAT > 0 produces SCF calculations for NSTAT states (see also:
SCF STATE DEFINITIONS below).
Remarks
The 10 default (NSTAT = 10) SCF hartree product functions are determined from a knowledge of the input parameters: SCF STATE DEFINITIONS. The initial parameter (= 0) indicates that NSTAT = 10 default representations are to be supplied as 10 records of 2 indices. The first index is the number of quanta in the mode given by the second index. All remaining quanta are defaulted to the value of the initial parameter (= 0). These 10 SCF states are then determined to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes (NMODE)), and is ignored on input. When the Normal Modes are supplied by the user (INORM = 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are included in the input parameters. Also, when the Normal Modes are supplied by the user (INORM = 0), the NORMAL COORDINATE DISPLACEMENTS are included in the input parameters.
2.5 H2CS TEST....SCF FOR 10 DEFAULT INPUT J=1 STATES WITH INPUT NORMAL MODES
Test SCF.5
Tests the calculation of ten default J=1 SCF states, using Normal Modes supplied by the user.
Key input parameters
JMAX = 1 indicates total angular momentum of J=1
ISCFCI = 0 forces `multimode' along the required SCF route.
INORM = 0 indicates Normal Modes to be supplied by the user.
NPOT = NMODE (suggested `book-keeping' value...see below...parameter ignored)
NSTAT > 0 produces SCF calculations for NSTAT states (see also:
SCF STATE DEFINITIONS below).
Remarks
The 10 default (NSTAT = 10) SCF hartree product functions are determined from a knowledge of the input parameters: SCF STATE DEFINITIONS. The initial parameter (= 0) indicates that NSTAT = 10 default representations are to be supplied as 10 records of 2 indices. The first index is the number of quanta in the mode given by the secondd index. All remaining quanta are defaulted to the value of the initial parameter (= 0). These 10 J=0 and J=JMAX SCF states are then determined to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). The J > 0 energies are calculated by the `Adiabatic Rotation' approximation, as explained in the Manual. For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes (NMODE)), and is ignored on input. When the Normal Modes are supplied by the user (INORM = 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are included in the input parameters. Also, when the Normal Modes are supplied by the user (INORM = 0), the NORMAL COORDINATE DISPLACEMENTS are included in the input parameters.
3.1 H2CS TEST....SCFCI FOR THE LOWEST 40 J=0 SCF STATES
Test SCFCI.1
Tests the CI mixing of the fourty lowest-energy J=0 SCF states, using Normal Modes determined by `multimode'.
Key input parameters
ISCFCI > 0 forces `multimode' along the CI route.
INORM > 0 indicates Normal Modes to be constructed by `multimode'.
NPOT = 0 (suggested `book-keeping' value...see below...parameter ignored)
NSTAT < 0 produces SCF calculations for |NSTAT| states.
ICI > 0 selects SCFCI option.
Remarks
The lowest 50 (NSTAT = -50) SCF hartree product functions are determined from a knowledge of the values of `omega' resulting from the Normal Mode analysis (INORM > 0). The first 40 of these SCF states are then determined (ICI = 40) to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). These 40 states are symmetry-blocked according to the symmetry parameters NVSYM,...,NSYM,...,MVSYM and mixed by non-orthogonal CI. The first 10 CI energies (ISCFCI = 10) of each block are printed. For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes), and is ignored on input, although an input value of NPOT = 0 is suggested...see below. When using the Normal Mode facility (INORM > 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are omitted from the input parameters. Also, when using the Normal Mode facility (INORM > 0), the NORMAL COORDINATE DISPLACEMENTS are omitted from the input parameters.
3.2 H2CS TEST....SCFCI FOR THE LOWEST 20 ADIABATIC J=1 SCF STATES WITH REAL*8 GRIDS
Test SCFCI.2
Tests the CI mixing of the twenty lowest-energy J=1 SCF states, using Normal
Modes determined by `multimode'.
Key input parameters
ISCFCI > 0 forces `multimode' along the CI route.
JMAX = 1 indicates total angular momentum of J=1
INORM > 0 indicates Normal Modes to be constructed by `multimode'.
NPOT = 0 (suggested `book-keeping' value...see below...parameter ignored)
NSTAT < 0 produces SCF calculations for |NSTAT| states.
ICI > 0 selects SCFCI option.
Remarks
The lowest 50 (NSTAT = -50) SCF hartree product functions are determined from a knowledge of the values of `omega' resulting from the Normal Mode analysis (INORM > 0). The first 20 of these SCF states are then determined (ICI = 20) to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). The SCF procedure is performed for J=0 and the 2J+1 values of K (the rigid-rotor eigenvalues) in the Adiabatic rotation scheme. For J=1 (JMAX = 1) these 40 states are symmetry-blocked for each K according to the symmetry parameters NVSYM,...,NSYM,...,MVSYM and mixed by non-orthogonal CI. The first 10 CI energies (ISCFCI = 10) of each block are printed relative to the J=0 zero point energy (EVLJ0 = 5375.4543 cm-1). For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes), and is ignored on input, although an input value of NPOT = 0 is suggested...see below. When using the Normal Mode facility (INORM > 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are omitted from the input parameters. Also, when using the Normal Mode facility (INORM > 0), the NORMAL COORDINATE DISPLACEMENTS are omitted from the input parameters.
3.3 H2CS TEST....SCFCI FOR THE LOWEST 20 ADIABATIC J=1 SCF STATES WITH REAL*4 GRIDS
Test SCFCI.3
Tests the CI mixing of the twenty lowest-energy J=1 SCF states, using Normal Modes determined by `multimode' and REAL*4 disc grids.
Key input parameters
ISCFCI > 0 forces `multimode' along the CI route.
ICOUPL < 0 uses REAL*4 instead of REAL*8 disc grids.
JMAX = 1 indicates total angular momentum of J=1.
INORM > 0 indicates Normal Modes to be constructed by `multimode'.
NPOT = 0 (suggested `book-keeping' value...see below...parameter ignored)
NSTAT < 0 produces SCF calculations for |NSTAT| states.
ICI > 0 selects SCFCI option.
Remarks
The lowest 50 (NSTAT = -50) SCF hartree product functions are determined from a knowledge of the values of `omega' resulting from the Normal Mode analysis (INORM > 0). The first 20 of these SCF states are then determined (ICI = 20) to a tolerance of CONV = 1.D-2 cm-1, for a maximum of ICOUPL = 3-mode coupling of the potential in internal coordinates (IWHICH > 0). The SCF procedure is performed for J=0 and the 2J+1 values of K (the rigid-rotor eigenvalues) in the Adiabatic rotation scheme. For J=1 (JMAX = 1) these 40 states are symmetry-blocked for each K according to the symmetry parameters NVSYM,...,NSYM,...,MVSYM and mixed by non-orthogonal CI. The first 10 CI energies (ISCFCI = 10) of each block are printed relative to the J=0 zero point energy (EVLJ0 = 5375.4543 cm-1). For potentials in internal coordinates (IWHICH > 0), the parameter NPOT is allocated by `multimode' to the number of Normal Modes of the system (see Manual for the use of NATOM and NOTTR to define the number of modes), and is ignored on input, although an input value of NPOT = 0 is suggested...see below. When using the Normal Mode facility (INORM > 0), the NPOT values of `omega' input as IPOT,JPOT,CPOT are omitted from the input parameters. Also, when using the Normal Mode facility (INORM > 0), the NORMAL COORDINATE DISPLACEMENTS are omitted from the input parameters.
4.1 H2CS TEST....VCI FOR 3 VIRTUAL QUANTA, MAXIMUM SUM OF 3 QUANTA
Test VCI.1
Tests the CI mixing of the three lowest quanta in each mode (restricted to a maximum sum of three quanta) of the virtual basis for the SCF zpe level.
Key input parameters
ISCFCI > 0 forces `multimode' along the CI route.
NSTAT < 0 produces SCF calculations for lowest-energy |NSTAT| states.
ICI < 0 selects VCI option.
NMAX > 0 constructs the sum-restricted VCI basis.
Remarks
The lowest 50 (NSTAT = -50) SCF hartree product functions are determined from a knowledge of the values of `omega' resulting from the Normal Mode analysis (INORM > 0). The virtual functions of ONLY the first (zpe) are used in the ensuing VCI calculation (note: NSTAT = -1 produces an identical virtual basis). The lowest 3 quanta (4 functions) of each 1-mode virtual basis are used to construct the orthonormal CI basis (ICI = -3), subject to the constraint that the total sum of quanta is not greater than 3 (NMAX = 3). The first 10 CI energies (ISCFCI = 10) of each block are printed, and the assignments include the states with the 2 largest CI coefficients (IPRINT = -2).
4.2 H2CS TEST....VCI FOR 3 VIRTUAL QUANTA, UNRESTRICTED
Test VCI.2
Tests the CI mixing of the three lowest quanta in each mode (unrestricted) of the virtual basis for the SCF zpe level.
Key input parameters
ISCFCI > 0 forces `multimode' along the CI route.
NSTAT < 0 produces SCF calculations for lowest-energy |NSTAT| states.
ICI < 0 selects VCI option.
NMAX = 0 constructs the unrestricted VCI basis.
Remarks
The lowest 50 (NSTAT = -50) SCF hartree product functions are determined from a knowledge of the values of `omega' resulting from the Normal Mode analysis (INORM > 0). The virtual functions of ONLY the first (zpe) are used in the ensuing VCI calculation (note: NSTAT = -1 produces an identical virtual basis). The lowest 3 (ICI = -3) quanta (4 functions) of each 1-mode virtual basis are used to construct the unrestricted orthonormal CI basis (NMAX = 0). The first test (2a) merely investigates the sizes of the CI symmetry blocks resulting from this algorithm (MATSIZ > 0) and terminates. The second test (2b) uses this basis in a VCI calculation (MATSIZ = 0). The first 10 CI energies (ISCFCI = 10) of each block are printed, and the assignments include the states with the 2 largest CI coefficients (IPRINT = -2).
4.3 H2CS TEST....VCI FOR MAXIMUM OF 3 COUPLED VIRTUAL QUANTA, RESTRICTED
Test VCI.3
Tests the CI mixing of 1, 2, 3 coupled modes in a selective virtual basis.
Key input parameters
ICI < 0 selects VCI option.
NMAX < 0 constructs the selective CI basis.
MAXSUM maximum sums of quanta for selective basis.
MAXBAS maximum quanta for selective basis.
Remarks
The VCI matrix is constructed from 1-, 2-, 3-mode coupled zpe virtual functions (NMAX = -3). The maximum sum of quanta in the 1-mode, 2-mode, 3-mode bases is 3 (MAXSUM(1) = MAXSUM(2) = MAXSUM(3) = 3). For each mode, the maximum quantum in the 1-mode basis is 3 (MAXBAS(1,1) = ... = MAXBAS(NMODE,1) = 3). For each mode, the maximum quantum in the 2-mode basis is 2 (MAXBAS(1,2) = ... = MAXBAS(NMODE,2) = 2). For each mode, the maximum quantum in the 3-mode basis is 1 (MAXBAS(1,3) = ... = MAXBAS(NMODE,3) = 1).
4.4 H2CS TEST....VCI FOR MAXIMUM OF 4 COUPLED VIRTUAL QUANTA, RESTRICTED
Test VCI.4
Tests the CI mixing of 1, 2, 3, 4 coupled modes in a selective virtual basis. Key input parameters
------------
ICI < 0 selects VCI option.
NMAX < 0 constructs the selective CI basis.
MAXSUM maximum sums of quanta for selective basis.
MAXBAS maximum quanta for selective basis.
Remarks
The VCI matrix is constructed from 1-, 2-, 3-, 4-mode coupled zpe virtual functions (NMAX = -4). The maximum sum of quanta in the 1-mode basis is 4 (MAXSUM(1) = 4). The maximum sum of quanta in the 2-mode basis is 5 (MAXSUM(2) = 5). The maximum sum of quanta in the 3-mode basis is 6 (MAXSUM(3) = 6). The maximum sum of quanta in the 4-mode basis is 7 (MAXSUM(3) = 7). For each mode, the maximum quantum in the 1-mode basis is 4 (MAXBAS(1,1) = ... = MAXBAS(NMODE,1) = 4). For each mode, the maximum quantum in the 2-mode basis is 3 (MAXBAS(1,2) = ... = MAXBAS(NMODE,2) = 3). For each mode, the maximum quantum in the 3-mode basis is 2 (MAXBAS(1,3) = ... = MAXBAS(NMODE,3) = 2). For each mode, the maximum quantum in the 4-mode basis is 1 (MAXBAS(1,4) = ... = MAXBAS(NMODE,4) = 1).
4.5 H2CS TEST....J=5 VCI FOR MAXIMUM OF 4 COUPLED VIRTUAL QUANTA, RESTRICTED
Test VCI.5
Performs a J=5 VCI calculation using 1, 2, 3, 4 coupled modes in a selective virtual basis.
Key input parameters
ICI < 0 selects VCI option.
NMAX < 0 constructs the selective CI basis.
MAXSUM maximum sums of quanta for selective basis.
MAXBAS maximum quanta for selective basis.
JMAX > 0 carries out exact Whitehead-Handy rotational analysis.
Remarks
The VCI matrix is constructed from 1-, 2-, 3-, 4-mode coupled zpe virtual functions (NMAX = -4). The maximum sum of quanta in the 1-mode basis is 4 (MAXSUM(1) = 4). The maximum sum of quanta in the 2-mode basis is 5 (MAXSUM(2) = 5). The maximum sum of quanta in the 3-mode basis is 6 (MAXSUM(3) = 6). The maximum sum of quanta in the 4-mode basis is 7 (MAXSUM(3) = 7). For each mode, the maximum quantum in the 1-mode basis is 4 (MAXBAS(1,1) = ... = MAXBAS(NMODE,1) = 4). For each mode, the maximum quantum in the 2-mode basis is 3 (MAXBAS(1,2) = ... = MAXBAS(NMODE,2) = 3). For each mode, the maximum quantum in the 3- mode basis is 2 (MAXBAS(1,3) = ... = MAXBAS(NMODE,3) = 2). For each mode, the maximum quantum in the 4-mode basis is 1 (MAXBAS(1,4) = ... = MAXBAS(NMODE,4) = 1). The VCI calculation is carried out using the exact algorithm for J=5 (JMAX = 5) where only the initial (Ka=0) K-diagonal basis is used (KSTEP = 0) to form the final rovibrational matrix. The energies are given relative to the J=0 zpe (EVLJ0 > 0).
4.6 H2CS TEST....J=5 VCI USING ALTERNATE K-DIAGONAL STEPS
Test VCI.6
Performs a J=5 VCI calculation using 1, 2, 3, 4 coupled modes in a selective virtual basis. K-diagonal steps are performed for alternate combinations of Ka,Kc.
Key input parameters
ICI < 0 selects VCI option.
NMAX < 0 constructs the selective CI basis.
MAXSUM maximum sums of quanta for selective basis.
MAXBAS maximum quanta for selective basis.
JMAX > 0 carries out exact Whitehead-Handy rotational analysis.
KSTEP > 0 indicates the increment in Ka,Kc for the K-diagonal steps.
Remarks
The VCI matrix is constructed from 1-, 2-, 3-, 4-mode coupled zpe virtual functions (NMAX = -4). The maximum sum of quanta in the 1-mode basis is 4 (MAXSUM(1) = 4). The maximum sum of quanta in the 2-mode basis is 5 (MAXSUM(2) = 5). The maximum sum of quanta in the 3-mode basis is 6 (MAXSUM(3) = 6). The maximum sum of quanta in the 4-mode basis is 7 (MAXSUM(3) = 7). For each mode, the maximum quantum in the 1-mode basis is 4 (MAXBAS(1,1) = ... = MAXBAS(NMODE,1) = 4). For each mode, the maximum quantum in the 2-mode basis is 3 (MAXBAS(1,2) = ... = MAXBAS(NMODE,2) = 3). For each mode, the maximum quantum in the 3-mode basis is 2 (MAXBAS(1,3) = ... = MAXBAS(NMODE,3) = 2). For each mode, the maximum quantum in the 4-mode basis is 1 (MAXBAS(1,4) = ... = MAXBAS(NMODE,4) = 1). The VCI calculation is carried out using the exact algorithm for J=5 (JMAX = 5) where alternate values of Ka,Kc are used to form the K-diagonal basis functions (KSTEP = 2) that go to form the final rovibrational matrix. The energies are given relative to the J=0 zpe (EVLJ0 > 0).
5.1 H2CS TEST....J=5 VCI USING ALTERNATE K-DIAGONAL STEPS AND DAVIDSON/LIU
Test LAN.1
Performs a J=5 VCI calculation using 1, 2, 3, 4 coupled modes in a selective virtual basis. K-diagonal steps are performed for alternate combinations of Ka,Kc. The CI matrices are diagonalised with the iterative Davidson/Liu procedure.
Key input parameters
ICI < 0 selects VCI option.
NMAX < 0 constructs the selective CI basis.
MAXSUM maximum sums of quanta for selective basis.
MAXBAS maximum quanta for selective basis.
NCYCLE > 0 number of iterations in Davidson/Liu diagonalisation.
TOLLAN > 0 tolerance of eigenvalues required.
LANMAX > 0 order of half-size matrix used to set up hamiltonian matrix.
Remarks
The VCI matrix is constructed from 1-, 2-, 3-, 4-mode coupled zpe virtual functions (NMAX = -4). The maximum sum of quanta in the 1-mode basis is 4 (MAXSUM(1) = 4). The maximum sum of quanta in the 2-mode basis is 5 (MAXSUM(2) = 5). The maximum sum of quanta in the 3-mode basis is 6 (MAXSUM(3) = 6). The maximum sum of quanta in the 4-mode basis is 7 (MAXSUM(3) = 7). For each mode, the maximum quantum in the 1-mode basis is 4 (MAXBAS(1,1) = ... = MAXBAS(NMODE,1) = 4). For each mode, the maximum quantum in the 2-mode basis is 3 (MAXBAS(1,2) = ... = MAXBAS(NMODE,2) = 3). For each mode, the maximum quantum in the 3-mode basis is 2 (MAXBAS(1,3) = ... = MAXBAS(NMODE,3) = 2). For each mode, the maximum quantum in the 4-mode basis is 1 (MAXBAS(1,4) = ... = MAXBAS(NMODE,4) = 1). The VCI calculation is carried out using the exact algorithm for J=5 (JMAX = 5) where alternate values of Ka,Kc are used to form the K-diagonal basis functions given relative to the J=0 zpe (EVLJ0 > 0). The Davidson/Liu diagonalisation scheme is used at the CI stages, each diagonalisation performing a maximum of 10 (NCYCLE = 10) iterations for a tolerance of TOLLAN = 1.D-3 cm-1 in the eigenvalues. The hamiltonian matrices are built up in a half-size work matrix whose order is 3000 (LANMAX = 3000).
5.2 H2CS TEST....J=0 VCI USING DAVIDSON/LIU WITH TOO SMALL A WORK ARRAY
Test LAN.2
Performs a J=0 VCI calculation using 1, 2, 3, 4 coupled modes in a selective virtual basis. The CI matrix is diagonalised with the iterative Davidson/Liu procedure. The work array is too small to hold the complete hamiltonian matrix.
Key input parameters
ICI < 0 selects VCI option.
NMAX < 0 constructs the selective CI basis.
MAXSUM maximum sums of quanta for selective basis.
MAXBAS maximum quanta for selective basis.
NCYCLE > 0 number of iterations in Davidson/Liu diagonalisation.
TOLLAN > 0 tolerance of eigenvalues required.
LANMAX > 0 order of half-size matrix used to set up hamiltonian matrix.
Remarks
The VCI matrix is constructed from 1-, 2-, 3-, 4-mode coupled zpe virtual functions (NMAX = -4). The maximum sum of quanta in the 1-mode basis is 4 (MAXSUM(1) = 4). The maximum sum of quanta in the 2-mode basis is 5 (MAXSUM(2) = 5). The maximum sum of quanta in the 3-mode basis is 6 (MAXSUM(3) = 6). The maximum sum of quanta in the 4-mode basis is 7 (MAXSUM(3) = 7). For each mode, the maximum quantum in the 1-mode basis is 4 (MAXBAS(1,1) = ... = MAXBAS(NMODE,1) = 4). For each mode, the maximum quantum in the 2-mode basis is 3 (MAXBAS(1,2) = ... = MAXBAS(NMODE,2) = 3). For each mode, the maximum quantum in the 3-mode basis is 2 (MAXBAS(1,3) = ... = MAXBAS(NMODE,3) = 2). For each mode, the maximum quantum in the 4-mode basis is 1 (MAXBAS(1,4) = ... = MAXBAS(NMODE,4) = 1). The Davidson/Liu diagonalisation scheme is used at the CI stages, each diagonalisation performing a maximum of 10 (NCYCLE = 10) iterations for a tolerance of TOLLAN = 1.D-3 cm-1 in the eigenvalues. The hamiltonian matrices are built up in a half-size work matrix whose order is 100 (LANMAX = 100). In some cases, this is not big enough to hold the complete array.
5.3 H2CS TEST....J=5 VCI USING ALTERNATE K-DIAGONAL STEPS AND DAVIDSON/LIU
Test LAN.3
Performs a J=5 VCI calculation using 1, 2, 3, 4 coupled modes in a selective virtual basis. K-diagonal steps are performed for alternate combinations of Ka,Kc. The CI matrices are diagonalised with the iterative Davidson/Liu procedure. The work array is too small to hold the complete hamiltonian matrix.
Key input parameters
ICI < 0 selects VCI option.
NMAX < 0 constructs the selective CI basis.
MAXSUM maximum sums of quanta for selective basis.
MAXBAS maximum quanta for selective basis.
NCYCLE > 0 number of iterations in Davidson/Liu diagonalisation.
TOLLAN > 0 tolerance of eigenvalues required.
LANMAX > 0 order of half-size matrix used to set up hamiltonian matrix.
Remarks
The VCI matrix is constructed from 1-, 2-, 3-, 4-mode coupled zpe virtual functions (NMAX = -4). The maximum sum of quanta in the 1-mode basis is 4 (MAXSUM(1) = 4). The maximum sum of quanta in the 2-mode basis is 5 (MAXSUM(2) = 5). The maximum sum of quanta in the 3-mode basis is 6 (MAXSUM(3) = 6). The maximum sum of quanta in the 4-mode basis is 7 (MAXSUM(3) = 7). For each mode, the maximum quantum in the 1-mode basis is 4 (MAXBAS(1,1) = ... = MAXBAS(NMODE,1) = 4). For each mode, the maximum quantum in the 2-mode basis is 3 (MAXBAS(1,2) = ... = MAXBAS(NMODE,2) = 3). For each mode, the maximum quantum in the 3-mode basis is 2 (MAXBAS(1,3) = ... = MAXBAS(NMODE,3) = 2). For each mode, the maximum quantum in the 4-mode basis is 1 (MAXBAS(1,4) = ... = MAXBAS(NMODE,4) = 1). The VCI calculation is carried out using the exact algorithm for J=5 (JMAX = 5) where alternate values of Ka,Kc are used to form the K-diagonal basis functions (KSTEP = 2) that go to form the final rovibrational matrix. The energies are given relative to the J=0 zpe (EVLJ0 > 0). The Davidson/Liu diagonalisation scheme is used at the CI stages, each diagonalisation performing a maximum of 10 (NCYCLE = 10) iterations for a tolerance of TOLLAN = 1.D-3 cm-1 in the eigenvalues. The hamiltonian matrices are built up in a half-size work matrix whose order is 100 (LANMAX = 100). In some cases, this is not big enough to hold the complete array.
|
|
|
|
|