QuantumBio, Inc.

A Quantum Breakthrough in Drug Discovery

• Home
• Company
• Products
• Support
  • Software Manuals
    • DivCon 4.5
    • tmp
  • Select Publications
  • Mailing Lists
• News & Events

• Home > 
• Support > 
• Software Manuals > 

General Quantum Mechanics

CARTESIAN

Cartesian coordinate format. DivCon reads Cartesian coordinates in the format shown in the following example:

1  N -0.26120 -0.98976 0.00000

2  C  0.64694 0.01940 0.00000

3  C -0.47100 1.06738 0.00000

4 C -1.44202 -0.13945 0.00000

5 O 1.83331 0.04003 0.00000

6 H -0.13870 -1.97802 0.00000

7 H -0.49385 1.68899 -0.88436

8 H -0.49385 1.68899 0.88436

9 H -2.05887 -0.23715 -0.88402

10 H -2.05887 -0.23715 0.88402

Coordinates are in Å. The specification of symbols and coordinates is format free and the maximum characters per line is 80.

RMIN=float

The minimum allowed distance between atoms (results in an error for single point calculations and geometry optimizations.

ECRIT=float

Used to set the convergence for the energy in units of eV. (default value is 4·10-6 eV). The actual value of ECRIT will be relaxed if the gradient norm is large and the structure is not tiny. This should speed up convergence without loss of accuracy. The DESCF keyword can be used instead of ECRIT in cases where this relaxation is undesired.

DCRIT=float

Used to set the convergence for the density matrix in atomic units (default value 5·10-4 e). The actual value of DCRIT will be relaxed if the gradient norm is large and the structure is not tiny. This should speed up convergence without any loss of accuracy. The DPSCF keyword can be used instead of DCRIT in cases where this relaxation is undesired.

DESCF=float

This keyword can be used instead of ECRIT as it defines the SCF energy convergence criterion. However, unlike ECRIT, this value is considered absolute (in eV). In affect, the SCF calculation will not stop until this criterion is reached.

DPSCF=float

This keyword can be used instead of DCRIT in that it defines the SCF energy convergence criterion. However, unlike DCRIT, this value is considered absolute (in eV). In affect, the SCF calculation will not stop until this criterion is reached.

CUTREPUL=float

Set the [xy|xy], [xz|xz], [xx|yy], [xx|zz], [zz|xx], [xx|xx] and [zz|zz] integrals to zero when the interatomic distance is larger than float. The CUTPREPUL keyword can be used to speed up a DivCon simulation by limited the number of calculations performed.

CUTBOND=float

Cutoff bonding for the H, P and F matrixes beyond float angstroms. The CUTBOND keyword can be used to speed up a DivCon simulation by limiting the number of calculations performed.

DIRECT

Using the DIRECT keyword will cause all 2-electron integrals to be kept in memory and recalculated at each step instead of being written out to file. This is the suggested approach as generally with how fast processors are today and how much memory users have at their disposal, accessing disk will actually be more expensive.

FULLSCF

In order to save in CPU time, there are times when the calculation does not need to perform a full diagonalization and instead performs a pseudo diagonalization. This keyword will turn off that functionality, and it will not allow DivCon to utilize this time saving function call.

RESIDUE

During a divide and conquer calculation, the system is divided according to residue delimiters. The ?RESIDUE? keyword is used to keep these residue pointers within DivCon. This keyword also requires that the user denotes the beginning of each residue in the divcon.in file by using the ?RES? delimiter immediately after the ?z? coordinate in each line.

CHKRES

Check inter-atomic distances for each residue.

TEMPK=float

The user may define a temperature for D&C. Units are Kelvin, and default is 1000K.

TESTRUN (-test)

Do setup work and stop before first energy evaluation.

TMAX=float

User defined maximum CPU time in seconds. The default limit is 2020 seconds.

SHIFT=float (-s ?)

User defined initial dynamic level shift[] parameter in eV.

XYZSPACE

Perform all operations in xyz space.

MAXIT=int

Set the maximum number of SCF iterations (default: 100). If it takes more then 100 SCF iterations to converge, it is generally thought that the system will probably not converge and is exhibiting problems.

DOUBLE=int

Do a double SCF step during a certain number (int) of SCF iterations. If int=0, then we do a double SCF step for every SCF iteration (only for non-geometry optimizations). This keyword will aid in convergence as it guarantees that the values calculated at each step are based completely on the current step, and not on the previous step. By default, the first step of an SCF calculation will be a double. Please note, using DOUBLE will significantly increase the amount of CPU time required to execute a DivCon job (assuming no difficulties in convergence).

1SCF

Perform only the first SCF iteration, i.e. calculate the energy through E = 0.5(H+F)P. Note that this is not equivalent to MAXIT=1, since no diagonalization is performed.

GUESS

Build the initial density matrix from one or more density matrix files. These files are listed by means of the GUESS parameter block as seen below. If no atom numbers are provided, the matrices are appended to one another to yield the starting matrix. The SHIFT keyword (see page ) in the following parameter will invoke the shift keyword for the simulation based upon the value of the shift stored in the density matrix. This keyword will allow the calculation to begin at the approximate point where the original guess density matrix was written. Use of the SHIFT keyword is especially useful when A.dmx and b.dmx are from a protein and an inhibitor respectively.

GUESS

  A.dmx SHIFT

  b.dmx

END_GUESS

Build the initial density matrix from the files A.dmx and b.dmx. The density matrix elements of atoms 1 through a are read from A.dmx, for atoms a+1 through n from file b.dmx.

GUESS

  2-10 A.dmx 33-41 f

  20-30 b.dmx 1-11

END_GUESS

Density matrix information for atoms 2-10 is read from the density matrix elements of atoms 33-41 of file A.dmx, density matrix information for atoms 20-30 is read from the density matrix elements of atoms 1-11 from file b.dmx. Missing density matrix elements are auto-initialized and a correction will be applied to constrain the total number of electrons. The density matrix elements of atoms 2-10 will be kept constant during the SCF iterations by using the Frozen Density Matrix approximation.[] It is imperative that the number of orbitals on a certain atom in the divcon.in file and density matrix file are the same, i.e. the number of orbitals on atom 2 from divcon.in and atom 33 from A.dmx should be identical.

Note that the maximum length of a density matrix file name is 20 characters, no dashes (?-?) are allowed in the density matrix file name.

PUSH

PUSH can only be used in combination with the CLUSTER keyword (page ) and when multiple cores (i.e. multiple cluster groups) are defined. Push the different cluster groups apart. The default value is 106 Å, and the user can define this distance by using PUSH=float. When more than two cluster groups are defined, each group is place on a grid point with grid spacing of 106 Å or the user defined value.

INTGLS=string

The INTGLS keyword can have two values. If the value is INTGLS=TALMAN then the integrals utilized using the DivCon calculation will use the Talman method. If the value is INTGLS=STEWART then the Steward integrals will be used. The default, and the recommended value, is INTGLS=TALMAN as this approach has been found to be most stable and accurate.

© 2001-2007 QuantumBio, Inc.