CamCASP/Interfaces - Revision history

import>Am592: /* Interface Details */

2011-12-08T17:54:55Z

Interface Details

@@ Line 202: / Line 202: @@
 MO(1) = D(1,1)*G(3d2-) + D(2,1)*G(3d1-) + D(3,1)*G(3d0) + ...
 </pre>
 The interface code in '''$CAMCASP/interfaces/gamessus/''' is based on the '''gamsintf.F''' code from SAPT2006, but has been heavily modified. It's infinitely better!

import>Am592: /* Interface Details */

2011-12-08T17:45:23Z

Interface Details

@@ Line 102: / Line 102: @@
     !I have encountered before. The following factors need to be included in the
     !transformation matrices:
     !  (3a) All coefficients in front of x^i y^j z^k get
     !  multiplied by sqrt[(2i-1)!!(2j-1)!!(2k-1)!!]
@@ Line 176: / Line 174: @@
    rt3 = sqrt(3) and blank terms are zeros.
 </pre>
 The interface code in '''$CAMCASP/interfaces/gamessus/''' is based on the '''gamsintf.F''' code from SAPT2006, but has been heavily modified. It's infinitely better!

import>Am592: /* Interface Details */

2011-12-08T17:33:41Z

Interface Details

← Older revision		Revision as of 17:33, 8 December 2011
Line 165:		Line 165:
	Ie., we get a 6 x 5 matrix C of MO coefficient. There are only 5 MOs, but 6 coefficients for each. What we need to do is determine a transformation matrix that transforms each of the sets of 6 coefficients into a set containing only 5, which would then stand in front of the fully normalized spherical GTOs used by CamCASP. To do this we use subroutine trans3 to define the matrix in the following manner.		Ie., we get a 6 x 5 matrix C of MO coefficient. There are only 5 MOs, but 6 coefficients for each. What we need to do is determine a transformation matrix that transforms each of the sets of 6 coefficients into a set containing only 5, which would then stand in front of the fully normalized spherical GTOs used by CamCASP. To do this we use subroutine trans3 to define the matrix in the following manner.

	First define the usual sort of Cartesian to Spherical transformation matrix		First define the usual sort of Cartesian to Spherical transformation matrix. See details of basis sets given in section [[ CamCASP/Programming/1 \| Basis functions ]]. For the d-function this matrix is (called BT2 in the code fragment above):
	<pre>		<pre>
			xx yy zz xy xz yz

			BT2 = 1 3d2- [ rt3 ]
			--- x 3d1- [ rt3 ]
			sqrt(3) 3d0 [ -1/2 -1/2 1 ]
			3d1+ [ rt3 ]
			3d2+ [rt3/2 -rt3/2 ]
			where
			rt3 = sqrt(3) and blank terms are zeros.
	</pre>		</pre>

import>Am592: /* Interface Details */

2011-12-08T17:21:51Z

Interface Details

← Older revision		Revision as of 17:21, 8 December 2011
Line 81:		Line 81:
	GAMESS(US) uses Spherical GTOs in a slightly different way from other codes. The MO coefficients of its spherical GTOs are not supplied as standing in front of the components of the spherical GTOs, but rather they stand in front of the Cartesian components. To make this concrete, rather than supply 5 sets of 5 MO coefficients for a spherical d-function (each set of 5 would correspond to an MO), GAMESS(US) will supply 5 sets of '''6''' coefficients. Each of these 6 would stand in front of Cartesian GTOs. This is OK, but the normalization of these Cartesian GTOs is rather awkward. I have worked out the transformation matrices from these Cartesian GTOs to an equivalent spherical set.		GAMESS(US) uses Spherical GTOs in a slightly different way from other codes. The MO coefficients of its spherical GTOs are not supplied as standing in front of the components of the spherical GTOs, but rather they stand in front of the Cartesian components. To make this concrete, rather than supply 5 sets of 5 MO coefficients for a spherical d-function (each set of 5 would correspond to an MO), GAMESS(US) will supply 5 sets of '''6''' coefficients. Each of these 6 would stand in front of Cartesian GTOs. This is OK, but the normalization of these Cartesian GTOs is rather awkward. I have worked out the transformation matrices from these Cartesian GTOs to an equivalent spherical set.

			Here are some comments from subroutine trans3 which is contained in subroutine transform_mos in module molecular_orbitals:
⚫	The interface code is based on the '''gamsintf.F''' code from SAPT2006, but has been heavily modified. ~~More~~ ~~on this later.~~
			<pre>
			!(1) GAMESS(US) writes out coefficients of MOs of spherical GTOs not as they
			!stand infront of the components of spherical GTOs (which would be logical),
			!but as they stand in front of Cartesian GTOs. So we get 6 coefficients for
			!a d-function, 7 for an f, etc.
			!
			!(2) The order of the Cartesian functions is the
			!same as that in CamCASP. See module basis_trans_mats. This is not
			!surprising as we use the GAMINT interal module for our integrals. This
			!module is the same as that used by GAMESS(US) - apart from modifications
			!made by Wojtek.
			!
			!We can take care of points (1) and (2) by defining the transformation
			!matrices as the Cartesian to spherical transformation matrices.
			!This choice was made above by the call to
			! subroutine init_basis_trans_mats(2)
			!
			!(3) But the normalization of the Cartesian GTOs does not seem to be anything
			!I have encountered before. The following factors need to be included in the
			!transformation matrices:
			! (3a) All coefficients in front of x^i y^j z^k get
			! multiplied by sqrt[(2i-1)!!(2j-1)!!(2k-1)!!]
			! (3a) All coefficients in front of x^i y^j z^k get
			! multiplied by sqrt[(2i-1)!!(2j-1)!!(2k-1)!!]
			! (3b) The term x^i in (3a) gets multiplied by
			! 1 i=0
			! 1 i=1
			! 2/3 i=2
			! 2/5 i=3
			! 8/35 i=4
			! With similar multipliers for y^j and z^k.
			! I.e., xxy gets a multiplier of (2/3)*1 = 2/3
			! and xxx gets a multiplied of (2/5)
			! and xxyy gets (2/3)*(2/3) = 4/9
			</pre>
			And here's how this is implemented (part of subroutine trans3 follows):
			<pre>
			use basis_trans_mats, only : BT0, BT1, BT2, BT3, BT4
			...
			real(dp), dimension(0:4) :: gmsfac = (/ &
			1.0_dp, 1.0_dp, 2.0_dp/3.0_dp, 2.0_dp/5.0_dp, 8.0_dp/35.0_dp &
			/)
			...
			select case(l)
			case(0)
			transmat(1:ncomp_trans,1:ncomp) = (BT0)
			case(1)
			transmat(1:ncomp_trans,1:ncomp) = (BT1)
			case(2)
			transmat(1:ncomp_trans,1:ncomp) = (BT2)
			case(3)
			transmat(1:ncomp_trans,1:ncomp) = (BT3)
			case(4)
			transmat(1:ncomp_trans,1:ncomp) = (BT4)
			case default
			print *,'INTERNAL ERROR in ',trim(this_routine)
			print *,'Illegal value of l received is ',l
			info = -1
			return
			end select
			!
			!Get Cartesian symmetry components (force GAMINT ordering):
			call map_symm_to_powers(l,i,j,k,order=par_order_symm_comp)
			do m = 1, ncomp
			fac = dfactorial(2i(m)-1)dfactorial(2j(m)-1)dfactorial(2*k(m)-1)
			fac = sqrt(fac)gmsfac(i(m))gmsfac(j(m))*gmsfac(k(m))
			! transmat(i,m) = t(i,m) fac(m)
			transmat(1:ncomp_trans,m) = transmat(1:ncomp_trans,m) * fac
			enddo
			!
			</pre>
			The effect of this subroutine is best seen with an example. Consider a basis of a single Spherical d-function. GAMESS(US) will supply MO coefficients as
			<pre>
			MO1 MO2 MO3 MO4 MO5
			xx C(1,1) C(2,1) C(1,3) ... ...
			yy C(2,1) C(2,2) ... ... ...
			zz ...
			xy ...
			xz ...
			yz ... C(6,5)
			</pre>
			Ie., we get a 6 x 5 matrix C of MO coefficient. There are only 5 MOs, but 6 coefficients for each. What we need to do is determine a transformation matrix that transforms each of the sets of 6 coefficients into a set containing only 5, which would then stand in front of the fully normalized spherical GTOs used by CamCASP. To do this we use subroutine trans3 to define the matrix in the following manner.

			First define the usual sort of Cartesian to Spherical transformation matrix
			<pre>

			</pre>


		⚫	The interface code in '''$CAMCASP/interfaces/gamessus/''' is based on the '''gamsintf.F''' code from SAPT2006, but has been heavily modified. It's infinitely better!

	==Examples==		==Examples==

import>Am592: /* GAMESS(US) */

2011-12-08T16:58:38Z

GAMESS(US)

@@ Line 78: / Line 78: @@
 =GAMESS(US)=
 ==Examples==
 ===User-defined basis sets with ghost-functions===
@@ Line 166: / Line 172: @@
 </pre>
 The '1.0's in the angular momenta lines does not seem to be needed.

import>Am592 at 16:56, 8 December 2011

2011-12-08T16:56:52Z

← Older revision		Revision as of 16:56, 8 December 2011
Line 1:		Line 1:
	[[ CamCASP ]] => Interfaces		[[ CamCASP ]] => Interfaces

	This page describes details of interfaces to various SCF codes.		This page describes details of interfaces to various SCF codes. Also see details of basis sets given in section [[ CamCASP/Programming/1 \| Basis functions ]].

	=NWChem=		=NWChem=

import>Am592: /* User-defined basis sets with ghost-functions */

2011-12-08T16:48:42Z

User-defined basis sets with ghost-functions

@@ Line 109: / Line 109: @@
 3.2532260   0.001640
 1.6151790  -0.000616
      D  1 1.0
 0.840       1.0
@@ Line 214: / Line 149: @@
 938.2595    0.008882
 389.2710    0.024232
      D  2 1.0
 2.9532      0.18353
@@ Line 269: / Line 165: @@
  $MOROKM MOROKM=.FALSE. $END
 </pre>
 ==Interface Details==

import>Am592: /* User-defined basis sets */

2011-12-08T16:46:42Z

User-defined basis sets

← Older revision		Revision as of 16:46, 8 December 2011
Line 74:		Line 74:
	task scf		task scf
	</pre>		</pre>

			The [https://bse.pnl.gov/bse/portal EMSL Basis portal] supplies basis sets with the atom symbol in place of a name. So perhaps we can use the symbol instead. We will need a converter to convert our GAMESS(US) formatted basis sets to NWChem format.

	=GAMESS(US)=		=GAMESS(US)=

import>Am592: /* NWChem */

2011-12-08T16:42:50Z

NWChem

← Older revision		Revision as of 16:42, 8 December 2011
Line 5:		Line 5:
	=NWChem=		=NWChem=
	==Interface==		==Interface==
	==NWChem==
	All we've used NWChem with are spherical GTOs. It supplied MO coefficients as standing in front of fully normalized spherical GTOs (like any sensible code should!). But the sign of some of these is different:		All we've used NWChem with are spherical GTOs. It supplied MO coefficients as standing in front of fully normalized spherical GTOs (like any sensible code should!). But the sign of some of these is different:
	<pre>		<pre>

import>Am592: /* NWChem */

2011-12-08T16:42:31Z

NWChem

@@ Line 4: / Line 4: @@
 =NWChem=
 ==User-defined basis sets==
 <pre>