ViewVC Help

View File | Revision Log | Show Annotations | View Changeset | Root Listing

root/group/trunk/OOPSE-4/src/UseTheForce/DarkSide/electrostatic.F90

Comparing trunk/OOPSE-4/src/UseTheForce/DarkSide/electrostatic.F90 (file contents):
Revision 2296 by chrisfen, Thu Sep 15 00:13:56 2005 UTC vs.
Revision 2843 by chrisfen, Fri Jun 9 18:26:18 2006 UTC

#	Line 47 | Line 47 | module electrostatic_module
47		use vector_class
48		use simulation
49		use status
50	+	use interpolation
51		#ifdef IS_MPI
52		use mpiSimulation
53		#endif
#	Line 54 | Line 55 | module electrostatic_module
55
56		PRIVATE
57
58	+
59	+	#define __FORTRAN90
60	+	#include "UseTheForce/DarkSide/fInteractionMap.h"
61	+	#include "UseTheForce/DarkSide/fElectrostaticSummationMethod.h"
62	+	#include "UseTheForce/DarkSide/fElectrostaticScreeningMethod.h"
63	+
64	+
65		!! these prefactors convert the multipole interactions into kcal / mol
66		!! all were computed assuming distances are measured in angstroms
67		!! Charge-Charge, assuming charges are measured in electrons
#	Line 68 | Line 76 | module electrostatic_module
76		!! This unit is also known affectionately as an esu centi-barn.
77		real(kind=dp), parameter :: pre14 = 69.13373_dp
78
79	+	real(kind=dp), parameter :: zero = 0.0_dp
80	+
81	+	!! number of points for electrostatic splines
82	+	integer, parameter :: np = 100
83	+
84	+	!! variables to handle different summation methods for long-range
85	+	!! electrostatics:
86	+	integer, save :: summationMethod = NONE
87	+	integer, save :: screeningMethod = UNDAMPED
88	+	logical, save :: summationMethodChecked = .false.
89	+	real(kind=DP), save :: defaultCutoff = 0.0_DP
90	+	real(kind=DP), save :: defaultCutoff2 = 0.0_DP
91	+	logical, save :: haveDefaultCutoff = .false.
92	+	real(kind=DP), save :: dampingAlpha = 0.0_DP
93	+	real(kind=DP), save :: alpha2 = 0.0_DP
94	+	real(kind=DP), save :: alpha4 = 0.0_DP
95	+	real(kind=DP), save :: alpha6 = 0.0_DP
96	+	real(kind=DP), save :: alpha8 = 0.0_DP
97	+	logical, save :: haveDampingAlpha = .false.
98	+	real(kind=DP), save :: dielectric = 1.0_DP
99	+	logical, save :: haveDielectric = .false.
100	+	real(kind=DP), save :: constEXP = 0.0_DP
101	+	real(kind=dp), save :: rcuti = 0.0_DP
102	+	real(kind=dp), save :: rcuti2 = 0.0_DP
103	+	real(kind=dp), save :: rcuti3 = 0.0_DP
104	+	real(kind=dp), save :: rcuti4 = 0.0_DP
105	+	real(kind=dp), save :: alphaPi = 0.0_DP
106	+	real(kind=dp), save :: invRootPi = 0.0_DP
107	+	real(kind=dp), save :: rrf = 1.0_DP
108	+	real(kind=dp), save :: rt = 1.0_DP
109	+	real(kind=dp), save :: rrfsq = 1.0_DP
110	+	real(kind=dp), save :: preRF = 0.0_DP
111	+	real(kind=dp), save :: preRF2 = 0.0_DP
112	+	real(kind=dp), save :: erfcVal = 1.0_DP
113	+	real(kind=dp), save :: derfcVal = 0.0_DP
114	+	type(cubicSpline), save :: erfcSpline
115	+	logical, save :: haveElectroSpline = .false.
116	+	real(kind=dp), save :: c1 = 1.0_DP
117	+	real(kind=dp), save :: c2 = 1.0_DP
118	+	real(kind=dp), save :: c3 = 0.0_DP
119	+	real(kind=dp), save :: c4 = 0.0_DP
120	+	real(kind=dp), save :: c5 = 0.0_DP
121	+	real(kind=dp), save :: c6 = 0.0_DP
122	+	real(kind=dp), save :: c1c = 1.0_DP
123	+	real(kind=dp), save :: c2c = 1.0_DP
124	+	real(kind=dp), save :: c3c = 0.0_DP
125	+	real(kind=dp), save :: c4c = 0.0_DP
126	+	real(kind=dp), save :: c5c = 0.0_DP
127	+	real(kind=dp), save :: c6c = 0.0_DP
128	+	real(kind=dp), save :: one_third = 1.0_DP / 3.0_DP
129	+
130	+	#if defined(__IFC) || defined(__PGI)
131	+	! error function for ifc version > 7.
132	+	real(kind=dp), external :: erfc
133	+	#endif
134	+
135	+	public :: setElectrostaticSummationMethod
136	+	public :: setScreeningMethod
137	+	public :: setElectrostaticCutoffRadius
138	+	public :: setDampingAlpha
139	+	public :: setReactionFieldDielectric
140	+	public :: buildElectroSpline
141		public :: newElectrostaticType
142		public :: setCharge
143		public :: setDipoleMoment
#	Line 76 | Line 146 | module electrostatic_module
146		public :: doElectrostaticPair
147		public :: getCharge
148		public :: getDipoleMoment
79	–	public :: pre22
149		public :: destroyElectrostaticTypes
150	+	public :: self_self
151	+	public :: rf_self_excludes
152
153	+
154		type :: Electrostatic
155		integer :: c_ident
156		logical :: is_Charge = .false.
#	Line 94 | Line 166 | contains
166
167		type(Electrostatic), dimension(:), allocatable :: ElectrostaticMap
168
169	+	logical, save :: hasElectrostaticMap
170	+
171		contains
172
173	+	subroutine setElectrostaticSummationMethod(the_ESM)
174	+	integer, intent(in) :: the_ESM
175	+
176	+	if ((the_ESM .le. 0) .or. (the_ESM .gt. REACTION_FIELD)) then
177	+	call handleError("setElectrostaticSummationMethod", "Unsupported Summation Method")
178	+	endif
179	+
180	+	summationMethod = the_ESM
181	+
182	+	end subroutine setElectrostaticSummationMethod
183	+
184	+	subroutine setScreeningMethod(the_SM)
185	+	integer, intent(in) :: the_SM
186	+	screeningMethod = the_SM
187	+	end subroutine setScreeningMethod
188	+
189	+	subroutine setElectrostaticCutoffRadius(thisRcut, thisRsw)
190	+	real(kind=dp), intent(in) :: thisRcut
191	+	real(kind=dp), intent(in) :: thisRsw
192	+	defaultCutoff = thisRcut
193	+	defaultCutoff2 = defaultCutoff*defaultCutoff
194	+	rrf = defaultCutoff
195	+	rt = thisRsw
196	+	haveDefaultCutoff = .true.
197	+	end subroutine setElectrostaticCutoffRadius
198	+
199	+	subroutine setDampingAlpha(thisAlpha)
200	+	real(kind=dp), intent(in) :: thisAlpha
201	+	dampingAlpha = thisAlpha
202	+	alpha2 = dampingAlpha*dampingAlpha
203	+	alpha4 = alpha2*alpha2
204	+	alpha6 = alpha4*alpha2
205	+	alpha8 = alpha4*alpha4
206	+	haveDampingAlpha = .true.
207	+	end subroutine setDampingAlpha
208	+
209	+	subroutine setReactionFieldDielectric(thisDielectric)
210	+	real(kind=dp), intent(in) :: thisDielectric
211	+	dielectric = thisDielectric
212	+	haveDielectric = .true.
213	+	end subroutine setReactionFieldDielectric
214	+
215	+	subroutine buildElectroSpline()
216	+	real( kind = dp ), dimension(np) :: xvals, yvals
217	+	real( kind = dp ) :: dx, rmin, rval
218	+	integer :: i
219	+
220	+	rmin = 0.0_dp
221	+
222	+	dx = (defaultCutoff-rmin) / dble(np-1)
223	+
224	+	do i = 1, np
225	+	rval = rmin + dble(i-1)*dx
226	+	xvals(i) = rval
227	+	yvals(i) = erfc(dampingAlpha*rval)
228	+	enddo
229	+
230	+	call newSpline(erfcSpline, xvals, yvals, .true.)
231	+
232	+	haveElectroSpline = .true.
233	+	end subroutine buildElectroSpline
234	+
235		subroutine newElectrostaticType(c_ident, is_Charge, is_Dipole, &
236		is_SplitDipole, is_Quadrupole, is_Tap, status)
237
#	Line 123 | Line 259 | contains
259		return
260		end if
261
262	<	if (.not. allocated(ElectrostaticMap)) then
127	<	allocate(ElectrostaticMap(nAtypes))
128	<	endif
262	>	allocate(ElectrostaticMap(nAtypes))
263
264		end if
265
#	Line 143 | Line 277 | contains
277		ElectrostaticMap(myATID)%is_Quadrupole = is_Quadrupole
278		ElectrostaticMap(myATID)%is_Tap = is_Tap
279
280	+	hasElectrostaticMap = .true.
281	+
282		end subroutine newElectrostaticType
283
284		subroutine setCharge(c_ident, charge, status)
#	Line 154 | Line 290 | contains
290		status = 0
291		myATID = getFirstMatchingElement(atypes, "c_ident", c_ident)
292
293	<	if (.not.allocated(ElectrostaticMap)) then
293	>	if (.not.hasElectrostaticMap) then
294		call handleError("electrostatic", "no ElectrostaticMap was present before first call of setCharge!")
295		status = -1
296		return
#	Line 184 | Line 320 | contains
320		status = 0
321		myATID = getFirstMatchingElement(atypes, "c_ident", c_ident)
322
323	<	if (.not.allocated(ElectrostaticMap)) then
323	>	if (.not.hasElectrostaticMap) then
324		call handleError("electrostatic", "no ElectrostaticMap was present before first call of setDipoleMoment!")
325		status = -1
326		return
#	Line 214 | Line 350 | contains
350		status = 0
351		myATID = getFirstMatchingElement(atypes, "c_ident", c_ident)
352
353	<	if (.not.allocated(ElectrostaticMap)) then
353	>	if (.not.hasElectrostaticMap) then
354		call handleError("electrostatic", "no ElectrostaticMap was present before first call of setSplitDipoleDistance!")
355		status = -1
356		return
#	Line 244 | Line 380 | contains
380		status = 0
381		myATID = getFirstMatchingElement(atypes, "c_ident", c_ident)
382
383	<	if (.not.allocated(ElectrostaticMap)) then
383	>	if (.not.hasElectrostaticMap) then
384		call handleError("electrostatic", "no ElectrostaticMap was present before first call of setQuadrupoleMoments!")
385		status = -1
386		return
#	Line 275 | Line 411 | contains
411		integer :: localError
412		real(kind=dp) :: c
413
414	<	if (.not.allocated(ElectrostaticMap)) then
414	>	if (.not.hasElectrostaticMap) then
415		call handleError("electrostatic", "no ElectrostaticMap was present before first call of getCharge!")
416		return
417		end if
#	Line 293 | Line 429 | contains
429		integer :: localError
430		real(kind=dp) :: dm
431
432	<	if (.not.allocated(ElectrostaticMap)) then
432	>	if (.not.hasElectrostaticMap) then
433		call handleError("electrostatic", "no ElectrostaticMap was present before first call of getDipoleMoment!")
434		return
435		end if
#	Line 306 | Line 442 | contains
442		dm = ElectrostaticMap(atid)%dipole_moment
443		end function getDipoleMoment
444
445	<	subroutine doElectrostaticPair(atom1, atom2, d, rij, r2, sw, &
310	<	vpair, fpair, pot, eFrame, f, t, do_pot, corrMethod, rcuti)
445	>	subroutine checkSummationMethod()
446
447	+	if (.not.haveDefaultCutoff) then
448	+	call handleError("checkSummationMethod", "no Default Cutoff set!")
449	+	endif
450	+
451	+	rcuti = 1.0_dp / defaultCutoff
452	+	rcuti2 = rcuti*rcuti
453	+	rcuti3 = rcuti2*rcuti
454	+	rcuti4 = rcuti2*rcuti2
455	+
456	+	if (screeningMethod .eq. DAMPED) then
457	+	if (.not.haveDampingAlpha) then
458	+	call handleError("checkSummationMethod", "no Damping Alpha set!")
459	+	endif
460	+
461	+	if (.not.haveDefaultCutoff) then
462	+	call handleError("checkSummationMethod", "no Default Cutoff set!")
463	+	endif
464	+
465	+	constEXP = exp(-alpha2*defaultCutoff2)
466	+	invRootPi = 0.56418958354775628695_dp
467	+	alphaPi = 2.0_dp*dampingAlpha*invRootPi
468	+
469	+	c1c = erfc(dampingAlpha*defaultCutoff) * rcuti
470	+	c2c = alphaPi*constEXP*rcuti + c1c*rcuti
471	+	c3c = 2.0_dp*alphaPi*alpha2 + 3.0_dp*c2c*rcuti
472	+	c4c = 4.0_dp*alphaPi*alpha4 + 5.0_dp*c3c*rcuti2
473	+	c5c = 8.0_dp*alphaPi*alpha6 + 7.0_dp*c4c*rcuti2
474	+	c6c = 16.0_dp*alphaPi*alpha8 + 9.0_dp*c5c*rcuti2
475	+	else
476	+	c1c = rcuti
477	+	c2c = c1c*rcuti
478	+	c3c = 3.0_dp*c2c*rcuti
479	+	c4c = 5.0_dp*c3c*rcuti2
480	+	c5c = 7.0_dp*c4c*rcuti2
481	+	c6c = 9.0_dp*c5c*rcuti2
482	+	endif
483	+
484	+	if (summationMethod .eq. REACTION_FIELD) then
485	+	if (haveDielectric) then
486	+	defaultCutoff2 = defaultCutoff*defaultCutoff
487	+	preRF = (dielectric-1.0_dp) / &
488	+	((2.0_dp*dielectric+1.0_dp)*defaultCutoff2*defaultCutoff)
489	+	preRF2 = 2.0_dp*preRF
490	+	else
491	+	call handleError("checkSummationMethod", "Dielectric not set")
492	+	endif
493	+
494	+	endif
495	+
496	+	if (.not.haveElectroSpline) then
497	+	call buildElectroSpline()
498	+	end if
499	+
500	+	summationMethodChecked = .true.
501	+	end subroutine checkSummationMethod
502	+
503	+
504	+	subroutine doElectrostaticPair(atom1, atom2, d, rij, r2, rcut, sw, &
505	+	vpair, fpair, pot, eFrame, f, t, do_pot)
506	+
507		logical, intent(in) :: do_pot
508
509		integer, intent(in) :: atom1, atom2
510		integer :: localError
316	–	integer, intent(in) :: corrMethod
511
512	<	real(kind=dp), intent(in) :: rij, r2, sw, rcuti
512	>	real(kind=dp), intent(in) :: rij, r2, sw, rcut
513		real(kind=dp), intent(in), dimension(3) :: d
514		real(kind=dp), intent(inout) :: vpair
515	<	real(kind=dp), intent(inout), dimension(3) :: fpair
515	>	real(kind=dp), intent(inout), dimension(3) :: fpair
516
517	<	real( kind = dp ) :: pot, swi
517	>	real( kind = dp ) :: pot
518		real( kind = dp ), dimension(9,nLocal) :: eFrame
519		real( kind = dp ), dimension(3,nLocal) :: f
520	+	real( kind = dp ), dimension(3,nLocal) :: felec
521		real( kind = dp ), dimension(3,nLocal) :: t
522
523		real (kind = dp), dimension(3) :: ux_i, uy_i, uz_i
#	Line 340 | Line 535 | contains
535		real (kind=dp) :: cx_i, cy_i, cz_i
536		real (kind=dp) :: cx_j, cy_j, cz_j
537		real (kind=dp) :: cx2, cy2, cz2
538	<	real (kind=dp) :: ct_i, ct_j, ct_ij, a1
538	>	real (kind=dp) :: ct_i, ct_j, ct_ij, a0, a1
539		real (kind=dp) :: riji, ri, ri2, ri3, ri4
540		real (kind=dp) :: pref, vterm, epot, dudr, vterm1, vterm2
541		real (kind=dp) :: xhat, yhat, zhat
542		real (kind=dp) :: dudx, dudy, dudz
543	<	real (kind=dp) :: scale, sc2, bigR, switcher, dswitcher
544	<	real (kind=dp) :: rcuti2, rcuti3, rcuti4
543	>	real (kind=dp) :: scale, sc2, bigR
544	>	real (kind=dp) :: varEXP
545	>	real (kind=dp) :: pot_term
546	>	real (kind=dp) :: preVal, rfVal
547	>	real (kind=dp) :: c2ri, c3ri, c4rij
548	>	real (kind=dp) :: cti3, ctj3, ctidotj
549	>	real (kind=dp) :: preSw, preSwSc
550	>	real (kind=dp) :: xhatdot2, yhatdot2, zhatdot2
551	>	real (kind=dp) :: xhatc4, yhatc4, zhatc4
552
553	<	if (.not.allocated(ElectrostaticMap)) then
554	<	call handleError("electrostatic", "no ElectrostaticMap was present before first call of do_electrostatic_pair!")
555	<	return
354	<	end if
553	>	if (.not.summationMethodChecked) then
554	>	call checkSummationMethod()
555	>	endif
556
557		#ifdef IS_MPI
558		me1 = atid_Row(atom1)
#	Line 363 | Line 564 | contains
564
565		!! some variables we'll need independent of electrostatic type:
566
567	<	riji = 1.0d0 / rij
568	<
567	>	riji = 1.0_dp / rij
568	>
569		xhat = d(1) * riji
570		yhat = d(2) * riji
571		zhat = d(3) * riji
572
372	–	rcuti2 = rcuti*rcuti
373	–	rcuti3 = rcuti2*rcuti
374	–	rcuti4 = rcuti2*rcuti2
375	–
376	–	swi = 1.0d0 / sw
377	–
573		!! logicals
574		i_is_Charge = ElectrostaticMap(me1)%is_Charge
575		i_is_Dipole = ElectrostaticMap(me1)%is_Dipole
#	Line 408 | Line 603 | contains
603		if (i_is_SplitDipole) then
604		d_i = ElectrostaticMap(me1)%split_dipole_distance
605		endif
606	<
606	>	duduz_i = zero
607		endif
608
609		if (i_is_Quadrupole) then
#	Line 439 | Line 634 | contains
634		cx_i = ux_i(1)*xhat + ux_i(2)*yhat + ux_i(3)*zhat
635		cy_i = uy_i(1)*xhat + uy_i(2)*yhat + uy_i(3)*zhat
636		cz_i = uz_i(1)*xhat + uz_i(2)*yhat + uz_i(3)*zhat
637	+	dudux_i = zero
638	+	duduy_i = zero
639	+	duduz_i = zero
640		endif
641
642		if (j_is_Charge) then
#	Line 461 | Line 659 | contains
659		if (j_is_SplitDipole) then
660		d_j = ElectrostaticMap(me2)%split_dipole_distance
661		endif
662	+	duduz_j = zero
663		endif
664
665		if (j_is_Quadrupole) then
#	Line 491 | Line 690 | contains
690		cx_j = ux_j(1)*xhat + ux_j(2)*yhat + ux_j(3)*zhat
691		cy_j = uy_j(1)*xhat + uy_j(2)*yhat + uy_j(3)*zhat
692		cz_j = uz_j(1)*xhat + uz_j(2)*yhat + uz_j(3)*zhat
693	+	dudux_j = zero
694	+	duduy_j = zero
695	+	duduz_j = zero
696		endif
697
698	<	!!$    switcher = 1.0d0
699	<	!!$    dswitcher = 0.0d0
700	<	!!$    ebalance = 0.0d0
701	<	!!$    ! weaken the dipole interaction at close range for TAP water
500	<	!!$    if (j_is_Tap .and. i_is_Tap) then
501	<	!!$      call calc_switch(rij, mu_i, switcher, dswitcher)
502	<	!!$    endif
698	>	epot = zero
699	>	dudx = zero
700	>	dudy = zero
701	>	dudz = zero
702
504	–	epot = 0.0_dp
505	–	dudx = 0.0_dp
506	–	dudy = 0.0_dp
507	–	dudz = 0.0_dp
508	–
509	–	dudux_i = 0.0_dp
510	–	duduy_i = 0.0_dp
511	–	duduz_i = 0.0_dp
512	–
513	–	dudux_j = 0.0_dp
514	–	duduy_j = 0.0_dp
515	–	duduz_j = 0.0_dp
516	–
703		if (i_is_Charge) then
704
705		if (j_is_Charge) then
706	+	if (screeningMethod .eq. DAMPED) then
707	+	! assemble the damping variables
708	+	call lookupUniformSpline1d(erfcSpline, rij, erfcVal, derfcVal)
709	+	c1 = erfcVal*riji
710	+	c2 = (-derfcVal + c1)*riji
711	+	else
712	+	c1 = riji
713	+	c2 = c1*riji
714	+	endif
715
716	<	if (corrMethod .eq. 1) then
522	<	vterm = pre11 * q_i * q_j * (riji - rcuti)
716	>	preVal = pre11 * q_i * q_j
717
718	<	vpair = vpair + vterm
719	<	epot = epot + sw * vterm
526	<
527	<	dudr  = - sw * pre11 * q_i * q_j * (riji*riji*riji - rcuti2*rcuti)
718	>	if (summationMethod .eq. SHIFTED_POTENTIAL) then
719	>	vterm = preVal * (c1 - c1c)
720
721	<	dudx = dudx + dudr * d(1)
722	<	dudy = dudy + dudr * d(2)
723	<	dudz = dudz + dudr * d(3)
724	<
533	<	else
534	<	vterm = pre11 * q_i * q_j * riji
535	<
536	<	vpair = vpair + vterm
537	<	epot = epot + sw * vterm
721	>	dudr  = -sw * preVal * c2
722	>
723	>	elseif (summationMethod .eq. SHIFTED_FORCE) then
724	>	vterm = preVal * ( c1 - c1c + c2c*(rij - defaultCutoff) )
725
726	<	dudr  = - sw * vterm * riji
726	>	dudr  = sw * preVal * (c2c - c2)
727	>
728	>	elseif (summationMethod .eq. REACTION_FIELD) then
729	>	rfVal = preRF*rij*rij
730	>	vterm = preVal * ( riji + rfVal )
731
732	<	dudx = dudx + dudr * xhat
733	<	dudy = dudy + dudr * yhat
734	<	dudz = dudz + dudr * zhat
735	<
732	>	dudr  = sw * preVal * ( 2.0_dp*rfVal - riji )*riji
733	>
734	>	else
735	>	vterm = preVal * riji*erfcVal
736	>
737	>	dudr  = - sw * preVal * c2
738	>
739		endif
740
741	+	vpair = vpair + vterm
742	+	epot = epot + sw*vterm
743	+
744	+	dudx = dudx + dudr * xhat
745	+	dudy = dudy + dudr * yhat
746	+	dudz = dudz + dudr * zhat
747	+
748		endif
749
750		if (j_is_Dipole) then
751	+	! pref is used by all the possible methods
752	+	pref = pre12 * q_i * mu_j
753	+	preSw = sw*pref
754
755	<	pref = sw * pre12 * q_i * mu_j
552	<
553	<	if (corrMethod .eq. 1) then
755	>	if (summationMethod .eq. REACTION_FIELD) then
756		ri2 = riji * riji
757		ri3 = ri2 * riji
758	<
759	<	vterm = - pref * ct_j * (ri2 - rcuti2)
760	<	vpair = vpair + swi*vterm
761	<	epot = epot + vterm
758	>
759	>	vterm = - pref * ct_j * ( ri2 - preRF2*rij )
760	>	vpair = vpair + vterm
761	>	epot = epot + sw*vterm
762
763	<	!! this has a + sign in the () because the rij vector is
764	<	!! r_j - r_i and the charge-dipole potential takes the origin
765	<	!! as the point dipole, which is atom j in this case.
766	<
767	<	dudx = dudx - pref * ( ri3*( uz_j(1) - 3.0d0*ct_j*xhat) &
768	<	- rcuti3*( uz_j(1) - 3.0d0*ct_j*d(1)*rcuti ) )
769	<	dudy = dudy - pref * ( ri3*( uz_j(2) - 3.0d0*ct_j*yhat) &
770	<	- rcuti3*( uz_j(2) - 3.0d0*ct_j*d(2)*rcuti ) )
771	<	dudz = dudz - pref * ( ri3*( uz_j(3) - 3.0d0*ct_j*zhat) &
570	<	- rcuti3*( uz_j(3) - 3.0d0*ct_j*d(3)*rcuti ) )
571	<
572	<	duduz_j(1) = duduz_j(1) - pref*( ri2*xhat - d(1)*rcuti3 )
573	<	duduz_j(2) = duduz_j(2) - pref*( ri2*yhat - d(2)*rcuti3 )
574	<	duduz_j(3) = duduz_j(3) - pref*( ri2*zhat - d(3)*rcuti3 )
763	>	dudx = dudx - preSw*( ri3*(uz_j(1) - 3.0_dp*ct_j*xhat) - &
764	>	preRF2*uz_j(1) )
765	>	dudy = dudy - preSw*( ri3*(uz_j(2) - 3.0_dp*ct_j*yhat) - &
766	>	preRF2*uz_j(2) )
767	>	dudz = dudz - preSw*( ri3*(uz_j(3) - 3.0_dp*ct_j*zhat) - &
768	>	preRF2*uz_j(3) )
769	>	duduz_j(1) = duduz_j(1) - preSw * xhat * ( ri2 - preRF2*rij )
770	>	duduz_j(2) = duduz_j(2) - preSw * yhat * ( ri2 - preRF2*rij )
771	>	duduz_j(3) = duduz_j(3) - preSw * zhat * ( ri2 - preRF2*rij )
772
773		else
774	+	! determine the inverse r used if we have split dipoles
775		if (j_is_SplitDipole) then
776		BigR = sqrt(r2 + 0.25_dp * d_j * d_j)
777		ri = 1.0_dp / BigR
#	Line 582 | Line 780 | contains
780		ri = riji
781		scale = 1.0_dp
782		endif
783	<
586	<	ri2 = ri * ri
587	<	ri3 = ri2 * ri
783	>
784		sc2 = scale * scale
785	+
786	+	if (screeningMethod .eq. DAMPED) then
787	+	! assemble the damping variables
788	+	call lookupUniformSpline1d(erfcSpline, rij, erfcVal, derfcVal)
789	+	c1 = erfcVal*ri
790	+	c2 = (-derfcVal + c1)*ri
791	+	c3 = -2.0_dp*derfcVal*alpha2 + 3.0_dp*c2*ri
792	+	else
793	+	c1 = ri
794	+	c2 = c1*ri
795	+	c3 = 3.0_dp*c2*ri
796	+	endif
797
798	<	vterm = - pref * ct_j * ri2 * scale
799	<	vpair = vpair + swi * vterm
800	<	epot = epot + vterm
798	>	c2ri = c2*ri
799	>
800	>	! calculate the potential
801	>	pot_term =  scale * c2
802	>	vterm = -pref * ct_j * pot_term
803	>	vpair = vpair + vterm
804	>	epot = epot + sw*vterm
805
806	<	!! this has a + sign in the () because the rij vector is
807	<	!! r_j - r_i and the charge-dipole potential takes the origin
808	<	!! as the point dipole, which is atom j in this case.
809	<
810	<	dudx = dudx - pref * ri3 * ( uz_j(1) - 3.0d0*ct_j*xhat*sc2)
811	<	dudy = dudy - pref * ri3 * ( uz_j(2) - 3.0d0*ct_j*yhat*sc2)
812	<	dudz = dudz - pref * ri3 * ( uz_j(3) - 3.0d0*ct_j*zhat*sc2)
813	<
602	<	duduz_j(1) = duduz_j(1) - pref * ri2 * xhat * scale
603	<	duduz_j(2) = duduz_j(2) - pref * ri2 * yhat * scale
604	<	duduz_j(3) = duduz_j(3) - pref * ri2 * zhat * scale
806	>	! calculate derivatives for forces and torques
807	>	dudx = dudx - preSw*( uz_j(1)*c2ri - ct_j*xhat*sc2*c3 )
808	>	dudy = dudy - preSw*( uz_j(2)*c2ri - ct_j*yhat*sc2*c3 )
809	>	dudz = dudz - preSw*( uz_j(3)*c2ri - ct_j*zhat*sc2*c3 )
810	>
811	>	duduz_j(1) = duduz_j(1) - preSw * pot_term * xhat
812	>	duduz_j(2) = duduz_j(2) - preSw * pot_term * yhat
813	>	duduz_j(3) = duduz_j(3) - preSw * pot_term * zhat
814
815		endif
816		endif
817
818		if (j_is_Quadrupole) then
819	<	ri2 = riji * riji
611	<	ri3 = ri2 * riji
612	<	ri4 = ri2 * ri2
819	>	! first precalculate some necessary variables
820		cx2 = cx_j * cx_j
821		cy2 = cy_j * cy_j
822		cz2 = cz_j * cz_j
823	+	pref =  pre14 * q_i * one_third
824	+
825	+	if (screeningMethod .eq. DAMPED) then
826	+	! assemble the damping variables
827	+	call lookupUniformSpline1d(erfcSpline, rij, erfcVal, derfcVal)
828	+	c1 = erfcVal*riji
829	+	c2 = (-derfcVal + c1)*riji
830	+	c3 = -2.0_dp*derfcVal*alpha2 + 3.0_dp*c2*riji
831	+	c4 = -4.0_dp*derfcVal*alpha4 + 5.0_dp*c3*riji*riji
832	+	else
833	+	c1 = riji
834	+	c2 = c1*riji
835	+	c3 = 3.0_dp*c2*riji
836	+	c4 = 5.0_dp*c3*riji*riji
837	+	endif
838
839	+	! precompute variables for convenience
840	+	preSw = sw*pref
841	+	c2ri = c2*riji
842	+	c3ri = c3*riji
843	+	c4rij = c4*rij
844	+	xhatdot2 = 2.0_dp*xhat*c3
845	+	yhatdot2 = 2.0_dp*yhat*c3
846	+	zhatdot2 = 2.0_dp*zhat*c3
847	+	xhatc4 = xhat*c4rij
848	+	yhatc4 = yhat*c4rij
849	+	zhatc4 = zhat*c4rij
850
851	<	pref =  sw * pre14 * q_i / 3.0_dp
851	>	! calculate the potential
852	>	pot_term = ( qxx_j*(cx2*c3 - c2ri) + qyy_j*(cy2*c3 - c2ri) + &
853	>	qzz_j*(cz2*c3 - c2ri) )
854	>	vterm = pref * pot_term
855	>	vpair = vpair + vterm
856	>	epot = epot + sw*vterm
857
858	<	if (corrMethod .eq. 1) then
859	<	vterm1 = pref * ri3*( qxx_j * (3.0_dp*cx2 - 1.0_dp) + &
860	<	qyy_j * (3.0_dp*cy2 - 1.0_dp) + &
861	<	qzz_j * (3.0_dp*cz2 - 1.0_dp) )
862	<	vterm2 = pref * rcuti3*( qxx_j * (3.0_dp*cx2 - 1.0_dp) + &
863	<	qyy_j * (3.0_dp*cy2 - 1.0_dp) + &
864	<	qzz_j * (3.0_dp*cz2 - 1.0_dp) )
865	<	vpair = vpair + swi*( vterm1 - vterm2 )
866	<	epot = epot + ( vterm1 - vterm2 )
867	<
868	<	dudx = dudx - (5.0_dp * &
869	<	(vterm1*riji*xhat - vterm2*rcuti2*d(1))) + pref * ( &
870	<	(ri4 - rcuti4)*(qxx_j*(6.0_dp*cx_j*ux_j(1)) - &
633	<	qxx_j*2.0_dp*(xhat - rcuti*d(1))) + &
634	<	(ri4 - rcuti4)*(qyy_j*(6.0_dp*cy_j*uy_j(1)) - &
635	<	qyy_j*2.0_dp*(xhat - rcuti*d(1))) + &
636	<	(ri4 - rcuti4)*(qzz_j*(6.0_dp*cz_j*uz_j(1)) - &
637	<	qzz_j*2.0_dp*(xhat - rcuti*d(1))) )
638	<	dudy = dudy - (5.0_dp * &
639	<	(vterm1*riji*yhat - vterm2*rcuti2*d(2))) + pref * ( &
640	<	(ri4 - rcuti4)*(qxx_j*(6.0_dp*cx_j*ux_j(2)) - &
641	<	qxx_j*2.0_dp*(yhat - rcuti*d(2))) + &
642	<	(ri4 - rcuti4)*(qyy_j*(6.0_dp*cy_j*uy_j(2)) - &
643	<	qyy_j*2.0_dp*(yhat - rcuti*d(2))) + &
644	<	(ri4 - rcuti4)*(qzz_j*(6.0_dp*cz_j*uz_j(2)) - &
645	<	qzz_j*2.0_dp*(yhat - rcuti*d(2))) )
646	<	dudz = dudz - (5.0_dp * &
647	<	(vterm1*riji*zhat - vterm2*rcuti2*d(3))) + pref * ( &
648	<	(ri4 - rcuti4)*(qxx_j*(6.0_dp*cx_j*ux_j(3)) - &
649	<	qxx_j*2.0_dp*(zhat - rcuti*d(3))) + &
650	<	(ri4 - rcuti4)*(qyy_j*(6.0_dp*cy_j*uy_j(3)) - &
651	<	qyy_j*2.0_dp*(zhat - rcuti*d(3))) + &
652	<	(ri4 - rcuti4)*(qzz_j*(6.0_dp*cz_j*uz_j(3)) - &
653	<	qzz_j*2.0_dp*(zhat - rcuti*d(3))) )
654	<
655	<	dudux_j(1) = dudux_j(1) + pref * (ri3*(qxx_j*6.0_dp*cx_j*xhat) - &
656	<	rcuti4*(qxx_j*6.0_dp*cx_j*d(1)))
657	<	dudux_j(2) = dudux_j(2) + pref * (ri3*(qxx_j*6.0_dp*cx_j*yhat) - &
658	<	rcuti4*(qxx_j*6.0_dp*cx_j*d(2)))
659	<	dudux_j(3) = dudux_j(3) + pref * (ri3*(qxx_j*6.0_dp*cx_j*zhat) - &
660	<	rcuti4*(qxx_j*6.0_dp*cx_j*d(3)))
661	<
662	<	duduy_j(1) = duduy_j(1) + pref * (ri3*(qyy_j*6.0_dp*cy_j*xhat) - &
663	<	rcuti4*(qyy_j*6.0_dp*cx_j*d(1)))
664	<	duduy_j(2) = duduy_j(2) + pref * (ri3*(qyy_j*6.0_dp*cy_j*yhat) - &
665	<	rcuti4*(qyy_j*6.0_dp*cx_j*d(2)))
666	<	duduy_j(3) = duduy_j(3) + pref * (ri3*(qyy_j*6.0_dp*cy_j*zhat) - &
667	<	rcuti4*(qyy_j*6.0_dp*cx_j*d(3)))
668	<
669	<	duduz_j(1) = duduz_j(1) + pref * (ri3*(qzz_j*6.0_dp*cz_j*xhat) - &
670	<	rcuti4*(qzz_j*6.0_dp*cx_j*d(1)))
671	<	duduz_j(2) = duduz_j(2) + pref * (ri3*(qzz_j*6.0_dp*cz_j*yhat) - &
672	<	rcuti4*(qzz_j*6.0_dp*cx_j*d(2)))
673	<	duduz_j(3) = duduz_j(3) + pref * (ri3*(qzz_j*6.0_dp*cz_j*zhat) - &
674	<	rcuti4*(qzz_j*6.0_dp*cx_j*d(3)))
675	<
676	<	else
677	<	vterm = pref * ri3 * (qxx_j * (3.0_dp*cx2 - 1.0_dp) + &
678	<	qyy_j * (3.0_dp*cy2 - 1.0_dp) + &
679	<	qzz_j * (3.0_dp*cz2 - 1.0_dp))
680	<	vpair = vpair + swi * vterm
681	<	epot = epot + vterm
682	<
683	<	dudx = dudx - 5.0_dp*vterm*riji*xhat + pref * ri4 * ( &
684	<	qxx_j*(6.0_dp*cx_j*ux_j(1) - 2.0_dp*xhat) + &
685	<	qyy_j*(6.0_dp*cy_j*uy_j(1) - 2.0_dp*xhat) + &
686	<	qzz_j*(6.0_dp*cz_j*uz_j(1) - 2.0_dp*xhat) )
687	<	dudy = dudy - 5.0_dp*vterm*riji*yhat + pref * ri4 * ( &
688	<	qxx_j*(6.0_dp*cx_j*ux_j(2) - 2.0_dp*yhat) + &
689	<	qyy_j*(6.0_dp*cy_j*uy_j(2) - 2.0_dp*yhat) + &
690	<	qzz_j*(6.0_dp*cz_j*uz_j(2) - 2.0_dp*yhat) )
691	<	dudz = dudz - 5.0_dp*vterm*riji*zhat + pref * ri4 * ( &
692	<	qxx_j*(6.0_dp*cx_j*ux_j(3) - 2.0_dp*zhat) + &
693	<	qyy_j*(6.0_dp*cy_j*uy_j(3) - 2.0_dp*zhat) + &
694	<	qzz_j*(6.0_dp*cz_j*uz_j(3) - 2.0_dp*zhat) )
695	<
696	<	dudux_j(1) = dudux_j(1) + pref * ri3*(qxx_j*6.0_dp*cx_j*xhat)
697	<	dudux_j(2) = dudux_j(2) + pref * ri3*(qxx_j*6.0_dp*cx_j*yhat)
698	<	dudux_j(3) = dudux_j(3) + pref * ri3*(qxx_j*6.0_dp*cx_j*zhat)
699	<
700	<	duduy_j(1) = duduy_j(1) + pref * ri3*(qyy_j*6.0_dp*cy_j*xhat)
701	<	duduy_j(2) = duduy_j(2) + pref * ri3*(qyy_j*6.0_dp*cy_j*yhat)
702	<	duduy_j(3) = duduy_j(3) + pref * ri3*(qyy_j*6.0_dp*cy_j*zhat)
703	<
704	<	duduz_j(1) = duduz_j(1) + pref * ri3*(qzz_j*6.0_dp*cz_j*xhat)
705	<	duduz_j(2) = duduz_j(2) + pref * ri3*(qzz_j*6.0_dp*cz_j*yhat)
706	<	duduz_j(3) = duduz_j(3) + pref * ri3*(qzz_j*6.0_dp*cz_j*zhat)
858	>	! calculate derivatives for the forces and torques
859	>	dudx = dudx - preSw * ( &
860	>	qxx_j*(cx2*xhatc4 - (2.0_dp*cx_j*ux_j(1) + xhat)*c3ri) + &
861	>	qyy_j*(cy2*xhatc4 - (2.0_dp*cy_j*uy_j(1) + xhat)*c3ri) + &
862	>	qzz_j*(cz2*xhatc4 - (2.0_dp*cz_j*uz_j(1) + xhat)*c3ri) )
863	>	dudy = dudy - preSw * ( &
864	>	qxx_j*(cx2*yhatc4 - (2.0_dp*cx_j*ux_j(2) + yhat)*c3ri) + &
865	>	qyy_j*(cy2*yhatc4 - (2.0_dp*cy_j*uy_j(2) + yhat)*c3ri) + &
866	>	qzz_j*(cz2*yhatc4 - (2.0_dp*cz_j*uz_j(2) + yhat)*c3ri) )
867	>	dudz = dudz - preSw * ( &
868	>	qxx_j*(cx2*zhatc4 - (2.0_dp*cx_j*ux_j(3) + zhat)*c3ri) + &
869	>	qyy_j*(cy2*zhatc4 - (2.0_dp*cy_j*uy_j(3) + zhat)*c3ri) + &
870	>	qzz_j*(cz2*zhatc4 - (2.0_dp*cz_j*uz_j(3) + zhat)*c3ri) )
871
872	<	endif
872	>	dudux_j(1) = dudux_j(1) + preSw*(qxx_j*cx_j*xhatdot2)
873	>	dudux_j(2) = dudux_j(2) + preSw*(qxx_j*cx_j*yhatdot2)
874	>	dudux_j(3) = dudux_j(3) + preSw*(qxx_j*cx_j*zhatdot2)
875	>
876	>	duduy_j(1) = duduy_j(1) + preSw*(qyy_j*cy_j*xhatdot2)
877	>	duduy_j(2) = duduy_j(2) + preSw*(qyy_j*cy_j*yhatdot2)
878	>	duduy_j(3) = duduy_j(3) + preSw*(qyy_j*cy_j*zhatdot2)
879	>
880	>	duduz_j(1) = duduz_j(1) + preSw*(qzz_j*cz_j*xhatdot2)
881	>	duduz_j(2) = duduz_j(2) + preSw*(qzz_j*cz_j*yhatdot2)
882	>	duduz_j(3) = duduz_j(3) + preSw*(qzz_j*cz_j*zhatdot2)
883	>
884	>
885		endif
886		endif
887	<
887	>
888		if (i_is_Dipole) then
889
890		if (j_is_Charge) then
891	+	! variables used by all the methods
892	+	pref = pre12 * q_j * mu_i
893	+	preSw = sw*pref
894
895	<	pref = sw * pre12 * q_j * mu_i
895	>	if (summationMethod .eq. REACTION_FIELD) then
896
718	–	if (corrMethod .eq. 1) then
897		ri2 = riji * riji
898		ri3 = ri2 * riji
899
900	<	vterm = pref * ct_i * (ri2 - rcuti2)
901	<	vpair = vpair + swi * vterm
902	<	epot = epot + vterm
900	>	vterm = pref * ct_i * ( ri2 - preRF2*rij )
901	>	vpair = vpair + vterm
902	>	epot = epot + sw*vterm
903
904	<	!! this has a + sign in the () because the rij vector is
905	<	!! r_j - r_i and the charge-dipole potential takes the origin
906	<	!! as the point dipole, which is atom j in this case.
904	>	dudx = dudx + preSw * ( ri3*(uz_i(1) - 3.0_dp*ct_i*xhat) - &
905	>	preRF2*uz_i(1) )
906	>	dudy = dudy + preSw * ( ri3*(uz_i(2) - 3.0_dp*ct_i*yhat) - &
907	>	preRF2*uz_i(2) )
908	>	dudz = dudz + preSw * ( ri3*(uz_i(3) - 3.0_dp*ct_i*zhat) - &
909	>	preRF2*uz_i(3) )
910
911	<	dudx = dudx + pref * ( ri3*( uz_i(1) - 3.0d0*ct_i*xhat) &
912	<	- rcuti3*( uz_i(1) - 3.0d0*ct_i*d(1)*rcuti ) )
913	<	dudy = dudy + pref * ( ri3*( uz_i(2) - 3.0d0*ct_i*yhat) &
733	<	- rcuti3*( uz_i(2) - 3.0d0*ct_i*d(2)*rcuti ) )
734	<	dudz = dudz + pref * ( ri3*( uz_i(3) - 3.0d0*ct_i*zhat) &
735	<	- rcuti3*( uz_i(3) - 3.0d0*ct_i*d(3)*rcuti ) )
736	<
737	<	duduz_i(1) = duduz_i(1) - pref*( ri2*xhat - d(1)*rcuti3 )
738	<	duduz_i(2) = duduz_i(2) - pref*( ri2*yhat - d(2)*rcuti3 )
739	<	duduz_i(3) = duduz_i(3) - pref*( ri2*zhat - d(3)*rcuti3 )
911	>	duduz_i(1) = duduz_i(1) + preSw * xhat * ( ri2 - preRF2*rij )
912	>	duduz_i(2) = duduz_i(2) + preSw * yhat * ( ri2 - preRF2*rij )
913	>	duduz_i(3) = duduz_i(3) + preSw * zhat * ( ri2 - preRF2*rij )
914
915		else
916	+	! determine inverse r if we are using split dipoles
917		if (i_is_SplitDipole) then
918		BigR = sqrt(r2 + 0.25_dp * d_i * d_i)
919		ri = 1.0_dp / BigR
#	Line 747 | Line 922 | contains
922		ri = riji
923		scale = 1.0_dp
924		endif
925	<
751	<	ri2 = ri * ri
752	<	ri3 = ri2 * ri
925	>
926		sc2 = scale * scale
927	+
928	+	if (screeningMethod .eq. DAMPED) then
929	+	! assemble the damping variables
930	+	call lookupUniformSpline1d(erfcSpline, rij, erfcVal, derfcVal)
931	+	c1 = erfcVal*ri
932	+	c2 = (-derfcVal + c1)*ri
933	+	c3 = -2.0_dp*derfcVal*alpha2 + 3.0_dp*c2*ri
934	+	else
935	+	c1 = ri
936	+	c2 = c1*ri
937	+	c3 = 3.0_dp*c2*ri
938	+	endif
939	+
940	+	c2ri = c2*ri
941	+
942	+	! calculate the potential
943	+	pot_term = c2 * scale
944	+	vterm = pref * ct_i * pot_term
945	+	vpair = vpair + vterm
946	+	epot = epot + sw*vterm
947	+
948	+	! calculate derivatives for the forces and torques
949	+	dudx = dudx + preSw * ( uz_i(1)*c2ri - ct_i*xhat*sc2*c3 )
950	+	dudy = dudy + preSw * ( uz_i(2)*c2ri - ct_i*yhat*sc2*c3 )
951	+	dudz = dudz + preSw * ( uz_i(3)*c2ri - ct_i*zhat*sc2*c3 )
952	+
953	+	duduz_i(1) = duduz_i(1) + preSw * pot_term * xhat
954	+	duduz_i(2) = duduz_i(2) + preSw * pot_term * yhat
955	+	duduz_i(3) = duduz_i(3) + preSw * pot_term * zhat
956
755	–	vterm = pref * ct_i * ri2 * scale
756	–	vpair = vpair + swi * vterm
757	–	epot = epot + vterm
758	–
759	–	dudx = dudx + pref * ri3 * ( uz_i(1) - 3.0d0 * ct_i * xhat*sc2)
760	–	dudy = dudy + pref * ri3 * ( uz_i(2) - 3.0d0 * ct_i * yhat*sc2)
761	–	dudz = dudz + pref * ri3 * ( uz_i(3) - 3.0d0 * ct_i * zhat*sc2)
762	–
763	–	duduz_i(1) = duduz_i(1) + pref * ri2 * xhat * scale
764	–	duduz_i(2) = duduz_i(2) + pref * ri2 * yhat * scale
765	–	duduz_i(3) = duduz_i(3) + pref * ri2 * zhat * scale
957		endif
958		endif
959	<
959	>
960		if (j_is_Dipole) then
961	+	! variables used by all methods
962	+	ct_ij = uz_i(1)*uz_j(1) + uz_i(2)*uz_j(2) + uz_i(3)*uz_j(3)
963	+	pref = pre22 * mu_i * mu_j
964	+	preSw = sw*pref
965
966	<	pref = sw * pre22 * mu_i * mu_j
772	<
773	<	if (corrMethod .eq. 1) then
966	>	if (summationMethod .eq. REACTION_FIELD) then
967		ri2 = riji * riji
968		ri3 = ri2 * riji
969		ri4 = ri2 * ri2
970
971	<	vterm = pref * (ri3 - rcuti3) * (ct_ij - 3.0d0 * ct_i * ct_j)
972	<	vpair = vpair + swi * vterm
973	<	epot = epot + vterm
971	>	vterm = pref*( ri3*(ct_ij - 3.0_dp * ct_i * ct_j) - &
972	>	preRF2*ct_ij )
973	>	vpair = vpair + vterm
974	>	epot = epot + sw*vterm
975
976	<	a1 = 5.0d0 * ct_i * ct_j - ct_ij
976	>	a1 = 5.0_dp * ct_i * ct_j - ct_ij
977
978	<	dudx = dudx + pref*3.0d0*ri4 &
979	<	*(a1*xhat-ct_i*uz_j(1)-ct_j*uz_i(1)) - &
980	<	pref*3.0d0*rcuti4*(a1*rcuti*d(1)-ct_i*uz_j(1)-ct_j*uz_i(1))
787	<	dudy = dudy + pref*3.0d0*ri4 &
788	<	*(a1*yhat-ct_i*uz_j(2)-ct_j*uz_i(2)) - &
789	<	pref*3.0d0*rcuti4*(a1*rcuti*d(2)-ct_i*uz_j(2)-ct_j*uz_i(2))
790	<	dudz = dudz + pref*3.0d0*ri4 &
791	<	*(a1*zhat-ct_i*uz_j(3)-ct_j*uz_i(3)) - &
792	<	pref*3.0d0*rcuti4*(a1*rcuti*d(3)-ct_i*uz_j(3)-ct_j*uz_i(3))
978	>	dudx = dudx + preSw*3.0_dp*ri4*(a1*xhat-ct_i*uz_j(1)-ct_j*uz_i(1))
979	>	dudy = dudy + preSw*3.0_dp*ri4*(a1*yhat-ct_i*uz_j(2)-ct_j*uz_i(2))
980	>	dudz = dudz + preSw*3.0_dp*ri4*(a1*zhat-ct_i*uz_j(3)-ct_j*uz_i(3))
981
982	<	duduz_i(1) = duduz_i(1) + pref*(ri3*(uz_j(1) - 3.0d0*ct_j*xhat) &
983	<	- rcuti3*(uz_j(1) - 3.0d0*ct_j*d(1)*rcuti))
984	<	duduz_i(2) = duduz_i(2) + pref*(ri3*(uz_j(2) - 3.0d0*ct_j*yhat) &
985	<	- rcuti3*(uz_j(2) - 3.0d0*ct_j*d(2)*rcuti))
986	<	duduz_i(3) = duduz_i(3) + pref*(ri3*(uz_j(3) - 3.0d0*ct_j*zhat) &
987	<	- rcuti3*(uz_j(3) - 3.0d0*ct_j*d(3)*rcuti))
988	<	duduz_j(1) = duduz_j(1) + pref*(ri3*(uz_i(1) - 3.0d0*ct_i*xhat) &
989	<	- rcuti3*(uz_i(1) - 3.0d0*ct_i*d(1)*rcuti))
990	<	duduz_j(2) = duduz_j(2) + pref*(ri3*(uz_i(2) - 3.0d0*ct_i*yhat) &
991	<	- rcuti3*(uz_i(2) - 3.0d0*ct_i*d(2)*rcuti))
992	<	duduz_j(3) = duduz_j(3) + pref*(ri3*(uz_i(3) - 3.0d0*ct_i*zhat) &
993	<	- rcuti3*(uz_i(3) - 3.0d0*ct_i*d(3)*rcuti))
982	>	duduz_i(1) = duduz_i(1) + preSw*(ri3*(uz_j(1)-3.0_dp*ct_j*xhat) &
983	>	- preRF2*uz_j(1))
984	>	duduz_i(2) = duduz_i(2) + preSw*(ri3*(uz_j(2)-3.0_dp*ct_j*yhat) &
985	>	- preRF2*uz_j(2))
986	>	duduz_i(3) = duduz_i(3) + preSw*(ri3*(uz_j(3)-3.0_dp*ct_j*zhat) &
987	>	- preRF2*uz_j(3))
988	>	duduz_j(1) = duduz_j(1) + preSw*(ri3*(uz_i(1)-3.0_dp*ct_i*xhat) &
989	>	- preRF2*uz_i(1))
990	>	duduz_j(2) = duduz_j(2) + preSw*(ri3*(uz_i(2)-3.0_dp*ct_i*yhat) &
991	>	- preRF2*uz_i(2))
992	>	duduz_j(3) = duduz_j(3) + preSw*(ri3*(uz_i(3)-3.0_dp*ct_i*zhat) &
993	>	- preRF2*uz_i(3))
994	>
995		else
807	–
996		if (i_is_SplitDipole) then
997		if (j_is_SplitDipole) then
998		BigR = sqrt(r2 + 0.25_dp * d_i * d_i + 0.25_dp * d_j * d_j)
#	Line 823 | Line 1011 | contains
1011		scale = 1.0_dp
1012		endif
1013		endif
1014	<
1015	<	ct_ij = uz_i(1)*uz_j(1) + uz_i(2)*uz_j(2) + uz_i(3)*uz_j(3)
1016	<
1017	<	ri2 = ri * ri
1018	<	ri3 = ri2 * ri
1019	<	ri4 = ri2 * ri2
1014	>
1015	>	if (screeningMethod .eq. DAMPED) then
1016	>	! assemble the damping variables
1017	>	call lookupUniformSpline1d(erfcSpline, rij, erfcVal, derfcVal)
1018	>	c1 = erfcVal*ri
1019	>	c2 = (-derfcVal + c1)*ri
1020	>	c3 = -2.0_dp*derfcVal*alpha2 + 3.0_dp*c2*ri
1021	>	c4 = -4.0_dp*derfcVal*alpha4 + 5.0_dp*c3*ri*ri
1022	>	else
1023	>	c1 = ri
1024	>	c2 = c1*ri
1025	>	c3 = 3.0_dp*c2*ri
1026	>	c4 = 5.0_dp*c3*ri*ri
1027	>	endif
1028	>
1029	>	! precompute variables for convenience
1030		sc2 = scale * scale
1031	+	cti3 = ct_i*sc2*c3
1032	+	ctj3 = ct_j*sc2*c3
1033	+	ctidotj = ct_i * ct_j * sc2
1034	+	preSwSc = preSw*scale
1035	+	c2ri = c2*ri
1036	+	c3ri = c3*ri
1037	+	c4rij = c4*rij
1038	+
1039	+
1040	+	! calculate the potential
1041	+	pot_term = (ct_ij*c2ri - ctidotj*c3)
1042	+	vterm = pref * pot_term
1043	+	vpair = vpair + vterm
1044	+	epot = epot + sw*vterm
1045	+
1046	+	! calculate derivatives for the forces and torques
1047	+	dudx = dudx + preSwSc * ( ctidotj*xhat*c4rij - &
1048	+	(ct_i*uz_j(1) + ct_j*uz_i(1) + ct_ij*xhat)*c3ri )
1049	+	dudy = dudy + preSwSc * ( ctidotj*yhat*c4rij - &
1050	+	(ct_i*uz_j(2) + ct_j*uz_i(2) + ct_ij*yhat)*c3ri )
1051	+	dudz = dudz + preSwSc * ( ctidotj*zhat*c4rij - &
1052	+	(ct_i*uz_j(3) + ct_j*uz_i(3) + ct_ij*zhat)*c3ri )
1053	+
1054	+	duduz_i(1) = duduz_i(1) + preSw * ( uz_j(1)*c2ri - ctj3*xhat )
1055	+	duduz_i(2) = duduz_i(2) + preSw * ( uz_j(2)*c2ri - ctj3*yhat )
1056	+	duduz_i(3) = duduz_i(3) + preSw * ( uz_j(3)*c2ri - ctj3*zhat )
1057
1058	<	vterm = pref * ri3 * (ct_ij - 3.0d0 * ct_i * ct_j * sc2)
1059	<	vpair = vpair + swi * vterm
1060	<	epot = epot + vterm
1061	<
838	<	a1 = 5.0d0 * ct_i * ct_j * sc2 - ct_ij
839	<
840	<	dudx = dudx + pref*3.0d0*ri4*scale &
841	<	*(a1*xhat-ct_i*uz_j(1)-ct_j*uz_i(1))
842	<	dudy = dudy + pref*3.0d0*ri4*scale &
843	<	*(a1*yhat-ct_i*uz_j(2)-ct_j*uz_i(2))
844	<	dudz = dudz + pref*3.0d0*ri4*scale &
845	<	*(a1*zhat-ct_i*uz_j(3)-ct_j*uz_i(3))
846	<
847	<	duduz_i(1) = duduz_i(1) + pref*ri3 &
848	<	*(uz_j(1) - 3.0d0*ct_j*xhat*sc2)
849	<	duduz_i(2) = duduz_i(2) + pref*ri3 &
850	<	*(uz_j(2) - 3.0d0*ct_j*yhat*sc2)
851	<	duduz_i(3) = duduz_i(3) + pref*ri3 &
852	<	*(uz_j(3) - 3.0d0*ct_j*zhat*sc2)
853	<
854	<	duduz_j(1) = duduz_j(1) + pref*ri3 &
855	<	*(uz_i(1) - 3.0d0*ct_i*xhat*sc2)
856	<	duduz_j(2) = duduz_j(2) + pref*ri3 &
857	<	*(uz_i(2) - 3.0d0*ct_i*yhat*sc2)
858	<	duduz_j(3) = duduz_j(3) + pref*ri3 &
859	<	*(uz_i(3) - 3.0d0*ct_i*zhat*sc2)
1058	>	duduz_j(1) = duduz_j(1) + preSw * ( uz_i(1)*c2ri - cti3*xhat )
1059	>	duduz_j(2) = duduz_j(2) + preSw * ( uz_i(2)*c2ri - cti3*yhat )
1060	>	duduz_j(3) = duduz_j(3) + preSw * ( uz_i(3)*c2ri - cti3*zhat )
1061	>
1062		endif
1063		endif
1064		endif
1065
1066		if (i_is_Quadrupole) then
1067		if (j_is_Charge) then
1068	<
1069	<	ri2 = riji * riji
1070	<	ri3 = ri2 * riji
1071	<	ri4 = ri2 * ri2
1068	>	if (screeningMethod .eq. DAMPED) then
1069	>	! assemble the damping variables
1070	>	call lookupUniformSpline1d(erfcSpline, rij, erfcVal, derfcVal)
1071	>	c1 = erfcVal*riji
1072	>	c2 = (-derfcVal + c1)*riji
1073	>	c3 = -2.0_dp*derfcVal*alpha2 + 3.0_dp*c2*riji
1074	>	c4 = -4.0_dp*derfcVal*alpha4 + 5.0_dp*c3*riji*riji
1075	>	else
1076	>	c1 = riji
1077	>	c2 = c1*riji
1078	>	c3 = 3.0_dp*c2*riji
1079	>	c4 = 5.0_dp*c3*riji*riji
1080	>	endif
1081	>
1082	>	! precompute some variables
1083		cx2 = cx_i * cx_i
1084		cy2 = cy_i * cy_i
1085		cz2 = cz_i * cz_i
1086	+	pref = pre14 * q_j * one_third
1087
1088	<	pref = sw * pre14 * q_j / 3.0_dp
1088	>	! calculate the potential
1089	>	pot_term = ( qxx_i * (cx2*c3 - c2ri) + qyy_i * (cy2*c3 - c2ri) + &
1090	>	qzz_i * (cz2*c3 - c2ri) )
1091
1092	<	if (corrMethod .eq. 1) then
1093	<	vterm1 = pref * ri3*( qxx_i * (3.0_dp*cx2 - 1.0_dp) + &
1094	<	qyy_i * (3.0_dp*cy2 - 1.0_dp) + &
1095	<	qzz_i * (3.0_dp*cz2 - 1.0_dp) )
1096	<	vterm2 = pref * rcuti3*( qxx_i * (3.0_dp*cx2 - 1.0_dp) + &
1097	<	qyy_i * (3.0_dp*cy2 - 1.0_dp) + &
1098	<	qzz_i * (3.0_dp*cz2 - 1.0_dp) )
1099	<	vpair = vpair + swi * ( vterm1 - vterm2 )
1100	<	epot = epot + ( vterm1 - vterm2 )
1101	<
1102	<	dudx = dudx - (5.0_dp*(vterm1*riji*xhat - vterm2*rcuti2*d(1))) + &
1103	<	pref * ( (ri4 - rcuti4)*(qxx_i*(6.0_dp*cx_i*ux_i(1)) - &
1104	<	qxx_i*2.0_dp*(xhat - rcuti*d(1))) + &
1105	<	(ri4 - rcuti4)*(qyy_i*(6.0_dp*cy_i*uy_i(1)) - &
1106	<	qyy_i*2.0_dp*(xhat - rcuti*d(1))) + &
891	<	(ri4 - rcuti4)*(qzz_i*(6.0_dp*cz_i*uz_i(1)) - &
892	<	qzz_i*2.0_dp*(xhat - rcuti*d(1))) )
893	<	dudy = dudy - (5.0_dp*(vterm1*riji*yhat - vterm2*rcuti2*d(2))) + &
894	<	pref * ( (ri4 - rcuti4)*(qxx_i*(6.0_dp*cx_i*ux_i(2)) - &
895	<	qxx_i*2.0_dp*(yhat - rcuti*d(2))) + &
896	<	(ri4 - rcuti4)*(qyy_i*(6.0_dp*cy_i*uy_i(2)) - &
897	<	qyy_i*2.0_dp*(yhat - rcuti*d(2))) + &
898	<	(ri4 - rcuti4)*(qzz_i*(6.0_dp*cz_i*uz_i(2)) - &
899	<	qzz_i*2.0_dp*(yhat - rcuti*d(2))) )
900	<	dudz = dudz - (5.0_dp*(vterm1*riji*zhat - vterm2*rcuti2*d(3))) + &
901	<	pref * ( (ri4 - rcuti4)*(qxx_i*(6.0_dp*cx_i*ux_i(3)) - &
902	<	qxx_i*2.0_dp*(zhat - rcuti*d(3))) + &
903	<	(ri4 - rcuti4)*(qyy_i*(6.0_dp*cy_i*uy_i(3)) - &
904	<	qyy_i*2.0_dp*(zhat - rcuti*d(3))) + &
905	<	(ri4 - rcuti4)*(qzz_i*(6.0_dp*cz_i*uz_i(3)) - &
906	<	qzz_i*2.0_dp*(zhat - rcuti*d(3))) )
907	<
908	<	dudux_i(1) = dudux_i(1) + pref * (ri3*(qxx_i*6.0_dp*cx_i*xhat) - &
909	<	rcuti4*(qxx_i*6.0_dp*cx_i*d(1)))
910	<	dudux_i(2) = dudux_i(2) + pref * (ri3*(qxx_i*6.0_dp*cx_i*yhat) - &
911	<	rcuti4*(qxx_i*6.0_dp*cx_i*d(2)))
912	<	dudux_i(3) = dudux_i(3) + pref * (ri3*(qxx_i*6.0_dp*cx_i*zhat) - &
913	<	rcuti4*(qxx_i*6.0_dp*cx_i*d(3)))
914	<
915	<	duduy_i(1) = duduy_i(1) + pref * (ri3*(qyy_i*6.0_dp*cy_i*xhat) - &
916	<	rcuti4*(qyy_i*6.0_dp*cx_i*d(1)))
917	<	duduy_i(2) = duduy_i(2) + pref * (ri3*(qyy_i*6.0_dp*cy_i*yhat) - &
918	<	rcuti4*(qyy_i*6.0_dp*cx_i*d(2)))
919	<	duduy_i(3) = duduy_i(3) + pref * (ri3*(qyy_i*6.0_dp*cy_i*zhat) - &
920	<	rcuti4*(qyy_i*6.0_dp*cx_i*d(3)))
921	<
922	<	duduz_i(1) = duduz_i(1) + pref * (ri3*(qzz_i*6.0_dp*cz_i*xhat) - &
923	<	rcuti4*(qzz_i*6.0_dp*cx_i*d(1)))
924	<	duduz_i(2) = duduz_i(2) + pref * (ri3*(qzz_i*6.0_dp*cz_i*yhat) - &
925	<	rcuti4*(qzz_i*6.0_dp*cx_i*d(2)))
926	<	duduz_i(3) = duduz_i(3) + pref * (ri3*(qzz_i*6.0_dp*cz_i*zhat) - &
927	<	rcuti4*(qzz_i*6.0_dp*cx_i*d(3)))
1092	>	vterm = pref * pot_term
1093	>	vpair = vpair + vterm
1094	>	epot = epot + sw*vterm
1095	>
1096	>	! precompute variables for convenience
1097	>	preSw = sw*pref
1098	>	c2ri = c2*riji
1099	>	c3ri = c3*riji
1100	>	c4rij = c4*rij
1101	>	xhatdot2 = 2.0_dp*xhat*c3
1102	>	yhatdot2 = 2.0_dp*yhat*c3
1103	>	zhatdot2 = 2.0_dp*zhat*c3
1104	>	xhatc4 = xhat*c4rij
1105	>	yhatc4 = yhat*c4rij
1106	>	zhatc4 = zhat*c4rij
1107
1108	<	else
1109	<	vterm = pref * ri3 * (qxx_i * (3.0_dp*cx2 - 1.0_dp) + &
1110	<	qyy_i * (3.0_dp*cy2 - 1.0_dp) + &
1111	<	qzz_i * (3.0_dp*cz2 - 1.0_dp))
1112	<	vpair = vpair + swi * vterm
1113	<	epot = epot + vterm
1114	<
1115	<	dudx = dudx - 5.0_dp*vterm*riji*xhat + pref * ri4 * ( &
1116	<	qxx_i*(6.0_dp*cx_i*ux_i(1) - 2.0_dp*xhat) + &
1117	<	qyy_i*(6.0_dp*cy_i*uy_i(1) - 2.0_dp*xhat) + &
1118	<	qzz_i*(6.0_dp*cz_i*uz_i(1) - 2.0_dp*xhat) )
1119	<	dudy = dudy - 5.0_dp*vterm*riji*yhat + pref * ri4 * ( &
1120	<	qxx_i*(6.0_dp*cx_i*ux_i(2) - 2.0_dp*yhat) + &
1121	<	qyy_i*(6.0_dp*cy_i*uy_i(2) - 2.0_dp*yhat) + &
1122	<	qzz_i*(6.0_dp*cz_i*uz_i(2) - 2.0_dp*yhat) )
1123	<	dudz = dudz - 5.0_dp*vterm*riji*zhat + pref * ri4 * ( &
1124	<	qxx_i*(6.0_dp*cx_i*ux_i(3) - 2.0_dp*zhat) + &
1125	<	qyy_i*(6.0_dp*cy_i*uy_i(3) - 2.0_dp*zhat) + &
1126	<	qzz_i*(6.0_dp*cz_i*uz_i(3) - 2.0_dp*zhat) )
1127	<
1128	<	dudux_i(1) = dudux_i(1) + pref * ri3*(qxx_i*6.0_dp*cx_i*xhat)
1129	<	dudux_i(2) = dudux_i(2) + pref * ri3*(qxx_i*6.0_dp*cx_i*yhat)
1130	<	dudux_i(3) = dudux_i(3) + pref * ri3*(qxx_i*6.0_dp*cx_i*zhat)
1131	<
1132	<	duduy_i(1) = duduy_i(1) + pref * ri3*(qyy_i*6.0_dp*cy_i*xhat)
954	<	duduy_i(2) = duduy_i(2) + pref * ri3*(qyy_i*6.0_dp*cy_i*yhat)
955	<	duduy_i(3) = duduy_i(3) + pref * ri3*(qyy_i*6.0_dp*cy_i*zhat)
956	<
957	<	duduz_i(1) = duduz_i(1) + pref * ri3*(qzz_i*6.0_dp*cz_i*xhat)
958	<	duduz_i(2) = duduz_i(2) + pref * ri3*(qzz_i*6.0_dp*cz_i*yhat)
959	<	duduz_i(3) = duduz_i(3) + pref * ri3*(qzz_i*6.0_dp*cz_i*zhat)
960	<	endif
1108	>	! calculate the derivatives for the forces and torques
1109	>	dudx = dudx - preSw * ( &
1110	>	qxx_i*(cx2*xhatc4 - (2.0_dp*cx_i*ux_i(1) + xhat)*c3ri) + &
1111	>	qyy_i*(cy2*xhatc4 - (2.0_dp*cy_i*uy_i(1) + xhat)*c3ri) + &
1112	>	qzz_i*(cz2*xhatc4 - (2.0_dp*cz_i*uz_i(1) + xhat)*c3ri) )
1113	>	dudy = dudy - preSw * ( &
1114	>	qxx_i*(cx2*yhatc4 - (2.0_dp*cx_i*ux_i(2) + yhat)*c3ri) + &
1115	>	qyy_i*(cy2*yhatc4 - (2.0_dp*cy_i*uy_i(2) + yhat)*c3ri) + &
1116	>	qzz_i*(cz2*yhatc4 - (2.0_dp*cz_i*uz_i(2) + yhat)*c3ri) )
1117	>	dudz = dudz - preSw * ( &
1118	>	qxx_i*(cx2*zhatc4 - (2.0_dp*cx_i*ux_i(3) + zhat)*c3ri) + &
1119	>	qyy_i*(cy2*zhatc4 - (2.0_dp*cy_i*uy_i(3) + zhat)*c3ri) + &
1120	>	qzz_i*(cz2*zhatc4 - (2.0_dp*cz_i*uz_i(3) + zhat)*c3ri) )
1121	>
1122	>	dudux_i(1) = dudux_i(1) + preSw*(qxx_i*cx_i*xhatdot2)
1123	>	dudux_i(2) = dudux_i(2) + preSw*(qxx_i*cx_i*yhatdot2)
1124	>	dudux_i(3) = dudux_i(3) + preSw*(qxx_i*cx_i*zhatdot2)
1125	>
1126	>	duduy_i(1) = duduy_i(1) + preSw*(qyy_i*cy_i*xhatdot2)
1127	>	duduy_i(2) = duduy_i(2) + preSw*(qyy_i*cy_i*yhatdot2)
1128	>	duduy_i(3) = duduy_i(3) + preSw*(qyy_i*cy_i*zhatdot2)
1129	>
1130	>	duduz_i(1) = duduz_i(1) + preSw*(qzz_i*cz_i*xhatdot2)
1131	>	duduz_i(2) = duduz_i(2) + preSw*(qzz_i*cz_i*yhatdot2)
1132	>	duduz_i(3) = duduz_i(3) + preSw*(qzz_i*cz_i*zhatdot2)
1133		endif
1134		endif
1135
1136
1137		if (do_pot) then
1138		#ifdef IS_MPI
1139	<	pot_row(atom1) = pot_row(atom1) + 0.5d0*epot
1140	<	pot_col(atom2) = pot_col(atom2) + 0.5d0*epot
1139	>	pot_row(ELECTROSTATIC_POT,atom1) = pot_row(ELECTROSTATIC_POT,atom1) + 0.5_dp*epot
1140	>	pot_col(ELECTROSTATIC_POT,atom2) = pot_col(ELECTROSTATIC_POT,atom2) + 0.5_dp*epot
1141		#else
1142		pot = pot + epot
1143		#endif
#	Line 1070 | Line 1242 | contains
1242		return
1243		end subroutine doElectrostaticPair
1244
1245	<	!! calculates the switching functions and their derivatives for a given
1074	<	subroutine calc_switch(r, mu, scale, dscale)
1245	>	subroutine destroyElectrostaticTypes()
1246
1247	<	real (kind=dp), intent(in) :: r, mu
1077	<	real (kind=dp), intent(inout) :: scale, dscale
1078	<	real (kind=dp) :: rl, ru, mulow, minRatio, temp, scaleVal
1247	>	if(allocated(ElectrostaticMap)) deallocate(ElectrostaticMap)
1248
1249	<	! distances must be in angstroms
1250	<	rl = 2.75d0
1251	<	ru = 3.75d0
1252	<	mulow = 0.0d0 !3.3856d0 ! 1.84 * 1.84
1253	<	minRatio = mulow / (mu*mu)
1254	<	scaleVal = 1.0d0 - minRatio
1249	>	end subroutine destroyElectrostaticTypes
1250	>
1251	>	subroutine self_self(atom1, eFrame, mypot, t, do_pot)
1252	>	logical, intent(in) :: do_pot
1253	>	integer, intent(in) :: atom1
1254	>	integer :: atid1
1255	>	real(kind=dp), dimension(9,nLocal) :: eFrame
1256	>	real(kind=dp), dimension(3,nLocal) :: t
1257	>	real(kind=dp) :: mu1, chg1
1258	>	real(kind=dp) :: preVal, epot, mypot
1259	>	real(kind=dp) :: eix, eiy, eiz
1260	>
1261	>	! this is a local only array, so we use the local atom type id's:
1262	>	atid1 = atid(atom1)
1263	>
1264	>	if (.not.summationMethodChecked) then
1265	>	call checkSummationMethod()
1266	>	endif
1267
1268	<	if (r.lt.rl) then
1269	<	scale = minRatio
1270	<	dscale = 0.0d0
1271	<	elseif (r.gt.ru) then
1272	<	scale = 1.0d0
1273	<	dscale = 0.0d0
1274	<	else
1275	<	scale = 1.0d0 - scaleVal*((ru + 2.0d0*r - 3.0d0*rl) * (ru-r)**2) &
1276	<	/ ((ru - rl)**3)
1277	<	dscale = -scaleVal * 6.0d0 * (r-ru)*(r-rl)/((ru - rl)**3)
1268	>	if (summationMethod .eq. REACTION_FIELD) then
1269	>	if (ElectrostaticMap(atid1)%is_Dipole) then
1270	>	mu1 = getDipoleMoment(atid1)
1271	>
1272	>	preVal = pre22 * preRF2 * mu1*mu1
1273	>	mypot = mypot - 0.5_dp*preVal
1274	>
1275	>	! The self-correction term adds into the reaction field vector
1276	>
1277	>	eix = preVal * eFrame(3,atom1)
1278	>	eiy = preVal * eFrame(6,atom1)
1279	>	eiz = preVal * eFrame(9,atom1)
1280	>
1281	>	! once again, this is self-self, so only the local arrays are needed
1282	>	! even for MPI jobs:
1283	>
1284	>	t(1,atom1)=t(1,atom1) - eFrame(6,atom1)*eiz + &
1285	>	eFrame(9,atom1)*eiy
1286	>	t(2,atom1)=t(2,atom1) - eFrame(9,atom1)*eix + &
1287	>	eFrame(3,atom1)*eiz
1288	>	t(3,atom1)=t(3,atom1) - eFrame(3,atom1)*eiy + &
1289	>	eFrame(6,atom1)*eix
1290	>
1291	>	endif
1292	>
1293	>	elseif ( (summationMethod .eq. SHIFTED_FORCE) .or. &
1294	>	(summationMethod .eq. SHIFTED_POTENTIAL) ) then
1295	>	if (ElectrostaticMap(atid1)%is_Charge) then
1296	>	chg1 = getCharge(atid1)
1297	>
1298	>	if (screeningMethod .eq. DAMPED) then
1299	>	mypot = mypot - (c1c * 0.5_dp + &
1300	>	dampingAlpha*invRootPi) * chg1 * chg1
1301	>
1302	>	else
1303	>	mypot = mypot - (rcuti * 0.5_dp * chg1 * chg1)
1304	>
1305	>	endif
1306	>	endif
1307		endif
1308	<
1308	>
1309		return
1310	<	end subroutine calc_switch
1310	>	end subroutine self_self
1311
1312	<	subroutine destroyElectrostaticTypes()
1312	>	subroutine rf_self_excludes(atom1, atom2, sw, eFrame, d, rij, vpair, myPot, &
1313	>	f, t, do_pot)
1314	>	logical, intent(in) :: do_pot
1315	>	integer, intent(in) :: atom1
1316	>	integer, intent(in) :: atom2
1317	>	logical :: i_is_Charge, j_is_Charge
1318	>	logical :: i_is_Dipole, j_is_Dipole
1319	>	integer :: atid1
1320	>	integer :: atid2
1321	>	real(kind=dp), intent(in) :: rij
1322	>	real(kind=dp), intent(in) :: sw
1323	>	real(kind=dp), intent(in), dimension(3) :: d
1324	>	real(kind=dp), intent(inout) :: vpair
1325	>	real(kind=dp), dimension(9,nLocal) :: eFrame
1326	>	real(kind=dp), dimension(3,nLocal) :: f
1327	>	real(kind=dp), dimension(3,nLocal) :: t
1328	>	real (kind = dp), dimension(3) :: duduz_i
1329	>	real (kind = dp), dimension(3) :: duduz_j
1330	>	real (kind = dp), dimension(3) :: uz_i
1331	>	real (kind = dp), dimension(3) :: uz_j
1332	>	real(kind=dp) :: q_i, q_j, mu_i, mu_j
1333	>	real(kind=dp) :: xhat, yhat, zhat
1334	>	real(kind=dp) :: ct_i, ct_j
1335	>	real(kind=dp) :: ri2, ri3, riji, vterm
1336	>	real(kind=dp) :: pref, preVal, rfVal, myPot
1337	>	real(kind=dp) :: dudx, dudy, dudz, dudr
1338
1339	<	if(allocated(ElectrostaticMap)) deallocate(ElectrostaticMap)
1339	>	if (.not.summationMethodChecked) then
1340	>	call checkSummationMethod()
1341	>	endif
1342
1343	<	end subroutine destroyElectrostaticTypes
1343	>	dudx = zero
1344	>	dudy = zero
1345	>	dudz = zero
1346
1347	+	riji = 1.0_dp/rij
1348	+
1349	+	xhat = d(1) * riji
1350	+	yhat = d(2) * riji
1351	+	zhat = d(3) * riji
1352	+
1353	+	! this is a local only array, so we use the local atom type id's:
1354	+	atid1 = atid(atom1)
1355	+	atid2 = atid(atom2)
1356	+	i_is_Charge = ElectrostaticMap(atid1)%is_Charge
1357	+	j_is_Charge = ElectrostaticMap(atid2)%is_Charge
1358	+	i_is_Dipole = ElectrostaticMap(atid1)%is_Dipole
1359	+	j_is_Dipole = ElectrostaticMap(atid2)%is_Dipole
1360	+
1361	+	if (i_is_Charge.and.j_is_Charge) then
1362	+	q_i = ElectrostaticMap(atid1)%charge
1363	+	q_j = ElectrostaticMap(atid2)%charge
1364	+
1365	+	preVal = pre11 * q_i * q_j
1366	+	rfVal = preRF*rij*rij
1367	+	vterm = preVal * rfVal
1368	+
1369	+	myPot = myPot + sw*vterm
1370	+
1371	+	dudr  = sw*preVal * 2.0_dp*rfVal*riji
1372	+
1373	+	dudx = dudx + dudr * xhat
1374	+	dudy = dudy + dudr * yhat
1375	+	dudz = dudz + dudr * zhat
1376	+
1377	+	elseif (i_is_Charge.and.j_is_Dipole) then
1378	+	q_i = ElectrostaticMap(atid1)%charge
1379	+	mu_j = ElectrostaticMap(atid2)%dipole_moment
1380	+	uz_j(1) = eFrame(3,atom2)
1381	+	uz_j(2) = eFrame(6,atom2)
1382	+	uz_j(3) = eFrame(9,atom2)
1383	+	ct_j = uz_j(1)*xhat + uz_j(2)*yhat + uz_j(3)*zhat
1384	+
1385	+	ri2 = riji * riji
1386	+	ri3 = ri2 * riji
1387	+
1388	+	pref = pre12 * q_i * mu_j
1389	+	vterm = - pref * ct_j * ( ri2 - preRF2*rij )
1390	+	myPot = myPot + sw*vterm
1391	+
1392	+	dudx = dudx - sw*pref*( ri3*(uz_j(1)-3.0_dp*ct_j*xhat) &
1393	+	- preRF2*uz_j(1) )
1394	+	dudy = dudy - sw*pref*( ri3*(uz_j(2)-3.0_dp*ct_j*yhat) &
1395	+	- preRF2*uz_j(2) )
1396	+	dudz = dudz - sw*pref*( ri3*(uz_j(3)-3.0_dp*ct_j*zhat) &
1397	+	- preRF2*uz_j(3) )
1398	+
1399	+	duduz_j(1) = duduz_j(1) - sw * pref * xhat * ( ri2 - preRF2*rij )
1400	+	duduz_j(2) = duduz_j(2) - sw * pref * yhat * ( ri2 - preRF2*rij )
1401	+	duduz_j(3) = duduz_j(3) - sw * pref * zhat * ( ri2 - preRF2*rij )
1402	+
1403	+	elseif (i_is_Dipole.and.j_is_Charge) then
1404	+	mu_i = ElectrostaticMap(atid1)%dipole_moment
1405	+	q_j = ElectrostaticMap(atid2)%charge
1406	+	uz_i(1) = eFrame(3,atom1)
1407	+	uz_i(2) = eFrame(6,atom1)
1408	+	uz_i(3) = eFrame(9,atom1)
1409	+	ct_i = uz_i(1)*xhat + uz_i(2)*yhat + uz_i(3)*zhat
1410	+
1411	+	ri2 = riji * riji
1412	+	ri3 = ri2 * riji
1413	+
1414	+	pref = pre12 * q_j * mu_i
1415	+	vterm = pref * ct_i * ( ri2 - preRF2*rij )
1416	+	myPot = myPot + sw*vterm
1417	+
1418	+	dudx = dudx + sw*pref*( ri3*(uz_i(1)-3.0_dp*ct_i*xhat) &
1419	+	- preRF2*uz_i(1) )
1420	+	dudy = dudy + sw*pref*( ri3*(uz_i(2)-3.0_dp*ct_i*yhat) &
1421	+	- preRF2*uz_i(2) )
1422	+	dudz = dudz + sw*pref*( ri3*(uz_i(3)-3.0_dp*ct_i*zhat) &
1423	+	- preRF2*uz_i(3) )
1424	+
1425	+	duduz_i(1) = duduz_i(1) + sw * pref * xhat * ( ri2 - preRF2*rij )
1426	+	duduz_i(2) = duduz_i(2) + sw * pref * yhat * ( ri2 - preRF2*rij )
1427	+	duduz_i(3) = duduz_i(3) + sw * pref * zhat * ( ri2 - preRF2*rij )
1428	+
1429	+	endif
1430	+
1431	+
1432	+	! accumulate the forces and torques resulting from the self term
1433	+	f(1,atom1) = f(1,atom1) + dudx
1434	+	f(2,atom1) = f(2,atom1) + dudy
1435	+	f(3,atom1) = f(3,atom1) + dudz
1436	+
1437	+	f(1,atom2) = f(1,atom2) - dudx
1438	+	f(2,atom2) = f(2,atom2) - dudy
1439	+	f(3,atom2) = f(3,atom2) - dudz
1440	+
1441	+	if (i_is_Dipole) then
1442	+	t(1,atom1)=t(1,atom1) - uz_i(2)*duduz_i(3) + uz_i(3)*duduz_i(2)
1443	+	t(2,atom1)=t(2,atom1) - uz_i(3)*duduz_i(1) + uz_i(1)*duduz_i(3)
1444	+	t(3,atom1)=t(3,atom1) - uz_i(1)*duduz_i(2) + uz_i(2)*duduz_i(1)
1445	+	elseif (j_is_Dipole) then
1446	+	t(1,atom2)=t(1,atom2) - uz_j(2)*duduz_j(3) + uz_j(3)*duduz_j(2)
1447	+	t(2,atom2)=t(2,atom2) - uz_j(3)*duduz_j(1) + uz_j(1)*duduz_j(3)
1448	+	t(3,atom2)=t(3,atom2) - uz_j(1)*duduz_j(2) + uz_j(2)*duduz_j(1)
1449	+	endif
1450	+
1451	+	return
1452	+	end subroutine rf_self_excludes
1453	+
1454		end module electrostatic_module

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines