ViewVC Help

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

root/group/trunk/OOPSE/libmdtools/do_Forces.F90

Comparing trunk/OOPSE/libmdtools/do_Forces.F90 (file contents):
Revision 388 by chuckv, Fri Mar 21 22:11:50 2003 UTC vs.
Revision 1192 by gezelter, Mon May 24 21:03:30 2004 UTC

#	Line 4 | Line 4
4
5		!! @author Charles F. Vardeman II
6		!! @author Matthew Meineke
7	<	!! @version $Id: do_Forces.F90,v 1.2 2003-03-21 22:11:50 chuckv Exp $, $Date: 2003-03-21 22:11:50 $, $Name: not supported by cvs2svn $, $Revision: 1.2 $
7	>	!! @version $Id: do_Forces.F90,v 1.61 2004-05-24 21:03:25 gezelter Exp $, $Date: 2004-05-24 21:03:25 $, $Name: not supported by cvs2svn $, $Revision: 1.61 $
8
9		module do_Forces
10		use force_globals
11		use simulation
12		use definitions
13		use atype_module
14	+	use switcheroo
15		use neighborLists
16		use lj
17		use sticky_pair
18		use dipole_dipole
19	+	use charge_charge
20		use reaction_field
21		use gb_pair
22	+	use vector_class
23	+	use eam
24	+	use status
25		#ifdef IS_MPI
26		use mpiSimulation
27		#endif
#	Line 26 | Line 31 | module do_Forces
31
32		#define __FORTRAN90
33		#include "fForceField.h"
34	+	#include "fSwitchingFunction.h"
35
36	<	logical, save :: do_forces_initialized = .false.
36	>	logical, save :: haveRlist = .false.
37	>	logical, save :: haveNeighborList = .false.
38	>	logical, save :: havePolicies = .false.
39	>	logical, save :: haveSIMvariables = .false.
40	>	logical, save :: havePropertyMap = .false.
41	>	logical, save :: haveSaneForceField = .false.
42		logical, save :: FF_uses_LJ
43		logical, save :: FF_uses_sticky
44	+	logical, save :: FF_uses_charges
45		logical, save :: FF_uses_dipoles
46		logical, save :: FF_uses_RF
47		logical, save :: FF_uses_GB
48		logical, save :: FF_uses_EAM
49	+	logical, save :: SIM_uses_LJ
50	+	logical, save :: SIM_uses_sticky
51	+	logical, save :: SIM_uses_charges
52	+	logical, save :: SIM_uses_dipoles
53	+	logical, save :: SIM_uses_RF
54	+	logical, save :: SIM_uses_GB
55	+	logical, save :: SIM_uses_EAM
56	+	logical, save :: SIM_requires_postpair_calc
57	+	logical, save :: SIM_requires_prepair_calc
58	+	logical, save :: SIM_uses_directional_atoms
59	+	logical, save :: SIM_uses_PBC
60	+	logical, save :: SIM_uses_molecular_cutoffs
61
62	+	real(kind=dp), save :: rlist, rlistsq
63	+
64		public :: init_FF
65		public :: do_force_loop
66	+	public :: setRlistDF
67
68	+	#ifdef PROFILE
69	+	public :: getforcetime
70	+	real, save :: forceTime = 0
71	+	real :: forceTimeInitial, forceTimeFinal
72	+	integer :: nLoops
73	+	#endif
74	+
75	+	type :: Properties
76	+	logical :: is_lj     = .false.
77	+	logical :: is_sticky = .false.
78	+	logical :: is_dp     = .false.
79	+	logical :: is_gb     = .false.
80	+	logical :: is_eam    = .false.
81	+	logical :: is_charge = .false.
82	+	real(kind=DP) :: charge = 0.0_DP
83	+	real(kind=DP) :: dipole_moment = 0.0_DP
84	+	end type Properties
85	+
86	+	type(Properties), dimension(:),allocatable :: PropertyMap
87	+
88		contains
89
90	+	subroutine setRlistDF( this_rlist )
91	+
92	+	real(kind=dp) :: this_rlist
93	+
94	+	rlist = this_rlist
95	+	rlistsq = rlist * rlist
96	+
97	+	haveRlist = .true.
98	+
99	+	end subroutine setRlistDF
100	+
101	+	subroutine createPropertyMap(status)
102	+	integer :: nAtypes
103	+	integer :: status
104	+	integer :: i
105	+	logical :: thisProperty
106	+	real (kind=DP) :: thisDPproperty
107	+
108	+	status = 0
109	+
110	+	nAtypes = getSize(atypes)
111	+
112	+	if (nAtypes == 0) then
113	+	status = -1
114	+	return
115	+	end if
116	+
117	+	if (.not. allocated(PropertyMap)) then
118	+	allocate(PropertyMap(nAtypes))
119	+	endif
120	+
121	+	do i = 1, nAtypes
122	+	call getElementProperty(atypes, i, "is_LJ", thisProperty)
123	+	PropertyMap(i)%is_LJ = thisProperty
124	+
125	+	call getElementProperty(atypes, i, "is_Charge", thisProperty)
126	+	PropertyMap(i)%is_Charge = thisProperty
127	+
128	+	if (thisProperty) then
129	+	call getElementProperty(atypes, i, "charge", thisDPproperty)
130	+	PropertyMap(i)%charge = thisDPproperty
131	+	endif
132	+
133	+	call getElementProperty(atypes, i, "is_DP", thisProperty)
134	+	PropertyMap(i)%is_DP = thisProperty
135	+
136	+	if (thisProperty) then
137	+	call getElementProperty(atypes, i, "dipole_moment", thisDPproperty)
138	+	PropertyMap(i)%dipole_moment = thisDPproperty
139	+	endif
140	+
141	+	call getElementProperty(atypes, i, "is_Sticky", thisProperty)
142	+	PropertyMap(i)%is_Sticky = thisProperty
143	+	call getElementProperty(atypes, i, "is_GB", thisProperty)
144	+	PropertyMap(i)%is_GB = thisProperty
145	+	call getElementProperty(atypes, i, "is_EAM", thisProperty)
146	+	PropertyMap(i)%is_EAM = thisProperty
147	+	end do
148	+
149	+	havePropertyMap = .true.
150	+
151	+	end subroutine createPropertyMap
152	+
153	+	subroutine setSimVariables()
154	+	SIM_uses_LJ = SimUsesLJ()
155	+	SIM_uses_sticky = SimUsesSticky()
156	+	SIM_uses_charges = SimUsesCharges()
157	+	SIM_uses_dipoles = SimUsesDipoles()
158	+	SIM_uses_RF = SimUsesRF()
159	+	SIM_uses_GB = SimUsesGB()
160	+	SIM_uses_EAM = SimUsesEAM()
161	+	SIM_requires_postpair_calc = SimRequiresPostpairCalc()
162	+	SIM_requires_prepair_calc = SimRequiresPrepairCalc()
163	+	SIM_uses_directional_atoms = SimUsesDirectionalAtoms()
164	+	SIM_uses_PBC = SimUsesPBC()
165	+	!SIM_uses_molecular_cutoffs = SimUsesMolecularCutoffs()
166	+
167	+	haveSIMvariables = .true.
168	+
169	+	return
170	+	end subroutine setSimVariables
171	+
172	+	subroutine doReadyCheck(error)
173	+	integer, intent(out) :: error
174	+
175	+	integer :: myStatus
176	+
177	+	error = 0
178	+
179	+	if (.not. havePropertyMap) then
180	+
181	+	myStatus = 0
182	+
183	+	call createPropertyMap(myStatus)
184	+
185	+	if (myStatus .ne. 0) then
186	+	write(default_error, *) 'createPropertyMap failed in do_Forces!'
187	+	error = -1
188	+	return
189	+	endif
190	+	endif
191	+
192	+	if (.not. haveSIMvariables) then
193	+	call setSimVariables()
194	+	endif
195	+
196	+	if (.not. haveRlist) then
197	+	write(default_error, *) 'rList has not been set in do_Forces!'
198	+	error = -1
199	+	return
200	+	endif
201	+
202	+	if (SIM_uses_LJ .and. FF_uses_LJ) then
203	+	if (.not. havePolicies) then
204	+	write(default_error, *) 'LJ mixing Policies have not been set in do_Forces!'
205	+	error = -1
206	+	return
207	+	endif
208	+	endif
209	+
210	+	if (.not. haveNeighborList) then
211	+	write(default_error, *) 'neighbor list has not been initialized in do_Forces!'
212	+	error = -1
213	+	return
214	+	end if
215	+
216	+	if (.not. haveSaneForceField) then
217	+	write(default_error, *) 'Force Field is not sane in do_Forces!'
218	+	error = -1
219	+	return
220	+	end if
221	+
222	+	#ifdef IS_MPI
223	+	if (.not. isMPISimSet()) then
224	+	write(default_error,*) "ERROR: mpiSimulation has not been initialized!"
225	+	error = -1
226	+	return
227	+	endif
228	+	#endif
229	+	return
230	+	end subroutine doReadyCheck
231	+
232	+
233		subroutine init_FF(LJMIXPOLICY, use_RF_c, thisStat)
234
235		integer, intent(in) :: LJMIXPOLICY
#	Line 64 | Line 254 | contains
254
255		FF_uses_LJ = .false.
256		FF_uses_sticky = .false.
257	+	FF_uses_charges = .false.
258		FF_uses_dipoles = .false.
259		FF_uses_GB = .false.
260		FF_uses_EAM = .false.
261
262		call getMatchingElementList(atypes, "is_LJ", .true., nMatches, MatchList)
263		if (nMatches .gt. 0) FF_uses_LJ = .true.
264	<
264	>
265	>	call getMatchingElementList(atypes, "is_Charge", .true., nMatches, MatchList)
266	>	if (nMatches .gt. 0) FF_uses_charges = .true.
267	>
268		call getMatchingElementList(atypes, "is_DP", .true., nMatches, MatchList)
269		if (nMatches .gt. 0) FF_uses_dipoles = .true.
270
#	Line 84 | Line 278 | contains
278		call getMatchingElementList(atypes, "is_EAM", .true., nMatches, MatchList)
279		if (nMatches .gt. 0) FF_uses_EAM = .true.
280
281	+	!! Assume sanity (for the sake of argument)
282	+	haveSaneForceField = .true.
283	+
284		!! check to make sure the FF_uses_RF setting makes sense
285
286		if (FF_uses_dipoles) then
90	–	rrf = getRrf()
91	–	rt = getRt()
92	–	call initialize_dipole(rrf, rt)
287		if (FF_uses_RF) then
288		dielect = getDielect()
289	<	call initialize_rf(rrf, rt, dielect)
289	>	call initialize_rf(dielect)
290		endif
291		else
292		if (FF_uses_RF) then
293		write(default_error,*) 'Using Reaction Field with no dipoles?  Huh?'
294		thisStat = -1
295	+	haveSaneForceField = .false.
296		return
297		endif
298	<	endif
298	>	endif
299
300		if (FF_uses_LJ) then
301
107	–	call getRcut(rcut)
108	–
302		select case (LJMIXPOLICY)
303		case (LB_MIXING_RULE)
304	<	call init_lj_FF(LB_MIXING_RULE, rcut, my_status)
304	>	call init_lj_FF(LB_MIXING_RULE, my_status)
305		case (EXPLICIT_MIXING_RULE)
306	<	call init_lj_FF(EXPLICIT_MIXING_RULE, rcut, my_status)
306	>	call init_lj_FF(EXPLICIT_MIXING_RULE, my_status)
307		case default
308		write(default_error,*) 'unknown LJ Mixing Policy!'
309		thisStat = -1
310	+	haveSaneForceField = .false.
311		return
312		end select
313		if (my_status /= 0) then
314		thisStat = -1
315	+	haveSaneForceField = .false.
316		return
317		end if
318	+	havePolicies = .true.
319		endif
320
321		if (FF_uses_sticky) then
322		call check_sticky_FF(my_status)
323		if (my_status /= 0) then
324		thisStat = -1
325	+	haveSaneForceField = .false.
326		return
327		end if
328		endif
329	<
329	>
330	>
331	>	if (FF_uses_EAM) then
332	>	call init_EAM_FF(my_status)
333	>	if (my_status /= 0) then
334	>	write(default_error, *) "init_EAM_FF returned a bad status"
335	>	thisStat = -1
336	>	haveSaneForceField = .false.
337	>	return
338	>	end if
339	>	endif
340	>
341		if (FF_uses_GB) then
342		call check_gb_pair_FF(my_status)
343		if (my_status .ne. 0) then
344		thisStat = -1
345	+	haveSaneForceField = .false.
346		return
347		endif
348		endif
349
350		if (FF_uses_GB .and. FF_uses_LJ) then
351		endif
352	+	if (.not. haveNeighborList) then
353	+	!! Create neighbor lists
354	+	call expandNeighborList(nLocal, my_status)
355	+	if (my_Status /= 0) then
356	+	write(default_error,*) "SimSetup: ExpandNeighborList returned error."
357	+	thisStat = -1
358	+	return
359	+	endif
360	+	haveNeighborList = .true.
361	+	endif
362
363	<
364	<	do_forces_initialized = .true.
146	<
363	>
364	>
365		end subroutine init_FF
366
367
368		!! Does force loop over i,j pairs. Calls do_pair to calculates forces.
369		!------------------------------------------------------------->
370	<	subroutine do_force_loop(q, A, u_l, f, t, tau, pot, do_pot_c, do_stress_c, &
371	<	error)
370	>	subroutine do_force_loop(q, q_group, A, u_l, f, t, tau, pot, &
371	>	do_pot_c, do_stress_c, error)
372		!! Position array provided by C, dimensioned by getNlocal
373	<	real ( kind = dp ), dimension(3,getNlocal()) :: q
373	>	real ( kind = dp ), dimension(3, nLocal) :: q
374	>	!! molecular center-of-mass position array
375	>	real ( kind = dp ), dimension(3, nGroup) :: q_group
376		!! Rotation Matrix for each long range particle in simulation.
377	<	real( kind = dp), dimension(9,getNlocal()) :: A
377	>	real( kind = dp), dimension(9, nLocal) :: A
378		!! Unit vectors for dipoles (lab frame)
379	<	real( kind = dp ), dimension(3,getNlocal()) :: u_l
379	>	real( kind = dp ), dimension(3,nLocal) :: u_l
380		!! Force array provided by C, dimensioned by getNlocal
381	<	real ( kind = dp ), dimension(3,getNlocal()) :: f
381	>	real ( kind = dp ), dimension(3,nLocal) :: f
382		!! Torsion array provided by C, dimensioned by getNlocal
383	<	real( kind = dp ), dimension(3,getNlocal()) :: t
383	>	real( kind = dp ), dimension(3,nLocal) :: t
384	>
385		!! Stress Tensor
386		real( kind = dp), dimension(9) :: tau
387		real ( kind = dp ) :: pot
388		logical ( kind = 2) :: do_pot_c, do_stress_c
389		logical :: do_pot
390		logical :: do_stress
391	<	#ifdef IS_MPI
391	>	logical :: in_switching_region
392	>	#ifdef IS_MPI
393		real( kind = DP ) :: pot_local
394		integer :: nrow
395		integer :: ncol
396	+	integer :: nprocs
397	+	integer :: nrow_group
398	+	integer :: ncol_group
399		#endif
175	–	integer :: nlocal
400		integer :: natoms
401		logical :: update_nlist
402		integer :: i, j, jbeg, jend, jnab
403	+	integer :: istart, iend, jstart
404	+	integer :: ia, jb, atom1, atom2
405		integer :: nlist
406	<	real( kind = DP ) ::  rijsq, rlistsq, rcutsq, rlist, rcut
407	<	real(kind=dp),dimension(3) :: d
406	>	real( kind = DP ) :: ratmsq, rgrpsq, rgrp, vpair, vij
407	>	real( kind = DP ) :: sw, dswdr, swderiv, mf
408	>	real(kind=dp),dimension(3) :: d_atm, d_grp, fpair, fij
409		real(kind=dp) :: rfpot, mu_i, virial
410	<	integer :: me_i
410	>	integer :: me_i, me_j, n_in_i, n_in_j
411		logical :: is_dp_i
412		integer :: neighborListSize
413		integer :: listerror, error
414		integer :: localError
415	+	integer :: propPack_i, propPack_j
416
417	+	real(kind=dp) :: listSkin = 1.0
418	+
419		!! initialize local variables
420	<
420	>
421		#ifdef IS_MPI
422	<	nlocal = getNlocal()
422	>	pot_local = 0.0_dp
423		nrow   = getNrow(plan_row)
424		ncol   = getNcol(plan_col)
425	+	nrow_group   = getNrowGroup(plan_row)
426	+	ncol_group   = getNcolGroup(plan_col)
427		#else
196	–	nlocal = getNlocal()
428		natoms = nlocal
429		#endif
199	–
200	–	call getRcut(rcut,rc2=rcutsq)
201	–	call getRlist(rlist,rlistsq)
430
431	<	call check_initialization(localError)
431	>	call doReadyCheck(localError)
432		if ( localError .ne. 0 ) then
433	+	call handleError("do_force_loop", "Not Initialized")
434		error = -1
435		return
436		end if
437		call zero_work_arrays()
438	<
438	>
439		do_pot = do_pot_c
440		do_stress = do_stress_c
441	<
441	>
442		! Gather all information needed by all force loops:
443
444		#ifdef IS_MPI
445	+
446	+	call gather(q, q_Row, plan_row3d)
447	+	call gather(q, q_Col, plan_col3d)
448
449	<	call gather(q,q_Row,plan_row3d)
450	<	call gather(q,q_Col,plan_col3d)
449	>	call gather(q_group, q_group_Row, plan_row_Group_3d)
450	>	call gather(q_group, q_group_Col, plan_col_Group_3d)
451
452	<	if (FF_UsesDirectionalAtoms() .and. SimUsesDirectionalAtoms()) then
452	>	if (FF_UsesDirectionalAtoms() .and. SIM_uses_directional_atoms) then
453		call gather(u_l,u_l_Row,plan_row3d)
454		call gather(u_l,u_l_Col,plan_col3d)
455
#	Line 227 | Line 459 | contains
459
460		#endif
461
462	<	if (FF_RequiresPrepairCalc() .and. SimRequiresPrepairCalc()) then
462	>	!! Begin force loop timing:
463	>	#ifdef PROFILE
464	>	call cpu_time(forceTimeInitial)
465	>	nloops = nloops + 1
466	>	#endif
467	>
468	>	if (FF_RequiresPrepairCalc() .and. SIM_requires_prepair_calc) then
469		!! See if we need to update neighbor lists
470	<	call checkNeighborList(nlocal, q, rcut, rlist, update_nlist)
470	>
471	>	call checkNeighborList(nGroup, q_group, listSkin, update_nlist)
472	>
473		!! if_mpi_gather_stuff_for_prepair
474		!! do_prepair_loop_if_needed
475		!! if_mpi_scatter_stuff_from_prepair
476		!! if_mpi_gather_stuff_from_prepair_to_main_loop
477	<	else
478	<	!! See if we need to update neighbor lists
479	<	call checkNeighborList(nlocal, q, rcut, rlist, update_nlist)
480	<	endif
481	<
477	>
478	>	!--------------------PREFORCE LOOP----------->>>>>>>>>>>>>>>>>>>>>>>>>>>
479	>
480	>	if (update_nlist) then
481	>
482	>	!! save current configuration, construct neighbor list,
483	>	!! and calculate forces
484	>
485	>	call saveNeighborList(nGroup, q_group)
486	>
487	>	neighborListSize = size(list)
488	>	nlist = 0
489	>
490	>	istart = 1
491		#ifdef IS_MPI
492	+	iend = nrow_group
493	+	#else
494	+	iend = nGroup - 1
495	+	#endif
496	+	do i = istart, iend
497	+
498	+	point(i) = nlist + 1
499	+
500	+	n_in_i = groupStart(i+1) - groupStart(i)
501	+
502	+	#ifdef IS_MPI
503	+	jstart = 1
504	+	jend = ncol_group
505	+	#else
506	+	jstart = i+1
507	+	jend = nGroup
508	+	#endif
509	+	do j = jstart, jend
510	+
511	+	#ifdef IS_MPI
512	+	call get_interatomic_vector(q_group_Row(:,i), &
513	+	q_group_Col(:,j), d_grp, rgrpsq)
514	+	#else
515	+	call get_interatomic_vector(q_group(:,i), &
516	+	q_group(:,j), d_grp, rgrpsq)
517	+	#endif
518	+	if (rgrpsq < rlistsq) then
519	+	nlist = nlist + 1
520	+
521	+	if (nlist > neighborListSize) then
522	+	call expandNeighborList(nGroup, listerror)
523	+	if (listerror /= 0) then
524	+	error = -1
525	+	write(DEFAULT_ERROR,*) "ERROR: nlist > list size and max allocations exceeded."
526	+	return
527	+	end if
528	+	neighborListSize = size(list)
529	+	endif
530	+
531	+	list(nlist) = j
532	+
533	+	call get_switch(rgrpsq, sw, dswdr, rgrp, group_switch, &
534	+	in_switching_region)
535	+
536	+	n_in_j = groupStart(j+1) - groupStart(j)
537	+
538	+	do ia = groupStart(i), groupStart(i+1)-1
539	+	atom1 = groupList(ia)
540	+
541	+	prepair_inner1: do jb = groupStart(j), groupStart(j+1)-1
542	+	atom2 = groupList(jb)
543	+
544	+	if (skipThisPair(atom1, atom2)) cycle prepair_inner1
545	+
546	+	if ((n_in_i .eq. 1).and.(n_in_j .eq. 1)) then
547	+	d_atm(1:3) = d_grp(1:3)
548	+	ratmsq = rgrpsq
549	+	else
550	+	#ifdef IS_MPI
551	+	call get_interatomic_vector(q_Row(:,atom1), &
552	+	q_Col(:,atom2), d_atm, ratmsq)
553	+	#else
554	+	call get_interatomic_vector(q(:,atom1), &
555	+	q(:,atom2), d_atm, ratmsq)
556	+	#endif
557	+	endif
558	+	#ifdef IS_MPI
559	+	call do_prepair(atom1, atom2, ratmsq, d_atm, sw, &
560	+	rgrpsq, d_grp, do_pot, do_stress, &
561	+	u_l, A, f, t, pot_local)
562	+	#else
563	+	call do_prepair(atom1, atom2, ratmsq, d_atm, sw, &
564	+	rgrpsq, d_grp, do_pot, do_stress, &
565	+	u_l, A, f, t, pot)
566	+	#endif
567	+	enddo prepair_inner1
568	+	enddo
569	+
570	+	end if
571	+	enddo
572	+	enddo
573	+
574	+	#ifdef IS_MPI
575	+	point(nrow_group + 1) = nlist + 1
576	+	#else
577	+	point(nGroup) = nlist + 1
578	+	#endif
579	+
580	+	else  !! (of update_check)
581	+
582	+	! use the list to find the neighbors
583	+
584	+	istart = 1
585	+	#ifdef IS_MPI
586	+	iend = nrow_group
587	+	#else
588	+	iend = nGroup - 1
589	+	#endif
590	+
591	+	do i = istart, iend
592	+
593	+	n_in_i = groupStart(i+1) - groupStart(i)
594	+
595	+	JBEG = POINT(i)
596	+	JEND = POINT(i+1) - 1
597	+	! check that group i has neighbors
598	+	if (jbeg .le. jend) then
599	+
600	+	do jnab = jbeg, jend
601	+	j = list(jnab)
602	+
603	+	#ifdef IS_MPI
604	+	call get_interatomic_vector(q_group_Row(:,i), &
605	+	q_group_Col(:,j), d_grp, rgrpsq)
606	+	#else
607	+	call get_interatomic_vector(q_group(:,i), &
608	+	q_group(:,j), d_grp, rgrpsq)
609	+	#endif
610	+	call get_switch(rgrpsq, sw, dswdr, rgrp, group_switch, &
611	+	in_switching_region)
612	+
613	+	n_in_j = groupStart(j+1) - groupStart(j)
614	+
615	+	do ia = groupStart(i), groupStart(i+1)-1
616	+	atom1 = groupList(ia)
617	+
618	+	prepair_inner2: do jb = groupStart(j), groupStart(j+1)-1
619	+
620	+	atom2 = groupList(jb)
621	+
622	+	if (skipThisPair(atom1, atom2)) cycle prepair_inner2
623	+
624	+	if ((n_in_i .eq. 1).and.(n_in_j .eq. 1)) then
625	+	d_atm(1:3) = d_grp(1:3)
626	+	ratmsq = rgrpsq
627	+	else
628	+	#ifdef IS_MPI
629	+	call get_interatomic_vector(q_Row(:,atom1), &
630	+	q_Col(:,atom2), d_atm, ratmsq)
631	+	#else
632	+	call get_interatomic_vector(q(:,atom1), &
633	+	q(:,atom2), d_atm, ratmsq)
634	+	#endif
635	+	endif
636	+
637	+	#ifdef IS_MPI
638	+	call do_prepair(atom1, atom2, ratmsq, d_atm, sw, &
639	+	rgrpsq, d_grp, do_pot, do_stress, &
640	+	u_l, A, f, t, pot_local)
641	+	#else
642	+	call do_prepair(atom1, atom2, ratmsq, d_atm, sw, &
643	+	rgrpsq, d_grp, do_pot, do_stress, &
644	+	u_l, A, f, t, pot)
645	+	#endif
646	+	enddo prepair_inner2
647	+	enddo
648	+	enddo
649	+	endif
650	+	enddo
651	+	endif
652	+	!! Do rest of preforce calculations
653	+	!! do necessary preforce calculations
654	+	call do_preforce(nlocal,pot)
655	+	! we have already updated the neighbor list set it to false...
656	+	update_nlist = .false.
657	+	else
658	+	!! See if we need to update neighbor lists for non pre-pair
659	+	call checkNeighborList(nGroup, q_group, listSkin, update_nlist)
660	+	end if
661
662	+	!---------------------------------MAIN Pair LOOP->>>>>>>>>>>>>
663	+
664		if (update_nlist) then
665
666		!! save current configuration, construct neighbor list,
667		!! and calculate forces
248	–	call saveNeighborList(q)
668
669	+	call saveNeighborList(nGroup, q_group)
670	+
671		neighborListSize = size(list)
672		nlist = 0
673	<
674	<	do i = 1, nrow
673	>
674	>	istart = 1
675	>	#ifdef IS_MPI
676	>	iend = nrow_group
677	>	#else
678	>	iend = nGroup - 1
679	>	#endif
680	>	do i = istart, iend
681	>
682		point(i) = nlist + 1
683	+
684	+	n_in_i = groupStart(i+1) - groupStart(i)
685
686	<	inner: do j = 1, ncol
686	>	#ifdef IS_MPI
687	>	jstart = 1
688	>	jend = ncol_group
689	>	#else
690	>	jstart = i+1
691	>	jend = nGroup
692	>	#endif
693	>	do j = jstart, jend
694
695	<	if (skipThisPair(i,j)) cycle inner
696	<
697	<	call get_interatomic_vector(q_Row(:,i), q_Col(:,j), d, rijsq)
698	<
699	<	if (rijsq <  rlistsq) then
700	<
695	>	#ifdef IS_MPI
696	>	call get_interatomic_vector(q_group_Row(:,i), &
697	>	q_group_Col(:,j), d_grp, rgrpsq)
698	>	#else
699	>	call get_interatomic_vector(q_group(:,i), &
700	>	q_group(:,j), d_grp, rgrpsq)
701	>	#endif
702	>	if (rgrpsq < rlistsq) then
703		nlist = nlist + 1
704
705		if (nlist > neighborListSize) then
706	<	call expandNeighborList(nlocal, listerror)
706	>	call expandNeighborList(nGroup, listerror)
707		if (listerror /= 0) then
708		error = -1
709		write(DEFAULT_ERROR,*) "ERROR: nlist > list size and max allocations exceeded."
#	Line 274 | Line 713 | contains
713		endif
714
715		list(nlist) = j
716	<
717	<	if (rijsq <  rcutsq) then
718	<	call do_pair(i, j, rijsq, d, do_pot, do_stress, &
719	<	u_l, A, f, t,pot)
716	>
717	>	vij = 0.0d0
718	>	fij(1:3) = 0.0d0
719	>
720	>	call get_switch(rgrpsq, sw, dswdr, rgrp, group_switch, &
721	>	in_switching_region)
722	>
723	>	n_in_j = groupStart(j+1) - groupStart(j)
724	>
725	>	do ia = groupStart(i), groupStart(i+1)-1
726	>	atom1 = groupList(ia)
727	>
728	>	inner1: do jb = groupStart(j), groupStart(j+1)-1
729	>	atom2 = groupList(jb)
730	>
731	>	if (skipThisPair(atom1, atom2)) cycle inner1
732	>
733	>	if ((n_in_i .eq. 1).and.(n_in_j .eq. 1)) then
734	>	d_atm(1:3) = d_grp(1:3)
735	>	ratmsq = rgrpsq
736	>	else
737	>	#ifdef IS_MPI
738	>	call get_interatomic_vector(q_Row(:,atom1), &
739	>	q_Col(:,atom2), d_atm, ratmsq)
740	>	#else
741	>	call get_interatomic_vector(q(:,atom1), &
742	>	q(:,atom2), d_atm, ratmsq)
743	>	#endif
744	>	endif
745	>	#ifdef IS_MPI
746	>	call do_pair(atom1, atom2, ratmsq, d_atm, sw, &
747	>	do_pot, &
748	>	u_l, A, f, t, pot_local, vpair, fpair)
749	>	#else
750	>	call do_pair(atom1, atom2, ratmsq, d_atm, sw, &
751	>	do_pot,  &
752	>	u_l, A, f, t, pot, vpair, fpair)
753	>	#endif
754	>	vij = vij + vpair
755	>	fij(1:3) = fij(1:3) + fpair(1:3)
756	>
757	>	enddo inner1
758	>	enddo
759	>
760	>	if (in_switching_region) then
761	>	swderiv = vij*dswdr/rgrp
762	>	fij(1) = fij(1) + swderiv*d_grp(1)
763	>	fij(2) = fij(2) + swderiv*d_grp(2)
764	>	fij(3) = fij(3) + swderiv*d_grp(3)
765	>
766	>	do ia=groupStart(i), groupStart(i+1)-1
767	>	atom1=groupList(ia)
768	>	mf = mfact(atom1)
769	>	#ifdef IS_MPI
770	>	f_Row(1,atom1) = f_Row(1,atom1) + swderiv*d_grp(1)*mf
771	>	f_Row(2,atom1) = f_Row(2,atom1) + swderiv*d_grp(2)*mf
772	>	f_Row(3,atom1) = f_Row(3,atom1) + swderiv*d_grp(3)*mf
773	>	#else
774	>	f(1,atom1) = f(1,atom1) + swderiv*d_grp(1)*mf
775	>	f(2,atom1) = f(2,atom1) + swderiv*d_grp(2)*mf
776	>	f(3,atom1) = f(3,atom1) + swderiv*d_grp(3)*mf
777	>	#endif
778	>	enddo
779	>
780	>	do jb=groupStart(j), groupStart(j+1)-1
781	>	atom2=groupList(jb)
782	>	mf = mfact(atom2)
783	>	#ifdef IS_MPI
784	>	f_Col(1,atom2) = f_Col(1,atom2) - swderiv*d_grp(1)*mf
785	>	f_Col(2,atom2) = f_Col(2,atom2) - swderiv*d_grp(2)*mf
786	>	f_Col(3,atom2) = f_Col(3,atom2) - swderiv*d_grp(3)*mf
787	>	#else
788	>	f(1,atom2) = f(1,atom2) - swderiv*d_grp(1)*mf
789	>	f(2,atom2) = f(2,atom2) - swderiv*d_grp(2)*mf
790	>	f(3,atom2) = f(3,atom2) - swderiv*d_grp(3)*mf
791	>	#endif
792	>	enddo
793		endif
794	<	endif
795	<	enddo inner
794	>
795	>	if (do_stress) call add_stress_tensor(d_grp, fij)
796	>
797	>	end if
798	>	enddo
799		enddo
800
801	<	point(nrow + 1) = nlist + 1
801	>	#ifdef IS_MPI
802	>	point(nrow_group + 1) = nlist + 1
803	>	#else
804	>	point(nGroup) = nlist + 1
805	>	#endif
806
807		else  !! (of update_check)
808	<
808	>
809		! use the list to find the neighbors
810	<	do i = 1, nrow
810	>
811	>	istart = 1
812	>	#ifdef IS_MPI
813	>	iend = nrow_group
814	>	#else
815	>	iend = nGroup - 1
816	>	#endif
817	>
818	>	do i = istart, iend
819	>
820	>	n_in_i = groupStart(i+1) - groupStart(i)
821	>
822		JBEG = POINT(i)
823		JEND = POINT(i+1) - 1
824	<	! check thiat molecule i has neighbors
824	>	! check that group i has neighbors
825		if (jbeg .le. jend) then
826
827		do jnab = jbeg, jend
828		j = list(jnab)
829	<
830	<	call get_interatomic_vector(q_Row(:,i), q_Col(:,j), d, rijsq)
831	<	call do_pair(i, j, rijsq, d, do_pot, do_stress, &
832	<	u_l, A, f, t,pot)
303	<
304	<	enddo
305	<	endif
306	<	enddo
307	<	endif
308	<
829	>
830	>	#ifdef IS_MPI
831	>	call get_interatomic_vector(q_group_Row(:,i), &
832	>	q_group_Col(:,j), d_grp, rgrpsq)
833		#else
834	<
835	<	if (update_nlist) then
836	<
837	<	! save current configuration, contruct neighbor list,
838	<	! and calculate forces
315	<	call saveNeighborList(q)
316	<
317	<	neighborListSize = size(list)
318	<
319	<	nlist = 0
320	<
321	<	do i = 1, natoms-1
322	<	point(i) = nlist + 1
323	<
324	<	inner: do j = i+1, natoms
325	<
326	<	if (skipThisPair(i,j))  cycle inner
327	<
328	<	call get_interatomic_vector(q(:,i), q(:,j), d, rijsq)
329	<
834	>	call get_interatomic_vector(q_group(:,i), &
835	>	q_group(:,j), d_grp, rgrpsq)
836	>	#endif
837	>	vij = 0.0d0
838	>	fij(1:3) = 0.0d0
839
840	<	if (rijsq <  rlistsq) then
840	>	call get_switch(rgrpsq, sw, dswdr, rgrp, group_switch, &
841	>	in_switching_region)
842
843	<	nlist = nlist + 1
334	<
335	<	if (nlist > neighborListSize) then
336	<	call expandNeighborList(natoms, listerror)
337	<	if (listerror /= 0) then
338	<	error = -1
339	<	write(DEFAULT_ERROR,*) "ERROR: nlist > list size and max allocations exceeded."
340	<	return
341	<	end if
342	<	neighborListSize = size(list)
343	<	endif
843	>	n_in_j = groupStart(j+1) - groupStart(j)
844
845	<	list(nlist) = j
845	>	do ia = groupStart(i), groupStart(i+1)-1
846	>	atom1 = groupList(ia)
847	>
848	>	inner2: do jb = groupStart(j), groupStart(j+1)-1
849	>
850	>	atom2 = groupList(jb)
851	>
852	>	if (skipThisPair(atom1, atom2)) cycle inner2
853	>
854	>	if ((n_in_i .eq. 1).and.(n_in_j .eq. 1)) then
855	>	d_atm(1:3) = d_grp(1:3)
856	>	ratmsq = rgrpsq
857	>	else
858	>	#ifdef IS_MPI
859	>	call get_interatomic_vector(q_Row(:,atom1), &
860	>	q_Col(:,atom2), d_atm, ratmsq)
861	>	#else
862	>	call get_interatomic_vector(q(:,atom1), &
863	>	q(:,atom2), d_atm, ratmsq)
864	>	#endif
865	>	endif
866	>	#ifdef IS_MPI
867	>	call do_pair(atom1, atom2, ratmsq, d_atm, sw, &
868	>	do_pot,  &
869	>	u_l, A, f, t, pot_local, vpair, fpair)
870	>	#else
871	>	call do_pair(atom1, atom2, ratmsq, d_atm, sw, &
872	>	do_pot,  &
873	>	u_l, A, f, t, pot, vpair, fpair)
874	>	#endif
875	>	vij = vij + vpair
876	>	fij(1:3) = fij(1:3) + fpair(1:3)
877	>
878	>	enddo inner2
879	>	enddo
880
881	<	if (rijsq <  rcutsq) then
882	<	call do_pair(i, j, rijsq, d, do_pot, do_stress, &
883	<	u_l, A, f, t,pot)
881	>	if (in_switching_region) then
882	>	swderiv = vij*dswdr/rgrp
883	>	fij(1) = fij(1) + swderiv*d_grp(1)
884	>	fij(2) = fij(2) + swderiv*d_grp(2)
885	>	fij(3) = fij(3) + swderiv*d_grp(3)
886	>
887	>	do ia=groupStart(i), groupStart(i+1)-1
888	>	atom1=groupList(ia)
889	>	mf = mfact(atom1)
890	>	#ifdef IS_MPI
891	>	f_Row(1,atom1) = f_Row(1,atom1) + swderiv*d_grp(1)*mf
892	>	f_Row(2,atom1) = f_Row(2,atom1) + swderiv*d_grp(2)*mf
893	>	f_Row(3,atom1) = f_Row(3,atom1) + swderiv*d_grp(3)*mf
894	>	#else
895	>	f(1,atom1) = f(1,atom1) + swderiv*d_grp(1)*mf
896	>	f(2,atom1) = f(2,atom1) + swderiv*d_grp(2)*mf
897	>	f(3,atom1) = f(3,atom1) + swderiv*d_grp(3)*mf
898	>	#endif
899	>	enddo
900	>
901	>	do jb=groupStart(j), groupStart(j+1)-1
902	>	atom2=groupList(jb)
903	>	mf = mfact(atom2)
904	>	#ifdef IS_MPI
905	>	f_Col(1,atom2) = f_Col(1,atom2) - swderiv*d_grp(1)*mf
906	>	f_Col(2,atom2) = f_Col(2,atom2) - swderiv*d_grp(2)*mf
907	>	f_Col(3,atom2) = f_Col(3,atom2) - swderiv*d_grp(3)*mf
908	>	#else
909	>	f(1,atom2) = f(1,atom2) - swderiv*d_grp(1)*mf
910	>	f(2,atom2) = f(2,atom2) - swderiv*d_grp(2)*mf
911	>	f(3,atom2) = f(3,atom2) - swderiv*d_grp(3)*mf
912	>	#endif
913	>	enddo
914		endif
351	–	endif
352	–	enddo inner
353	–	enddo
354	–
355	–	point(natoms) = nlist + 1
356	–
357	–	else !! (update)
358	–
359	–	! use the list to find the neighbors
360	–	do i = 1, natoms-1
361	–	JBEG = POINT(i)
362	–	JEND = POINT(i+1) - 1
363	–	! check thiat molecule i has neighbors
364	–	if (jbeg .le. jend) then
365	–
366	–	do jnab = jbeg, jend
367	–	j = list(jnab)
915
916	<	call get_interatomic_vector(q(:,i), q(:,j), d, rijsq)
370	<	call do_pair(i, j, rijsq, d, do_pot, do_stress, &
371	<	u_l, A, f, t,pot)
916	>	if (do_stress) call add_stress_tensor(d_grp, fij)
917
918		enddo
919		endif
920		enddo
921		endif
922
378	–	#endif
379	–
923		! phew, done with main loop.
924
925	+	!! Do timing
926	+	#ifdef PROFILE
927	+	call cpu_time(forceTimeFinal)
928	+	forceTime = forceTime + forceTimeFinal - forceTimeInitial
929	+	#endif
930	+
931		#ifdef IS_MPI
932		!!distribute forces
933
934	<	call scatter(f_Row,f,plan_row3d)
934	>	f_temp = 0.0_dp
935	>	call scatter(f_Row,f_temp,plan_row3d)
936	>	do i = 1,nlocal
937	>	f(1:3,i) = f(1:3,i) + f_temp(1:3,i)
938	>	end do
939	>
940	>	f_temp = 0.0_dp
941		call scatter(f_Col,f_temp,plan_col3d)
942		do i = 1,nlocal
943		f(1:3,i) = f(1:3,i) + f_temp(1:3,i)
944		end do
945
946	<	if (FF_UsesDirectionalAtoms() .and. SimUsesDirectionalAtoms()) then
947	<	call scatter(t_Row,t,plan_row3d)
946	>	if (FF_UsesDirectionalAtoms() .and. SIM_uses_directional_atoms) then
947	>	t_temp = 0.0_dp
948	>	call scatter(t_Row,t_temp,plan_row3d)
949	>	do i = 1,nlocal
950	>	t(1:3,i) = t(1:3,i) + t_temp(1:3,i)
951	>	end do
952	>	t_temp = 0.0_dp
953		call scatter(t_Col,t_temp,plan_col3d)
954
955		do i = 1,nlocal
#	Line 406 | Line 966 | contains
966		do i = 1, nlocal
967		pot_local = pot_local + pot_Temp(i)
968		enddo
969	<
970	<	pot_Temp = 0.0_DP
971	<
969	>
970	>	pot_Temp = 0.0_DP
971	>
972		call scatter(pot_Col, pot_Temp, plan_col)
973		do i = 1, nlocal
974		pot_local = pot_local + pot_Temp(i)
975		enddo
976
977	<	endif
977	>	endif
978		#endif
979	<
980	<	if (FF_RequiresPostpairCalc() .and. SimRequiresPostpairCalc()) then
979	>
980	>	if (FF_RequiresPostpairCalc() .and. SIM_requires_postpair_calc) then
981
982	<	if (FF_uses_RF .and. SimUsesRF()) then
982	>	if (FF_uses_RF .and. SIM_uses_RF) then
983
984		#ifdef IS_MPI
985		call scatter(rf_Row,rf,plan_row3d)
#	Line 429 | Line 989 | contains
989		end do
990		#endif
991
992	<	do i = 1, getNlocal()
993	<
992	>	do i = 1, nLocal
993	>
994		rfpot = 0.0_DP
995		#ifdef IS_MPI
996		me_i = atid_row(i)
997		#else
998		me_i = atid(i)
999		#endif
1000	<	call getElementProperty(atypes, me_i, "is_DP", is_DP_i)
1001	<	if ( is_DP_i ) then
1002	<	call getElementProperty(atypes, me_i, "dipole_moment", mu_i)
1000	>
1001	>	if (PropertyMap(me_i)%is_DP) then
1002	>
1003	>	mu_i = PropertyMap(me_i)%dipole_moment
1004	>
1005		!! The reaction field needs to include a self contribution
1006		!! to the field:
1007	<	call accumulate_self_rf(i, mu_i, u_l)
1007	>	call accumulate_self_rf(i, mu_i, u_l)
1008		!! Get the reaction field contribution to the
1009		!! potential and torques:
1010		call reaction_field_final(i, mu_i, u_l, rfpot, t, do_pot)
#	Line 456 | Line 1018 | contains
1018		enddo
1019		endif
1020		endif
1021	<
1022	<
1021	>
1022	>
1023		#ifdef IS_MPI
1024	<
1024	>
1025		if (do_pot) then
1026	<	pot = pot_local
1026	>	pot = pot + pot_local
1027		!! we assume the c code will do the allreduce to get the total potential
1028		!! we could do it right here if we needed to...
1029		endif
1030	<
1030	>
1031		if (do_stress) then
1032	<	call mpi_allreduce(tau, tau_Temp,9,mpi_double_precision,mpi_sum, &
1032	>	call mpi_allreduce(tau_Temp, tau, 9,mpi_double_precision,mpi_sum, &
1033		mpi_comm_world,mpi_err)
1034	<	call mpi_allreduce(virial, virial_Temp,1,mpi_double_precision,mpi_sum, &
1034	>	call mpi_allreduce(virial_Temp, virial,1,mpi_double_precision,mpi_sum, &
1035		mpi_comm_world,mpi_err)
1036		endif
1037	<
1037	>
1038		#else
1039	<
1039	>
1040		if (do_stress) then
1041		tau = tau_Temp
1042		virial = virial_Temp
1043		endif
1044	<
1044	>
1045		#endif
1046
1047	+
1048		end subroutine do_force_loop
1049
1050	<	subroutine do_pair(i, j, rijsq, d, do_pot, do_stress, u_l, A, f, t,pot)
1050	>
1051	>	subroutine do_pair(i, j, rijsq, d, sw, do_pot, &
1052	>	u_l, A, f, t, pot, vpair, fpair)
1053
1054	<	real( kind = dp ) :: pot
1055	<	real( kind = dp ), dimension(:,:) :: u_l
1056	<	real (kind=dp), dimension(:,:) :: A
1057	<	real (kind=dp), dimension(:,:) :: f
1058	<	real (kind=dp), dimension(:,:) :: t
1054	>	real( kind = dp ) :: pot, vpair, sw
1055	>	real( kind = dp ), dimension(3) :: fpair
1056	>	real( kind = dp ), dimension(nLocal)   :: mfact
1057	>	real( kind = dp ), dimension(3,nLocal) :: u_l
1058	>	real( kind = dp ), dimension(9,nLocal) :: A
1059	>	real( kind = dp ), dimension(3,nLocal) :: f
1060	>	real( kind = dp ), dimension(3,nLocal) :: t
1061
1062	<	logical, intent(inout) :: do_pot, do_stress
1062	>	logical, intent(inout) :: do_pot
1063		integer, intent(in) :: i, j
1064	<	real ( kind = dp ), intent(inout)    :: rijsq
1064	>	real ( kind = dp ), intent(inout) :: rijsq
1065		real ( kind = dp )                :: r
1066		real ( kind = dp ), intent(inout) :: d(3)
500	–	logical :: is_LJ_i, is_LJ_j
501	–	logical :: is_DP_i, is_DP_j
502	–	logical :: is_GB_i, is_GB_j
503	–	logical :: is_Sticky_i, is_Sticky_j
1067		integer :: me_i, me_j
1068
1069		r = sqrt(rijsq)
1070	+	vpair = 0.0d0
1071	+	fpair(1:3) = 0.0d0
1072
1073		#ifdef IS_MPI
1074	<
1074	>	if (tagRow(i) .eq. tagColumn(j)) then
1075	>	write(0,*) 'do_pair is doing', i , j, tagRow(i), tagColumn(j)
1076	>	endif
1077		me_i = atid_row(i)
1078		me_j = atid_col(j)
512	–
1079		#else
514	–
1080		me_i = atid(i)
1081		me_j = atid(j)
517	–
1082		#endif
1083	+
1084	+	if (FF_uses_LJ .and. SIM_uses_LJ) then
1085	+
1086	+	if ( PropertyMap(me_i)%is_LJ .and. PropertyMap(me_j)%is_LJ ) then
1087	+	!write(*,*) 'calling lj with'
1088	+	!write(*,*) i, j, r, rijsq
1089	+	!write(*,'(3es12.3)') d(1), d(2), d(3)
1090	+	!write(*,'(3es12.3)') sw, vpair, pot
1091	+	!write(*,*)
1092
1093	<	if (FF_uses_LJ .and. SimUsesLJ()) then
1094	<	call getElementProperty(atypes, me_i, "is_LJ", is_LJ_i)
1095	<	call getElementProperty(atypes, me_j, "is_LJ", is_LJ_j)
523	<
524	<	if ( is_LJ_i .and. is_LJ_j ) &
525	<	call do_lj_pair(i, j, d, r, rijsq, pot, f, do_pot, do_stress)
1093	>	call do_lj_pair(i, j, d, r, rijsq, sw, vpair, fpair, pot, f, do_pot)
1094	>	endif
1095	>
1096		endif
1097	<
1098	<	if (FF_uses_dipoles .and. SimUsesDipoles()) then
529	<	call getElementProperty(atypes, me_i, "is_DP", is_DP_i)
530	<	call getElementProperty(atypes, me_j, "is_DP", is_DP_j)
1097	>
1098	>	if (FF_uses_charges .and. SIM_uses_charges) then
1099
1100	<	if ( is_DP_i .and. is_DP_j ) then
1101	<
534	<	call do_dipole_pair(i, j, d, r, rijsq, pot, u_l, f, t, &
535	<	do_pot, do_stress)
536	<	if (FF_uses_RF .and. SimUsesRF()) then
537	<	call accumulate_rf(i, j, r, u_l)
538	<	call rf_correct_forces(i, j, d, r, u_l, f, do_stress)
539	<	endif
540	<
1100	>	if (PropertyMap(me_i)%is_Charge .and. PropertyMap(me_j)%is_Charge) then
1101	>	call do_charge_pair(i, j, d, r, rijsq, sw, vpair, fpair, pot, f, do_pot)
1102		endif
1103	+
1104		endif
1105	+
1106	+	if (FF_uses_dipoles .and. SIM_uses_dipoles) then
1107	+
1108	+	if ( PropertyMap(me_i)%is_DP .and. PropertyMap(me_j)%is_DP) then
1109	+	call do_dipole_pair(i, j, d, r, rijsq, sw, vpair, fpair, pot, u_l, f, t, &
1110	+	do_pot)
1111	+	if (FF_uses_RF .and. SIM_uses_RF) then
1112	+	call accumulate_rf(i, j, r, u_l, sw)
1113	+	call rf_correct_forces(i, j, d, r, u_l, sw, f, fpair)
1114	+	endif
1115	+	endif
1116
1117	<	if (FF_uses_Sticky .and. SimUsesSticky()) then
1117	>	endif
1118
1119	<	call getElementProperty(atypes, me_i, "is_Sticky", is_Sticky_i)
547	<	call getElementProperty(atypes, me_j, "is_Sticky", is_Sticky_j)
1119	>	if (FF_uses_Sticky .and. SIM_uses_sticky) then
1120
1121	<	if ( is_Sticky_i .and. is_Sticky_j ) then
1122	<	call do_sticky_pair(i, j, d, r, rijsq, A, pot, f, t, &
1123	<	do_pot, do_stress)
1121	>	if ( PropertyMap(me_i)%is_Sticky .and. PropertyMap(me_j)%is_Sticky) then
1122	>	call do_sticky_pair(i, j, d, r, rijsq, sw, vpair, fpair, pot, A, f, t, &
1123	>	do_pot)
1124		endif
1125	+
1126		endif
1127
1128
1129	<	if (FF_uses_GB .and. SimUsesGB()) then
1129	>	if (FF_uses_GB .and. SIM_uses_GB) then
1130	>
1131	>	if ( PropertyMap(me_i)%is_GB .and. PropertyMap(me_j)%is_GB) then
1132	>	call do_gb_pair(i, j, d, r, rijsq, sw, vpair, fpair, pot, u_l, f, t, &
1133	>	do_pot)
1134	>	endif
1135
1136	<	call getElementProperty(atypes, me_i, "is_GB", is_GB_i)
559	<	call getElementProperty(atypes, me_j, "is_GB", is_GB_j)
1136	>	endif
1137
1138	<	if ( is_GB_i .and. is_GB_j ) then
1139	<	call do_gb_pair(i, j, d, r, rijsq, u_l, pot, f, t, &
1140	<	do_pot, do_stress)
1138	>	if (FF_uses_EAM .and. SIM_uses_EAM) then
1139	>
1140	>	if ( PropertyMap(me_i)%is_EAM .and. PropertyMap(me_j)%is_EAM) then
1141	>	call do_eam_pair(i, j, d, r, rijsq, sw, vpair, fpair, pot, f, &
1142	>	do_pot)
1143		endif
1144	+
1145		endif
1146	<
1146	>
1147		end subroutine do_pair
1148
1149	+	subroutine do_prepair(i, j, rijsq, d, sw, rcijsq, dc, &
1150	+	do_pot, do_stress, u_l, A, f, t, pot)
1151
1152	<	subroutine get_interatomic_vector(q_i, q_j, d, r_sq)
1153	<
1154	<	real (kind = dp), dimension(3) :: q_i
1155	<	real (kind = dp), dimension(3) :: q_j
1156	<	real ( kind = dp ), intent(out) :: r_sq
1157	<	real( kind = dp ) :: d(3)
1158	<	real( kind = dp ) :: d_old(3)
1159	<	d(1:3) = q_i(1:3) - q_j(1:3)
1160	<	d_old = d
1161	<	! Wrap back into periodic box if necessary
1162	<	if ( SimUsesPBC() ) then
1152	>	real( kind = dp ) :: pot, sw
1153	>	real( kind = dp ), dimension(3,nLocal) :: u_l
1154	>	real (kind=dp), dimension(9,nLocal) :: A
1155	>	real (kind=dp), dimension(3,nLocal) :: f
1156	>	real (kind=dp), dimension(3,nLocal) :: t
1157	>
1158	>	logical, intent(inout) :: do_pot, do_stress
1159	>	integer, intent(in) :: i, j
1160	>	real ( kind = dp ), intent(inout)    :: rijsq, rcijsq
1161	>	real ( kind = dp )                :: r, rc
1162	>	real ( kind = dp ), intent(inout) :: d(3), dc(3)
1163	>
1164	>	logical :: is_EAM_i, is_EAM_j
1165	>
1166	>	integer :: me_i, me_j
1167	>
1168
1169	<	d(1:3) = d(1:3) - box(1:3) * sign(1.0_dp,d(1:3)) * &
1170	<	int(abs(d(1:3)/box(1:3)) + 0.5_dp)
1171	<
1169	>	r = sqrt(rijsq)
1170	>	if (SIM_uses_molecular_cutoffs) then
1171	>	rc = sqrt(rcijsq)
1172	>	else
1173	>	rc = r
1174		endif
1175	<	r_sq = dot_product(d,d)
587	<
588	<	end subroutine get_interatomic_vector
1175	>
1176
590	–	subroutine check_initialization(error)
591	–	integer, intent(out) :: error
592	–
593	–	error = 0
594	–	! Make sure we are properly initialized.
595	–	if (.not. do_forces_initialized) then
596	–	error = -1
597	–	return
598	–	endif
599	–
1177		#ifdef IS_MPI
1178	<	if (.not. isMPISimSet()) then
1179	<	write(default_error,*) "ERROR: mpiSimulation has not been initialized!"
1180	<	error = -1
1181	<	return
1182	<	endif
1178	>	if (tagRow(i) .eq. tagColumn(j)) then
1179	>	write(0,*) 'do_prepair is doing', i , j, tagRow(i), tagColumn(j)
1180	>	endif
1181	>
1182	>	me_i = atid_row(i)
1183	>	me_j = atid_col(j)
1184	>
1185	>	#else
1186	>
1187	>	me_i = atid(i)
1188	>	me_j = atid(j)
1189	>
1190		#endif
1191	<
1192	<	return
609	<	end subroutine check_initialization
610	<
611	<
612	<	subroutine zero_work_arrays()
613	<
614	<	#ifdef IS_MPI
615	<
616	<	q_Row = 0.0_dp
617	<	q_Col = 0.0_dp
618	<
619	<	u_l_Row = 0.0_dp
620	<	u_l_Col = 0.0_dp
621	<
622	<	A_Row = 0.0_dp
623	<	A_Col = 0.0_dp
624	<
625	<	f_Row = 0.0_dp
626	<	f_Col = 0.0_dp
627	<	f_Temp = 0.0_dp
1191	>
1192	>	if (FF_uses_EAM .and. SIM_uses_EAM) then
1193
1194	<	t_Row = 0.0_dp
1195	<	t_Col = 0.0_dp
1196	<	t_Temp = 0.0_dp
1194	>	if (PropertyMap(me_i)%is_EAM .and. PropertyMap(me_j)%is_EAM) &
1195	>	call calc_EAM_prepair_rho(i, j, d, r, rijsq )
1196	>
1197	>	endif
1198	>
1199	>	end subroutine do_prepair
1200
1201	<	pot_Row = 0.0_dp
1202	<	pot_Col = 0.0_dp
1203	<	pot_Temp = 0.0_dp
1204	<
1205	<	rf_Row = 0.0_dp
1206	<	rf_Col = 0.0_dp
1207	<	rf_Temp = 0.0_dp
1201	>
1202	>	subroutine do_preforce(nlocal,pot)
1203	>	integer :: nlocal
1204	>	real( kind = dp ) :: pot
1205	>
1206	>	if (FF_uses_EAM .and. SIM_uses_EAM) then
1207	>	call calc_EAM_preforce_Frho(nlocal,pot)
1208	>	endif
1209	>
1210	>
1211	>	end subroutine do_preforce
1212	>
1213	>
1214	>	subroutine get_interatomic_vector(q_i, q_j, d, r_sq)
1215	>
1216	>	real (kind = dp), dimension(3) :: q_i
1217	>	real (kind = dp), dimension(3) :: q_j
1218	>	real ( kind = dp ), intent(out) :: r_sq
1219	>	real( kind = dp ) :: d(3), scaled(3)
1220	>	integer i
1221	>
1222	>	d(1:3) = q_j(1:3) - q_i(1:3)
1223	>
1224	>	! Wrap back into periodic box if necessary
1225	>	if ( SIM_uses_PBC ) then
1226	>
1227	>	if( .not.boxIsOrthorhombic ) then
1228	>	! calc the scaled coordinates.
1229	>
1230	>	scaled = matmul(HmatInv, d)
1231	>
1232	>	! wrap the scaled coordinates
1233	>
1234	>	scaled = scaled  - anint(scaled)
1235	>
1236	>
1237	>	! calc the wrapped real coordinates from the wrapped scaled
1238	>	! coordinates
1239	>
1240	>	d = matmul(Hmat,scaled)
1241	>
1242	>	else
1243	>	! calc the scaled coordinates.
1244	>
1245	>	do i = 1, 3
1246	>	scaled(i) = d(i) * HmatInv(i,i)
1247	>
1248	>	! wrap the scaled coordinates
1249	>
1250	>	scaled(i) = scaled(i) - anint(scaled(i))
1251	>
1252	>	! calc the wrapped real coordinates from the wrapped scaled
1253	>	! coordinates
1254	>
1255	>	d(i) = scaled(i)*Hmat(i,i)
1256	>	enddo
1257	>	endif
1258	>
1259	>	endif
1260	>
1261	>	r_sq = dot_product(d,d)
1262	>
1263	>	end subroutine get_interatomic_vector
1264	>
1265	>	subroutine zero_work_arrays()
1266	>
1267	>	#ifdef IS_MPI
1268	>
1269	>	q_Row = 0.0_dp
1270	>	q_Col = 0.0_dp
1271
1272	+	q_group_Row = 0.0_dp
1273	+	q_group_Col = 0.0_dp
1274	+
1275	+	u_l_Row = 0.0_dp
1276	+	u_l_Col = 0.0_dp
1277	+
1278	+	A_Row = 0.0_dp
1279	+	A_Col = 0.0_dp
1280	+
1281	+	f_Row = 0.0_dp
1282	+	f_Col = 0.0_dp
1283	+	f_Temp = 0.0_dp
1284	+
1285	+	t_Row = 0.0_dp
1286	+	t_Col = 0.0_dp
1287	+	t_Temp = 0.0_dp
1288	+
1289	+	pot_Row = 0.0_dp
1290	+	pot_Col = 0.0_dp
1291	+	pot_Temp = 0.0_dp
1292	+
1293	+	rf_Row = 0.0_dp
1294	+	rf_Col = 0.0_dp
1295	+	rf_Temp = 0.0_dp
1296	+
1297		#endif
1298	<
1299	<	rf = 0.0_dp
1300	<	tau_Temp = 0.0_dp
1301	<	virial_Temp = 0.0_dp
1302	<
1303	<	end subroutine zero_work_arrays
1304	<
1305	<	function skipThisPair(atom1, atom2) result(skip_it)
1306	<	integer, intent(in) :: atom1
1307	<	integer, intent(in), optional :: atom2
1308	<	logical :: skip_it
1309	<	integer :: unique_id_1, unique_id_2
1310	<	integer :: me_i,me_j
1311	<	integer :: i
1312	<
1313	<	skip_it = .false.
1314	<
1315	<	!! there are a number of reasons to skip a pair or a particle
1316	<	!! mostly we do this to exclude atoms who are involved in short
1317	<	!! range interactions (bonds, bends, torsions), but we also need
1318	<	!! to exclude some overcounted interactions that result from
1319	<	!! the parallel decomposition
1320	<
1298	>
1299	>	if (FF_uses_EAM .and. SIM_uses_EAM) then
1300	>	call clean_EAM()
1301	>	endif
1302	>
1303	>	rf = 0.0_dp
1304	>	tau_Temp = 0.0_dp
1305	>	virial_Temp = 0.0_dp
1306	>	end subroutine zero_work_arrays
1307	>
1308	>	function skipThisPair(atom1, atom2) result(skip_it)
1309	>	integer, intent(in) :: atom1
1310	>	integer, intent(in), optional :: atom2
1311	>	logical :: skip_it
1312	>	integer :: unique_id_1, unique_id_2
1313	>	integer :: me_i,me_j
1314	>	integer :: i
1315	>
1316	>	skip_it = .false.
1317	>
1318	>	!! there are a number of reasons to skip a pair or a particle
1319	>	!! mostly we do this to exclude atoms who are involved in short
1320	>	!! range interactions (bonds, bends, torsions), but we also need
1321	>	!! to exclude some overcounted interactions that result from
1322	>	!! the parallel decomposition
1323	>
1324		#ifdef IS_MPI
1325	<	!! in MPI, we have to look up the unique IDs for each atom
1326	<	unique_id_1 = tagRow(atom1)
1325	>	!! in MPI, we have to look up the unique IDs for each atom
1326	>	unique_id_1 = tagRow(atom1)
1327		#else
1328	<	!! in the normal loop, the atom numbers are unique
1329	<	unique_id_1 = atom1
1328	>	!! in the normal loop, the atom numbers are unique
1329	>	unique_id_1 = atom1
1330		#endif
1331	<
1332	<	!! We were called with only one atom, so just check the global exclude
1333	<	!! list for this atom
1334	<	if (.not. present(atom2)) then
1335	<	do i = 1, nExcludes_global
1336	<	if (excludesGlobal(i) == unique_id_1) then
1337	<	skip_it = .true.
1338	<	return
1339	<	end if
1340	<	end do
1341	<	return
1342	<	end if
1343	<
1331	>
1332	>	!! We were called with only one atom, so just check the global exclude
1333	>	!! list for this atom
1334	>	if (.not. present(atom2)) then
1335	>	do i = 1, nExcludes_global
1336	>	if (excludesGlobal(i) == unique_id_1) then
1337	>	skip_it = .true.
1338	>	return
1339	>	end if
1340	>	end do
1341	>	return
1342	>	end if
1343	>
1344		#ifdef IS_MPI
1345	<	unique_id_2 = tagColumn(atom2)
1345	>	unique_id_2 = tagColumn(atom2)
1346		#else
1347	<	unique_id_2 = atom2
1347	>	unique_id_2 = atom2
1348		#endif
1349	<
1349	>
1350		#ifdef IS_MPI
1351	<	!! this situation should only arise in MPI simulations
1352	<	if (unique_id_1 == unique_id_2) then
1353	<	skip_it = .true.
1354	<	return
1355	<	end if
1356	<
1357	<	!! this prevents us from doing the pair on multiple processors
1358	<	if (unique_id_1 < unique_id_2) then
1359	<	if (mod(unique_id_1 + unique_id_2,2) == 0) skip_it = .true.
1360	<	return
1361	<	else
1362	<	if (mod(unique_id_1 + unique_id_2,2) == 1) skip_it = .true.
1363	<	return
1364	<	endif
1351	>	!! this situation should only arise in MPI simulations
1352	>	if (unique_id_1 == unique_id_2) then
1353	>	skip_it = .true.
1354	>	return
1355	>	end if
1356	>
1357	>	!! this prevents us from doing the pair on multiple processors
1358	>	if (unique_id_1 < unique_id_2) then
1359	>	if (mod(unique_id_1 + unique_id_2,2) == 0) then
1360	>	skip_it = .true.
1361	>	return
1362	>	endif
1363	>	else
1364	>	if (mod(unique_id_1 + unique_id_2,2) == 1) then
1365	>	skip_it = .true.
1366	>	return
1367	>	endif
1368	>	endif
1369		#endif
1370	<
1371	<	!! the rest of these situations can happen in all simulations:
1372	<	do i = 1, nExcludes_global
1373	<	if ((excludesGlobal(i) == unique_id_1) .or. &
1374	<	(excludesGlobal(i) == unique_id_2)) then
1375	<	skip_it = .true.
1376	<	return
1377	<	endif
1378	<	enddo
1370	>
1371	>	!! the rest of these situations can happen in all simulations:
1372	>	do i = 1, nExcludes_global
1373	>	if ((excludesGlobal(i) == unique_id_1) .or. &
1374	>	(excludesGlobal(i) == unique_id_2)) then
1375	>	skip_it = .true.
1376	>	return
1377	>	endif
1378	>	enddo
1379	>
1380	>	do i = 1, nSkipsForAtom(unique_id_1)
1381	>	if (skipsForAtom(unique_id_1, i) .eq. unique_id_2) then
1382	>	skip_it = .true.
1383	>	return
1384	>	endif
1385	>	end do
1386	>
1387	>	return
1388	>	end function skipThisPair
1389
1390	<	do i = 1, nExcludes_local
1391	<	if (excludesLocal(1,i) == unique_id_1) then
1392	<	if (excludesLocal(2,i) == unique_id_2) then
1393	<	skip_it = .true.
1394	<	return
1395	<	endif
1396	<	else
1397	<	if (excludesLocal(1,i) == unique_id_2) then
1398	<	if (excludesLocal(2,i) == unique_id_1) then
1399	<	skip_it = .true.
1400	<	return
1401	<	endif
1402	<	endif
1403	<	endif
1404	<	end do
1405	<
1406	<	return
1407	<	end function skipThisPair
1390	>	function FF_UsesDirectionalAtoms() result(doesit)
1391	>	logical :: doesit
1392	>	doesit = FF_uses_dipoles .or. FF_uses_sticky .or. &
1393	>	FF_uses_GB .or. FF_uses_RF
1394	>	end function FF_UsesDirectionalAtoms
1395	>
1396	>	function FF_RequiresPrepairCalc() result(doesit)
1397	>	logical :: doesit
1398	>	doesit = FF_uses_EAM
1399	>	end function FF_RequiresPrepairCalc
1400	>
1401	>	function FF_RequiresPostpairCalc() result(doesit)
1402	>	logical :: doesit
1403	>	doesit = FF_uses_RF
1404	>	end function FF_RequiresPostpairCalc
1405	>
1406	>	#ifdef PROFILE
1407	>	function getforcetime() result(totalforcetime)
1408	>	real(kind=dp) :: totalforcetime
1409	>	totalforcetime = forcetime
1410	>	end function getforcetime
1411	>	#endif
1412	>
1413	>	!! This cleans componets of force arrays belonging only to fortran
1414
1415	<	function FF_UsesDirectionalAtoms() result(doesit)
1416	<	logical :: doesit
1417	<	doesit = FF_uses_dipoles .or. FF_uses_sticky .or. &
1418	<	FF_uses_GB .or. FF_uses_RF
1419	<	end function FF_UsesDirectionalAtoms
1420	<
1421	<	function FF_RequiresPrepairCalc() result(doesit)
1422	<	logical :: doesit
1423	<	doesit = FF_uses_EAM
1424	<	end function FF_RequiresPrepairCalc
1425	<
1426	<	function FF_RequiresPostpairCalc() result(doesit)
1427	<	logical :: doesit
1428	<	doesit = FF_uses_RF
1429	<	end function FF_RequiresPostpairCalc
1430	<
1415	>	subroutine add_stress_tensor(dpair, fpair)
1416	>
1417	>	real( kind = dp ), dimension(3), intent(in) :: dpair, fpair
1418	>
1419	>	! because the d vector is the rj - ri vector, and
1420	>	! because fx, fy, fz are the force on atom i, we need a
1421	>	! negative sign here:
1422	>
1423	>	tau_Temp(1) = tau_Temp(1) - dpair(1) * fpair(1)
1424	>	tau_Temp(2) = tau_Temp(2) - dpair(1) * fpair(2)
1425	>	tau_Temp(3) = tau_Temp(3) - dpair(1) * fpair(3)
1426	>	tau_Temp(4) = tau_Temp(4) - dpair(2) * fpair(1)
1427	>	tau_Temp(5) = tau_Temp(5) - dpair(2) * fpair(2)
1428	>	tau_Temp(6) = tau_Temp(6) - dpair(2) * fpair(3)
1429	>	tau_Temp(7) = tau_Temp(7) - dpair(3) * fpair(1)
1430	>	tau_Temp(8) = tau_Temp(8) - dpair(3) * fpair(2)
1431	>	tau_Temp(9) = tau_Temp(9) - dpair(3) * fpair(3)
1432	>
1433	>	!write(*,'(6es12.3)')  fpair(1:3), tau_Temp(1), tau_Temp(5), tau_temp(9)
1434	>	virial_Temp = virial_Temp + &
1435	>	(tau_Temp(1) + tau_Temp(5) + tau_Temp(9))
1436	>
1437	>	end subroutine add_stress_tensor
1438	>
1439		end module do_Forces
1440	+

Diff Legend

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