ViewVC Help

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

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

Comparing trunk/OOPSE/libmdtools/calc_eam.F90 (file contents):
Revision 497 by chuckv, Mon Apr 14 21:16:37 2003 UTC vs.
Revision 747 by gezelter, Fri Sep 5 21:28:52 2003 UTC

#	Line 1 | Line 1
1	<	module calc_eam
2	<	use definitions, ONLY : DP
1	>	module eam
2	>	use definitions, ONLY : DP,default_error
3	>	use simulation
4		use force_globals
5	<	#ifdef MPI
5	>	use status
6	>	use atype_module
7	>	use Vector_class
8	>	#ifdef IS_MPI
9		use mpiSimulation
10		#endif
11	+	implicit none
12		PRIVATE
13
14
15		logical, save :: EAM_FF_initialized = .false.
16		integer, save :: EAM_Mixing_Policy
17	<	integer, save :: EAM_rcut
17	>	real(kind = dp), save :: EAM_rcut
18	>	logical, save :: haveRcut = .false.
19
20	+	character(len = statusMsgSize) :: errMesg
21	+	integer :: eam_err
22
23	+	character(len = 200) :: errMsg
24	+	character(len=*), parameter :: RoutineName =  "EAM MODULE"
25	+	!! Logical that determines if eam arrays should be zeroed
26	+	logical :: cleanme = .true.
27	+	logical :: nmflag  = .false.
28
29	+
30	+	type, private :: EAMtype
31	+	integer           :: eam_atype
32	+	real( kind = DP ) :: eam_dr
33	+	integer           :: eam_nr
34	+	integer           :: eam_nrho
35	+	real( kind = DP ) :: eam_lattice
36	+	real( kind = DP ) :: eam_drho
37	+	real( kind = DP ) :: eam_rcut
38	+	integer           :: eam_atype_map
39	+
40	+	real( kind = DP ), pointer, dimension(:) :: eam_rvals        => null()
41	+	real( kind = DP ), pointer, dimension(:) :: eam_rhovals      => null()
42	+	real( kind = DP ), pointer, dimension(:) :: eam_F_rho        => null()
43	+	real( kind = DP ), pointer, dimension(:) :: eam_Z_r          => null()
44	+	real( kind = DP ), pointer, dimension(:) :: eam_rho_r        => null()
45	+	real( kind = DP ), pointer, dimension(:) :: eam_phi_r        => null()
46	+	real( kind = DP ), pointer, dimension(:) :: eam_F_rho_pp     => null()
47	+	real( kind = DP ), pointer, dimension(:) :: eam_Z_r_pp       => null()
48	+	real( kind = DP ), pointer, dimension(:) :: eam_rho_r_pp     => null()
49	+	real( kind = DP ), pointer, dimension(:) :: eam_phi_r_pp     => null()
50	+	end type EAMtype
51
17	–	!! standard eam stuff
18	–	integer                                        :: n_eam_atypes
19	–	integer,           allocatable, dimension(:)   :: eam_atype
20	–	real( kind = DP ), allocatable, dimension(:)   :: eam_dr
21	–	integer, allocatable, dimension(:)   :: eam_nr
22	–	integer, allocatable, dimension(:)   :: eam_nrho
23	–	real( kind = DP ), allocatable, dimension(:)   :: eam_lattice
24	–	real( kind = DP ), allocatable, dimension(:)   :: eam_drho
25	–	integer          , allocatable, dimension(:)   :: eam_atype_map
26	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_rvals
27	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_rhovals
28	–	real( kind = DP ), allocatable, dimension(:)   :: eam_rcut
29	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_F_rho
30	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_Z_r
31	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_rho_r
32	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_phi_r
33	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_F_rho_pp
34	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_Z_r_pp
35	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_rho_r_pp
36	–	real( kind = DP ), allocatable, dimension(:,:) :: eam_phi_r_pp
52
53	+	!! Arrays for derivatives used in force calculation
54	+	real( kind = dp), dimension(:), allocatable :: frho
55	+	real( kind = dp), dimension(:), allocatable :: rho
56
57	<	public :: init_EAM_FF
58	<	public :: EAM_new_rcut
41	<	public :: do_EAM_pair
57	>	real( kind = dp), dimension(:), allocatable :: dfrhodrho
58	>	real( kind = dp), dimension(:), allocatable :: d2frhodrhodrho
59
60
61	+	!! Arrays for MPI storage
62	+	#ifdef IS_MPI
63	+	real( kind = dp),save, dimension(:), allocatable :: dfrhodrho_col
64	+	real( kind = dp),save, dimension(:), allocatable :: dfrhodrho_row
65	+	real( kind = dp),save, dimension(:), allocatable :: frho_row
66	+	real( kind = dp),save, dimension(:), allocatable :: frho_col
67	+	real( kind = dp),save, dimension(:), allocatable :: rho_row
68	+	real( kind = dp),save, dimension(:), allocatable :: rho_col
69	+	real( kind = dp),save, dimension(:), allocatable :: d2frhodrhodrho_col
70	+	real( kind = dp),save, dimension(:), allocatable :: d2frhodrhodrho_row
71	+	#endif
72
73	<	contains
74	<	subroutine init_EAM_FF()
73	>	type, private :: EAMTypeList
74	>	integer           :: n_eam_types = 0
75	>	integer           :: currentAddition = 0
76	>
77	>	type (EAMtype), pointer  :: EAMParams(:) => null()
78	>	end type EAMTypeList
79
80
81	+	type (eamTypeList), save :: EAMList
82
83	<	character(len=80) :: eam_pot_file
51	<	integer :: i, j,  max_size, prev_max_size
52	<	integer :: number_rho, number_r
53	<	integer :: eam_unit
54	<	integer :: this_error
55	<	character(len=300) :: msg
56	<	integer, external :: nfiles
57	<	!for mpi
83	>	!! standard eam stuff
84
85
86	<	#ifdef MPI
87	<	if (node == 0)  &
88	<	n_eam_atypes = nfiles(trim(eam_pot_dir)//char(0))
86	>	public :: init_EAM_FF
87	>	public :: setCutoffEAM
88	>	public :: do_eam_pair
89	>	public :: newEAMtype
90	>	public :: calc_eam_prepair_rho
91	>	public :: calc_eam_preforce_Frho
92	>	public :: clean_EAM
93
94	<	call mpi_bcast(n_eam_atypes,1,mpi_integer,0,mpi_comm_world,mpi_err)
65	<	if (n_eam_atypes == -1) then
66	<	call error("INITIALIZE_EAM","NO EAM potentials found!")
67	<	endif
68	<	write(msg,'(a5,i4,a12,i5,a14)') 'Node: ',node,' reading ...', &
69	<	n_eam_atypes, ' eam atom types'
70	<	call info('INITIALIZE_EAM', trim(msg))
71	<	#else
72	<	n_eam_atypes = nfiles(trim(eam_pot_dir)//char(0))
73	<	if (n_eam_atypes == -1) then
74	<	call error("INITIALIZE_EAM","NO EAM potentials found!")
75	<	endif
94	>	contains
95
77	–	write(msg,'(a12,i5,a14)') ' Reading ...', &
78	–	n_eam_atypes, ' eam atom types'
79	–	call info('INITIALIZE_EAM', trim(msg))
80	–	#endif
96
97	+	subroutine newEAMtype(lattice_constant,eam_nrho,eam_drho,eam_nr,&
98	+	eam_dr,rcut,eam_Z_r,eam_rho_r,eam_F_rho,&
99	+	eam_ident,status)
100	+	real (kind = dp )                      :: lattice_constant
101	+	integer                                :: eam_nrho
102	+	real (kind = dp )                      :: eam_drho
103	+	integer                                :: eam_nr
104	+	real (kind = dp )                      :: eam_dr
105	+	real (kind = dp )                      :: rcut
106	+	real (kind = dp ), dimension(eam_nr)   :: eam_Z_r
107	+	real (kind = dp ), dimension(eam_nr)   :: eam_rho_r
108	+	real (kind = dp ), dimension(eam_nrho) :: eam_F_rho
109	+	integer                                :: eam_ident
110	+	integer                                :: status
111
112	<	call allocate_eam_atype(n_eam_atypes)
112	>	integer                                :: nAtypes
113	>	integer                                :: maxVals
114	>	integer                                :: alloc_stat
115	>	integer                                :: current
116	>	integer,pointer                        :: Matchlist(:) => null()
117
118	+	status = 0
119
120
121	<	!! get largest number of data points for any potential
122	<	#ifdef MPI
123	<	if (node == 0) then
124	<	#endif
125	<	prev_max_size = 0
126	<	do i = 1, n_eam_atypes
127	<	call getfilename(i, eam_pot_file)
128	<	max_size = max(get_eam_sizes( &
129	<	trim(eam_pot_dir) // '/' // eam_pot_file), &
96	<	prev_max_size)
97	<	prev_max_size = max_size
98	<	end do
99	<	#ifdef MPI
121	>	!! Assume that atypes has already been set and get the total number of types in atypes
122	>	!! Also assume that every member of atypes is a EAM model.
123	>
124	>
125	>	! check to see if this is the first time into
126	>	if (.not.associated(EAMList%EAMParams)) then
127	>	call getMatchingElementList(atypes, "is_EAM", .true., nAtypes, MatchList)
128	>	EAMList%n_eam_types = nAtypes
129	>	allocate(EAMList%EAMParams(nAtypes))
130		end if
131
132	+	EAMList%currentAddition = EAMList%currentAddition + 1
133	+	current = EAMList%currentAddition
134	+
135
136	<	call mpi_bcast(max_size,1,mpi_integer,0,mpi_comm_world,mpi_err)
137	<	#endif
136	>	call allocate_EAMType(eam_nrho,eam_nr,EAMList%EAMParams(current),stat=alloc_stat)
137	>	if (alloc_stat /= 0) then
138	>	status = -1
139	>	return
140	>	end if
141
142	<	call allocate_eam_module(n_eam_atypes,max_size)
143	<	allocate(eam_atype_map(get_max_atype()))
142	>	! this is a possible bug, we assume a correspondence between the vector atypes and
143	>	! EAMAtypes
144	>
145	>	EAMList%EAMParams(current)%eam_atype    = eam_ident
146	>	EAMList%EAMParams(current)%eam_lattice  = lattice_constant
147	>	EAMList%EAMParams(current)%eam_nrho     = eam_nrho
148	>	EAMList%EAMParams(current)%eam_drho     = eam_drho
149	>	EAMList%EAMParams(current)%eam_nr       = eam_nr
150	>	EAMList%EAMParams(current)%eam_dr       = eam_dr
151	>	EAMList%EAMParams(current)%eam_rcut     = rcut
152	>	EAMList%EAMParams(current)%eam_Z_r      = eam_Z_r
153	>	EAMList%EAMParams(current)%eam_rho_r    = eam_rho_r
154	>	EAMList%EAMParams(current)%eam_F_rho    = eam_F_rho
155
156	<	#ifdef MPI
110	<	if (node == 0) then
111	<	#endif
112	<	do i = 1, n_eam_atypes
113	<	call getfilename(i, eam_pot_file)
114	<	call read_eam_pot(i,trim(eam_pot_dir) // '/' // eam_pot_file, &
115	<	this_error)
156	>	end subroutine newEAMtype
157
117	–	do j = 1, eam_nr(i)
118	–	eam_rvals(j,i) = dble(j-1)*eam_dr(i)
119	–	enddo
158
121	–	do j = 1, eam_nrho(i)
122	–	eam_rhovals(j,i) = dble(j-1)*eam_drho(i)
123	–	enddo
159
160	+	subroutine init_EAM_FF(status)
161	+	integer :: status
162	+	integer :: i,j
163	+	real(kind=dp) :: current_rcut_max
164	+	integer :: alloc_stat
165	+	integer :: number_r, number_rho
166	+
167	+	if (EAMList%currentAddition == 0) then
168	+	call handleError("init_EAM_FF","No members in EAMList")
169	+	status = -1
170	+	return
171	+	end if
172	+
173	+
174	+
175	+	do i = 1, EAMList%currentAddition
176	+
177	+	! Build array of r values
178	+
179	+	do j = 1,EAMList%EAMParams(i)%eam_nr
180	+	EAMList%EAMParams(i)%eam_rvals(j) = &
181	+	real(j-1,kind=dp)* &
182	+	EAMList%EAMParams(i)%eam_dr
183	+	end do
184	+	! Build array of rho values
185	+	do j = 1,EAMList%EAMParams(i)%eam_nrho
186	+	EAMList%EAMParams(i)%eam_rhovals(j) = &
187	+	real(j-1,kind=dp)* &
188	+	EAMList%EAMParams(i)%eam_drho
189	+	end do
190		! convert from eV to kcal / mol:
191	<	do j = 1, eam_nrho(i)
127	<	eam_F_rho(j,i) = eam_F_rho(j,i)*23.06054E0_DP
128	<	enddo
191	>	EAMList%EAMParams(i)%eam_F_rho = EAMList%EAMParams(i)%eam_F_rho * 23.06054E0_DP
192
193		! precompute the pair potential and get it into kcal / mol:
194	<	eam_phi_r(1,i) = 0.0E0_DP
195	<	do j = 2, eam_nr(i)
196	<	eam_phi_r(j,i) = (eam_Z_r(j,i)**2)/eam_rvals(j,i)
197	<	eam_phi_r(j,i) = eam_phi_r(j,i)*331.999296E0_DP
194	>	EAMList%EAMParams(i)%eam_phi_r(1) = 0.0E0_DP
195	>	do j = 2, EAMList%EAMParams(i)%eam_nr
196	>	EAMList%EAMParams(i)%eam_phi_r(j) = (EAMList%EAMParams(i)%eam_Z_r(j)**2)/EAMList%EAMParams(i)%eam_rvals(j)
197	>	EAMList%EAMParams(i)%eam_phi_r(j) =  EAMList%EAMParams(i)%eam_phi_r(j)*331.999296E0_DP
198		enddo
136	–
199		end do
200	<	#ifdef MPI
139	<	call info('INITIALIZE_EAM','NODE 0: Distributing spline arrays')
140	<	endif
200	>
201
202	<	call mpi_bcast(this_error,n_eam_atypes,mpi_integer,0, &
203	<	mpi_comm_world,mpi_err)
204	<	if (this_error /= 0) then
205	<	call error('INITIALIZE_EAM',"Cannot read eam files")
206	<	endif
202	>	do i = 1,  EAMList%currentAddition
203	>	number_r   = EAMList%EAMParams(i)%eam_nr
204	>	number_rho = EAMList%EAMParams(i)%eam_nrho
205	>
206	>	call eam_spline(number_r, EAMList%EAMParams(i)%eam_rvals, &
207	>	EAMList%EAMParams(i)%eam_rho_r, &
208	>	EAMList%EAMParams(i)%eam_rho_r_pp, &
209	>	0.0E0_DP, 0.0E0_DP, 'N')
210	>	call eam_spline(number_r, EAMList%EAMParams(i)%eam_rvals, &
211	>	EAMList%EAMParams(i)%eam_Z_r, &
212	>	EAMList%EAMParams(i)%eam_Z_r_pp, &
213	>	0.0E0_DP, 0.0E0_DP, 'N')
214	>	call eam_spline(number_rho, EAMList%EAMParams(i)%eam_rhovals, &
215	>	EAMList%EAMParams(i)%eam_F_rho, &
216	>	EAMList%EAMParams(i)%eam_F_rho_pp, &
217	>	0.0E0_DP, 0.0E0_DP, 'N')
218	>	call eam_spline(number_r, EAMList%EAMParams(i)%eam_rvals, &
219	>	EAMList%EAMParams(i)%eam_phi_r, &
220	>	EAMList%EAMParams(i)%eam_phi_r_pp, &
221	>	0.0E0_DP, 0.0E0_DP, 'N')
222	>	enddo
223
148	–	call mpi_bcast(eam_atype,n_eam_atypes,mpi_integer,0, &
149	–	mpi_comm_world,mpi_err)
224
225	<	!! distribute values to cluster......
226	<	call mpi_bcast(eam_nr,n_eam_atypes,mpi_integer,&
227	<	0,mpi_comm_world,mpi_err)
228	<	call mpi_bcast(eam_nrho,n_eam_atypes,mpi_integer,&
229	<	0,mpi_comm_world,mpi_err)
156	<	call mpi_bcast(eam_rvals,n_eam_atypes*max_size,mpi_double_precision, &
157	<	0,mpi_comm_world,mpi_err)
158	<	call mpi_bcast(eam_rcut,n_eam_atypes,mpi_double_precision, &
159	<	0,mpi_comm_world,mpi_err)
160	<	call mpi_bcast(eam_rhovals,n_eam_atypes*max_size,mpi_double_precision, &
161	<	0,mpi_comm_world,mpi_err)
225	>	!       current_rcut_max = EAMList%EAMParams(1)%eam_rcut
226	>	!! find the smallest rcut for any eam atype
227	>	!       do i = 2, EAMList%currentAddition
228	>	!          current_rcut_max =max(current_rcut_max,EAMList%EAMParams(i)%eam_rcut)
229	>	!       end do
230
231	<	!! distribute arrays
232	<	call mpi_bcast(eam_rho_r,n_eam_atypes*max_size,mpi_double_precision, &
233	<	0,mpi_comm_world,mpi_err)
234	<	call mpi_bcast(eam_Z_r,n_eam_atypes*max_size,mpi_double_precision, &
235	<	0,mpi_comm_world,mpi_err)
236	<	call mpi_bcast(eam_F_rho,n_eam_atypes*max_size,mpi_double_precision, &
237	<	0,mpi_comm_world,mpi_err)
238	<	call mpi_bcast(eam_phi_r,n_eam_atypes*max_size,mpi_double_precision, &
239	<	0,mpi_comm_world,mpi_err)
231	>	!       EAM_rcut = current_rcut_max
232	>	!       EAM_rcut_orig = current_rcut_max
233	>	!       do i = 1, EAMList%currentAddition
234	>	!          EAMList%EAMParam(i)s%eam_atype_map(eam_atype(i)) = i
235	>	!       end do
236	>	!! Allocate arrays for force calculation
237	>	call allocateEAM(alloc_stat)
238	>	if (alloc_stat /= 0 ) then
239	>	status = -1
240	>	return
241	>	endif
242
243	<	#endif
174	<	call info('INITIALIZE_EAM', 'creating splines')
243	>	end subroutine init_EAM_FF
244
245	<	do i = 1, n_eam_atypes
246	<	number_r = eam_nr(i)
247	<	number_rho = eam_nrho(i)
245	>	!! routine checks to see if array is allocated, deallocates array if allocated
246	>	!! and then creates the array to the required size
247	>	subroutine allocateEAM(status)
248	>	integer, intent(out) :: status
249
250	<	call eam_spline(i, number_r, eam_rvals, eam_rho_r, eam_rho_r_pp, &
251	<	0.0E0_DP, 0.0E0_DP, 'N')
252	<	call eam_spline(i, number_r, eam_rvals, eam_Z_r, eam_Z_r_pp, &
253	<	0.0E0_DP, 0.0E0_DP, 'N')
254	<	call eam_spline(i, number_rho, eam_rhovals, eam_F_rho, eam_F_rho_pp, &
255	<	0.0E0_DP, 0.0E0_DP, 'N')
186	<	call eam_spline(i, number_r, eam_rvals, eam_phi_r, eam_phi_r_pp, &
187	<	0.0E0_DP, 0.0E0_DP, 'N')
188	<	enddo
250	>	integer :: nlocal
251	>	#ifdef IS_MPI
252	>	integer :: nrow
253	>	integer :: ncol
254	>	#endif
255	>	integer :: alloc_stat
256
190	–	do i = 1, n_eam_atypes
191	–	eam_atype_map(eam_atype(i)) = i
192	–	end do
257
258	+	nlocal = getNlocal()
259
260	+	#ifdef IS_MPI
261	+	nrow = getNrow(plan_row)
262	+	ncol = getNcol(plan_col)
263	+	#endif
264
265	<	call info('INITIALIZE_EAM','Done creating splines')
265	>	if (allocated(frho)) deallocate(frho)
266	>	allocate(frho(nlocal),stat=alloc_stat)
267	>	if (alloc_stat /= 0) then
268	>	status = -1
269	>	return
270	>	end if
271	>	if (allocated(rho)) deallocate(rho)
272	>	allocate(rho(nlocal),stat=alloc_stat)
273	>	if (alloc_stat /= 0) then
274	>	status = -1
275	>	return
276	>	end if
277
278	<	return
279	<	end subroutine initialize_eam
278	>	if (allocated(dfrhodrho)) deallocate(dfrhodrho)
279	>	allocate(dfrhodrho(nlocal),stat=alloc_stat)
280	>	if (alloc_stat /= 0) then
281	>	status = -1
282	>	return
283	>	end if
284
285	+	if (allocated(d2frhodrhodrho)) deallocate(d2frhodrhodrho)
286	+	allocate(d2frhodrhodrho(nlocal),stat=alloc_stat)
287	+	if (alloc_stat /= 0) then
288	+	status = -1
289	+	return
290	+	end if
291	+
292	+	#ifdef IS_MPI
293
294	<	subroutine allocate_eam_atype(n_size_atype)
295	<	integer, intent(in) :: n_size_atype
294	>	if (allocated(frho_row)) deallocate(frho_row)
295	>	allocate(frho_row(nrow),stat=alloc_stat)
296	>	if (alloc_stat /= 0) then
297	>	status = -1
298	>	return
299	>	end if
300	>	if (allocated(rho_row)) deallocate(rho_row)
301	>	allocate(rho_row(nrow),stat=alloc_stat)
302	>	if (alloc_stat /= 0) then
303	>	status = -1
304	>	return
305	>	end if
306	>	if (allocated(dfrhodrho_row)) deallocate(dfrhodrho_row)
307	>	allocate(dfrhodrho_row(nrow),stat=alloc_stat)
308	>	if (alloc_stat /= 0) then
309	>	status = -1
310	>	return
311	>	end if
312	>	if (allocated(d2frhodrhodrho_row)) deallocate(d2frhodrhodrho_row)
313	>	allocate(d2frhodrhodrho_row(nrow),stat=alloc_stat)
314	>	if (alloc_stat /= 0) then
315	>	status = -1
316	>	return
317	>	end if
318
205	–	allocate(eam_atype(n_size_atype))
206	–	allocate(eam_drho(n_size_atype))
207	–	allocate(eam_dr(n_size_atype))
208	–	allocate(eam_nr(n_size_atype))
209	–	allocate(eam_nrho(n_size_atype))
210	–	allocate(eam_lattice(n_size_atype))
211	–	allocate(eam_rcut(n_size_atype))
319
320	<	end subroutine allocate_eam_atype
320	>	! Now do column arrays
321
322	<	subroutine allocate_eam_module(n_size_atype,n_eam_points)
323	<	integer, intent(in) :: n_eam_points
324	<	integer, intent(in) :: n_size_atype
322	>	if (allocated(frho_col)) deallocate(frho_col)
323	>	allocate(frho_col(ncol),stat=alloc_stat)
324	>	if (alloc_stat /= 0) then
325	>	status = -1
326	>	return
327	>	end if
328	>	if (allocated(rho_col)) deallocate(rho_col)
329	>	allocate(rho_col(ncol),stat=alloc_stat)
330	>	if (alloc_stat /= 0) then
331	>	status = -1
332	>	return
333	>	end if
334	>	if (allocated(dfrhodrho_col)) deallocate(dfrhodrho_col)
335	>	allocate(dfrhodrho_col(ncol),stat=alloc_stat)
336	>	if (alloc_stat /= 0) then
337	>	status = -1
338	>	return
339	>	end if
340	>	if (allocated(d2frhodrhodrho_col)) deallocate(d2frhodrhodrho_col)
341	>	allocate(d2frhodrhodrho_col(ncol),stat=alloc_stat)
342	>	if (alloc_stat /= 0) then
343	>	status = -1
344	>	return
345	>	end if
346	>
347	>	#endif
348
349	<	allocate(eam_rvals(n_eam_points,n_size_atype))
220	<	allocate(eam_rhovals(n_eam_points,n_size_atype))
221	<	allocate(eam_F_rho(n_eam_points,n_size_atype))
222	<	allocate(eam_Z_r(n_eam_points,n_size_atype))
223	<	allocate(eam_rho_r(n_eam_points,n_size_atype))
224	<	allocate(eam_phi_r(n_eam_points,n_size_atype))
225	<	allocate(eam_F_rho_pp(n_eam_points,n_size_atype))
226	<	allocate(eam_Z_r_pp(n_eam_points,n_size_atype))
227	<	allocate(eam_rho_r_pp(n_eam_points,n_size_atype))
228	<	allocate(eam_phi_r_pp(n_eam_points,n_size_atype))
349	>	end subroutine allocateEAM
350
351	<	end subroutine allocate_eam_module
351	>	!! C sets rcut to be the largest cutoff of any atype
352	>	!! present in this simulation. Doesn't include all atypes
353	>	!! sim knows about, just those in the simulation.
354	>	subroutine setCutoffEAM(rcut, status)
355	>	real(kind=dp) :: rcut
356	>	integer :: status
357	>	status = 0
358
359	<	subroutine deallocate_eam_module()
359	>	EAM_rcut = rcut
360
361	<	deallocate(eam_atype)
362	<	deallocate(eam_drho)
363	<	deallocate(eam_dr)
364	<	deallocate(eam_nr)
365	<	deallocate(eam_nrho)
366	<	deallocate(eam_lattice)
367	<	deallocate(eam_atype_map)
368	<	deallocate(eam_rvals)
369	<	deallocate(eam_rhovals)
370	<	deallocate(eam_rcut)
371	<	deallocate(eam_Z_r)
372	<	deallocate(eam_rho_r)
373	<	deallocate(eam_phi_r)
374	<	deallocate(eam_F_rho_pp)
375	<	deallocate(eam_Z_r_pp)
376	<	deallocate(eam_rho_r_pp)
377	<	deallocate(eam_phi_r_pp)
361	>	end subroutine setCutoffEAM
362	>
363	>
364	>
365	>	subroutine clean_EAM()
366	>
367	>	! clean non-IS_MPI first
368	>	frho = 0.0_dp
369	>	rho  = 0.0_dp
370	>	dfrhodrho = 0.0_dp
371	>	! clean MPI if needed
372	>	#ifdef IS_MPI
373	>	frho_row = 0.0_dp
374	>	frho_col = 0.0_dp
375	>	rho_row  = 0.0_dp
376	>	rho_col  = 0.0_dp
377	>	dfrhodrho_row = 0.0_dp
378	>	dfrhodrho_col = 0.0_dp
379	>	#endif
380	>	end subroutine clean_EAM
381	>
382	>
383	>
384	>	subroutine allocate_EAMType(eam_n_rho,eam_n_r,thisEAMType,stat)
385	>	integer, intent(in)          :: eam_n_rho
386	>	integer, intent(in)          :: eam_n_r
387	>	type (EAMType)               :: thisEAMType
388	>	integer, optional   :: stat
389	>	integer             :: alloc_stat
390	>
391	>
392	>
393	>	if (present(stat)) stat = 0
394	>
395	>	allocate(thisEAMType%eam_rvals(eam_n_r),stat=alloc_stat)
396	>	if (alloc_stat /= 0 ) then
397	>	if (present(stat)) stat = -1
398	>	return
399	>	end if
400	>	allocate(thisEAMType%eam_rhovals(eam_n_rho),stat=alloc_stat)
401	>	if (alloc_stat /= 0 ) then
402	>	if (present(stat)) stat = -1
403	>	return
404	>	end if
405	>	allocate(thisEAMType%eam_F_rho(eam_n_rho),stat=alloc_stat)
406	>	if (alloc_stat /= 0 ) then
407	>	if (present(stat)) stat = -1
408	>	return
409	>	end if
410	>	allocate(thisEAMType%eam_Z_r(eam_n_r),stat=alloc_stat)
411	>	if (alloc_stat /= 0 ) then
412	>	if (present(stat)) stat = -1
413	>	return
414	>	end if
415	>	allocate(thisEAMType%eam_rho_r(eam_n_r),stat=alloc_stat)
416	>	if (alloc_stat /= 0 ) then
417	>	if (present(stat)) stat = -1
418	>	return
419	>	end if
420	>	allocate(thisEAMType%eam_phi_r(eam_n_r),stat=alloc_stat)
421	>	if (alloc_stat /= 0 ) then
422	>	if (present(stat)) stat = -1
423	>	return
424	>	end if
425	>	allocate(thisEAMType%eam_F_rho_pp(eam_n_rho),stat=alloc_stat)
426	>	if (alloc_stat /= 0 ) then
427	>	if (present(stat)) stat = -1
428	>	return
429	>	end if
430	>	allocate(thisEAMType%eam_Z_r_pp(eam_n_r),stat=alloc_stat)
431	>	if (alloc_stat /= 0 ) then
432	>	if (present(stat)) stat = -1
433	>	return
434	>	end if
435	>	allocate(thisEAMType%eam_rho_r_pp(eam_n_r),stat=alloc_stat)
436	>	if (alloc_stat /= 0 ) then
437	>	if (present(stat)) stat = -1
438	>	return
439	>	end if
440	>	allocate(thisEAMType%eam_phi_r_pp(eam_n_r),stat=alloc_stat)
441	>	if (alloc_stat /= 0 ) then
442	>	if (present(stat)) stat = -1
443	>	return
444	>	end if
445	>
446	>
447	>	end subroutine allocate_EAMType
448	>
449	>
450	>	subroutine deallocate_EAMType(thisEAMType)
451	>	type (EAMtype), pointer :: thisEAMType
452	>
453	>	! free Arrays in reverse order of allocation...
454	>	deallocate(thisEAMType%eam_phi_r_pp)
455	>	deallocate(thisEAMType%eam_rho_r_pp)
456	>	deallocate(thisEAMType%eam_Z_r_pp)
457	>	deallocate(thisEAMType%eam_F_rho_pp)
458	>	deallocate(thisEAMType%eam_phi_r)
459	>	deallocate(thisEAMType%eam_rho_r)
460	>	deallocate(thisEAMType%eam_Z_r)
461	>	deallocate(thisEAMType%eam_F_rho)
462	>	deallocate(thisEAMType%eam_rhovals)
463	>	deallocate(thisEAMType%eam_rvals)
464	>
465	>	end subroutine deallocate_EAMType
466	>
467	>	!! Calculates rho_r
468	>	subroutine calc_eam_prepair_rho(atom1,atom2,d,r,rijsq)
469	>	integer :: atom1,atom2
470	>	real(kind = dp), dimension(3) :: d
471	>	real(kind = dp), intent(inout)               :: r
472	>	real(kind = dp), intent(inout)               :: rijsq
473	>	! value of electron density rho do to atom i at atom j
474	>	real(kind = dp) :: rho_i_at_j
475	>	! value of electron density rho do to atom j at atom i
476	>	real(kind = dp) :: rho_j_at_i
477	>
478	>	! we don't use the derivatives, dummy variables
479	>	real( kind = dp) :: drho,d2rho
480	>	integer :: eam_err
481	>
482	>	integer :: myid_atom1
483	>	integer :: myid_atom2
484	>
485	>	! check to see if we need to be cleaned at the start of a force loop
486	>
487	>
488	>
489	>
490	>	#ifdef IS_MPI
491	>	myid_atom1 = atid_Row(atom1)
492	>	myid_atom2 = atid_Col(atom2)
493	>	#else
494	>	myid_atom1 = atid(atom1)
495	>	myid_atom2 = atid(atom2)
496	>	#endif
497	>
498	>	if (r.lt.EAMList%EAMParams(myid_atom1)%eam_rcut) then
499	>
500	>
501	>
502	>	call eam_splint(EAMList%EAMParams(myid_atom1)%eam_nr, &
503	>	EAMList%EAMParams(myid_atom1)%eam_rvals, &
504	>	EAMList%EAMParams(myid_atom1)%eam_rho_r, &
505	>	EAMList%EAMParams(myid_atom1)%eam_rho_r_pp, &
506	>	r, rho_i_at_j,drho,d2rho)
507	>
508	>
509	>
510	>	#ifdef  IS_MPI
511	>	rho_col(atom2) = rho_col(atom2) + rho_i_at_j
512	>	#else
513	>	rho(atom2) = rho(atom2) + rho_i_at_j
514	>	#endif
515	>	!       write(*,*) atom1,atom2,r,rho_i_at_j
516	>	endif
517	>
518	>	if (r.lt.EAMList%EAMParams(myid_atom2)%eam_rcut) then
519	>	call eam_splint(EAMList%EAMParams(myid_atom2)%eam_nr, &
520	>	EAMList%EAMParams(myid_atom2)%eam_rvals, &
521	>	EAMList%EAMParams(myid_atom2)%eam_rho_r, &
522	>	EAMList%EAMParams(myid_atom2)%eam_rho_r_pp, &
523	>	r, rho_j_at_i,drho,d2rho)
524	>
525	>
526	>
527	>
528	>	#ifdef  IS_MPI
529	>	rho_row(atom1) = rho_row(atom1) + rho_j_at_i
530	>	#else
531	>	rho(atom1) = rho(atom1) + rho_j_at_i
532	>	#endif
533	>	endif
534	>
535	>
536	>	end subroutine calc_eam_prepair_rho
537	>
538	>
539	>
540	>
541	>	!! Calculate the functional F(rho) for all local atoms
542	>	subroutine calc_eam_preforce_Frho(nlocal,pot)
543	>	integer :: nlocal
544	>	real(kind=dp) :: pot
545	>	integer :: i,j
546	>	integer :: atom
547	>	real(kind=dp) :: U,U1,U2
548	>	integer :: atype1
549	>	integer :: me
550	>	integer :: n_rho_points
551	>
552	>
553	>	cleanme = .true.
554	>	!! Scatter the electron density from  pre-pair calculation back to local atoms
555	>	#ifdef IS_MPI
556	>	call scatter(rho_row,rho,plan_row,eam_err)
557	>	if (eam_err /= 0 ) then
558	>	write(errMsg,*) " Error scattering rho_row into rho"
559	>	call handleError(RoutineName,errMesg)
560	>	endif
561	>	call scatter(rho_col,rho,plan_col,eam_err)
562	>	if (eam_err /= 0 ) then
563	>	write(errMsg,*) " Error scattering rho_col into rho"
564	>	call handleError(RoutineName,errMesg)
565	>	endif
566	>	#endif
567
252	–	end subroutine deallocate_eam_module
568
569	+	!! Calculate F(rho) and derivative
570	+	do atom = 1, nlocal
571	+	me = atid(atom)
572	+	n_rho_points = EAMList%EAMParams(me)%eam_nrho
573	+	!  Check to see that the density is not greater than the larges rho we have calculated
574	+	if (rho(atom) < EAMList%EAMParams(me)%eam_rhovals(n_rho_points)) then
575	+	call eam_splint(n_rho_points, &
576	+	EAMList%EAMParams(me)%eam_rhovals, &
577	+	EAMList%EAMParams(me)%eam_f_rho, &
578	+	EAMList%EAMParams(me)%eam_f_rho_pp, &
579	+	rho(atom), & ! Actual Rho
580	+	u, u1, u2)
581	+	else
582	+	! Calculate F(rho with the largest available rho value
583	+	call eam_splint(n_rho_points, &
584	+	EAMList%EAMParams(me)%eam_rhovals, &
585	+	EAMList%EAMParams(me)%eam_f_rho, &
586	+	EAMList%EAMParams(me)%eam_f_rho_pp, &
587	+	EAMList%EAMParams(me)%eam_rhovals(n_rho_points), & ! Largest rho
588	+	u,u1,u2)
589	+	end if
590
591	<	subroutine calc_eam_pair(atom1,atom2,d,rij,r2,pot,f,do_pot,do_stress)
592	<	!Arguments
591	>
592	>	frho(atom) = u
593	>	dfrhodrho(atom) = u1
594	>	d2frhodrhodrho(atom) = u2
595	>	pot = pot + u
596	>
597	>	enddo
598	>
599	>
600	>
601	>	#ifdef IS_MPI
602	>	!! communicate f(rho) and derivatives back into row and column arrays
603	>	call gather(frho,frho_row,plan_row, eam_err)
604	>	if (eam_err /=  0) then
605	>	call handleError("cal_eam_forces()","MPI gather frho_row failure")
606	>	endif
607	>	call gather(dfrhodrho,dfrhodrho_row,plan_row, eam_err)
608	>	if (eam_err /=  0) then
609	>	call handleError("cal_eam_forces()","MPI gather dfrhodrho_row failure")
610	>	endif
611	>	call gather(frho,frho_col,plan_col, eam_err)
612	>	if (eam_err /=  0) then
613	>	call handleError("cal_eam_forces()","MPI gather frho_col failure")
614	>	endif
615	>	call gather(dfrhodrho,dfrhodrho_col,plan_col, eam_err)
616	>	if (eam_err /=  0) then
617	>	call handleError("cal_eam_forces()","MPI gather dfrhodrho_col failure")
618	>	endif
619	>
620	>
621	>
622	>
623	>
624	>	if (nmflag) then
625	>	call gather(d2frhodrhodrho,d2frhodrhodrho_row,plan_row)
626	>	call gather(d2frhodrhodrho,d2frhodrhodrho_col,plan_col)
627	>	endif
628	>	#endif
629	>
630	>
631	>	end subroutine calc_eam_preforce_Frho
632	>
633	>
634	>
635	>
636	>	!! Does EAM pairwise Force calculation.
637	>	subroutine do_eam_pair(atom1,atom2,d,rij,r2,pot,f,do_pot,do_stress)
638	>	!Arguments
639		integer, intent(in) ::  atom1, atom2
640		real( kind = dp ), intent(in) :: rij, r2
641		real( kind = dp ) :: pot
642		real( kind = dp ), dimension(3,getNlocal()) :: f
643		real( kind = dp ), intent(in), dimension(3) :: d
644		logical, intent(in) :: do_pot, do_stress
645	+
646	+	real( kind = dp ) :: drdx,drdy,drdz
647	+	real( kind = dp ) :: d2
648	+	real( kind = dp ) :: phab,pha,dvpdr,d2vpdrdr
649	+	real( kind = dp ) :: rha,drha,d2rha, dpha
650	+	real( kind = dp ) :: rhb,drhb,d2rhb, dphb
651	+	real( kind = dp ) :: dudr
652	+	real( kind = dp ) :: rci,rcj
653	+	real( kind = dp ) :: drhoidr,drhojdr
654	+	real( kind = dp ) :: d2rhoidrdr
655	+	real( kind = dp ) :: d2rhojdrdr
656	+	real( kind = dp ) :: Fx,Fy,Fz
657	+	real( kind = dp ) :: r,d2pha,phb,d2phb
658
659	+	integer :: id1,id2
660	+	integer  :: mytype_atom1
661	+	integer  :: mytype_atom2
662	+
663		!Local Variables
664
665	+	! write(*,*) "Frho: ", Frho(atom1)
666
667	+	phab = 0.0E0_DP
668	+	dvpdr = 0.0E0_DP
669	+	d2vpdrdr = 0.0E0_DP
670	+
671		if (rij .lt. EAM_rcut) then
672	+	#ifdef IS_MPI
673	+	!!!!! FIX ME
674	+	mytype_atom1 = atid_row(atom1)
675	+	#else
676	+	mytype_atom1 = atid(atom1)
677	+	#endif
678	+
679	+	drdx = d(1)/rij
680	+	drdy = d(2)/rij
681	+	drdz = d(3)/rij
682	+
683
684	<	r = dsqrt(rijsq)
685	<	efr(1,j) = -rxij
686	<	efr(2,j) = -ryij
687	<	efr(3,j) = -rzij
684	>	call eam_splint(EAMList%EAMParams(mytype_atom1)%eam_nr, &
685	>	EAMList%EAMParams(mytype_atom1)%eam_rvals, &
686	>	EAMList%EAMParams(mytype_atom1)%eam_rho_r, &
687	>	EAMList%EAMParams(mytype_atom1)%eam_rho_r_pp, &
688	>	rij, rha,drha,d2rha)
689
690	+	!! Calculate Phi(r) for atom1.
691	+	call eam_splint(EAMList%EAMParams(mytype_atom1)%eam_nr, &
692	+	EAMList%EAMParams(mytype_atom1)%eam_rvals, &
693	+	EAMList%EAMParams(mytype_atom1)%eam_phi_r, &
694	+	EAMList%EAMParams(mytype_atom1)%eam_phi_r_pp, &
695	+	rij, pha,dpha,d2pha)
696
697	<	call calc_eam_rho(r, rha, drha, d2rha, atype1)
698	<	call calc_eam_phi(r, pha, dpha, d2pha, atype1)
699	<	rci = eam_rcut(eam_atype_map(atype1))
700	<	#ifdef MPI
701	<	atype2 = ident_col(j)
697	>
698	>	! get cutoff for atom 1
699	>	rci = EAMList%EAMParams(mytype_atom1)%eam_rcut
700	>	#ifdef IS_MPI
701	>	mytype_atom2 = atid_col(atom2)
702		#else
703	<	atype2 = ident(j)
703	>	mytype_atom2 = atid(atom2)
704		#endif
705
706	<	call calc_eam_rho(r, rhb, drhb, d2rhb, atype2)
707	<	call calc_eam_phi(r, phb, dphb, d2phb, atype2)
708	<	rcj = eam_rcut(eam_atype_map(atype2))
706	>	! Calculate rho,drho and d2rho for atom1
707	>	call eam_splint(EAMList%EAMParams(mytype_atom2)%eam_nr, &
708	>	EAMList%EAMParams(mytype_atom2)%eam_rvals, &
709	>	EAMList%EAMParams(mytype_atom2)%eam_rho_r, &
710	>	EAMList%EAMParams(mytype_atom2)%eam_rho_r_pp, &
711	>	rij, rhb,drhb,d2rhb)
712
713	<	phab = 0.0E0_DP
714	<	dvpdr = 0.0E0_DP
715	<	d2vpdrdr = 0.0E0_DP
713	>	!! Calculate Phi(r) for atom2.
714	>	call eam_splint(EAMList%EAMParams(mytype_atom2)%eam_nr, &
715	>	EAMList%EAMParams(mytype_atom2)%eam_rvals, &
716	>	EAMList%EAMParams(mytype_atom2)%eam_phi_r, &
717	>	EAMList%EAMParams(mytype_atom2)%eam_phi_r_pp, &
718	>	rij, phb,dphb,d2phb)
719
720	<	if (r.lt.rci) then
720	>
721	>	! get type specific cutoff for atom 2
722	>	rcj = EAMList%EAMParams(mytype_atom1)%eam_rcut
723	>
724	>
725	>
726	>	if (rij.lt.rci) then
727		phab = phab + 0.5E0_DP*(rhb/rha)*pha
728		dvpdr = dvpdr + 0.5E0_DP*((rhb/rha)*dpha + &
729		pha*((drhb/rha) - (rhb*drha/rha/rha)))
#	Line 298 | Line 732 | contains
732		pha*((d2rhb/rha) - 2.0E0_DP*(drhb*drha/rha/rha) + &
733		(2.0E0_DP*rhb*drha*drha/rha/rha/rha) - (rhb*d2rha/rha/rha)))
734		endif
735	+
736
737	<
303	<	if (r.lt.rcj) then
737	>	if (rij.lt.rcj) then
738		phab = phab + 0.5E0_DP*(rha/rhb)*phb
739		dvpdr = dvpdr + 0.5E0_DP*((rha/rhb)*dphb + &
740		phb*((drha/rhb) - (rha*drhb/rhb/rhb)))
#	Line 309 | Line 743 | contains
743		phb*((d2rha/rhb) - 2.0E0_DP*(drha*drhb/rhb/rhb) + &
744		(2.0E0_DP*rha*drhb*drhb/rhb/rhb/rhb) - (rha*d2rhb/rhb/rhb)))
745		endif
746	<
313	<
314	<	#ifdef MPI
315	<
316	<	e_row(i) = e_row(i) + phab*0.5
317	<	e_col(i) = e_col(i) + phab*0.5
318	<	#else
319	<	if (do_pot) pot = pot + phab
320	<	#endif
321	<
746	>
747		drhoidr = drha
748		drhojdr = drhb
749
750		d2rhoidrdr = d2rha
751		d2rhojdrdr = d2rhb
752	<	#ifdef MPI
753	<	dudr = drhojdr*dfrhodrho_row(i)+drhoidr*dfrhodrho_col(j) &
752	>
753	>
754	>	#ifdef IS_MPI
755	>	dudr = drhojdr*dfrhodrho_row(atom1)+drhoidr*dfrhodrho_col(atom2) &
756		+ dvpdr
757
331	–	if (nmflag) then
332	–	d2 = d2vpdrdr + &
333	–	d2rhoidrdr*dfrhodrho_col(j) + &
334	–	d2rhojdrdr*dfrhodrho_row(i) + &
335	–	drhoidr*drhoidr*d2frhodrhodrho_col(j) + &
336	–	drhojdr*drhojdr*d2frhodrhodrho_row(i)
337	–	endif
758		#else
759	<	dudr = drhojdr*dfrhodrho(i)+drhoidr*dfrhodrho(j) &
759	>	dudr = drhojdr*dfrhodrho(atom1)+drhoidr*dfrhodrho(atom2) &
760		+ dvpdr
761	<
342	<	d2 = d2vpdrdr + &
343	<	d2rhoidrdr*dfrhodrho(j) + &
344	<	d2rhojdrdr*dfrhodrho(i) + &
345	<	drhoidr*drhoidr*d2frhodrhodrho(j) + &
346	<	drhojdr*drhojdr*d2frhodrhodrho(i)
761	>	! write(*,*) "Atom1,Atom2, dfrhodrho(atom1) dfrhodrho(atom2): ", atom1,atom2,dfrhodrho(atom1),dfrhodrho(atom2)
762		#endif
763
764	+	fx = dudr * drdx
765	+	fy = dudr * drdy
766	+	fz = dudr * drdz
767
350	–	do dim = 1, 3
768
769	<	drdx1 = efr(dim,j) / r
770	<	ftmp = dudr * drdx1
769	>	#ifdef IS_MPI
770	>	if (do_pot) then
771	>	pot_Row(atom1) = pot_Row(atom1) + phab*0.5
772	>	pot_Col(atom2) = pot_Col(atom2) + phab*0.5
773	>	end if
774
775	<	#ifdef MPI
776	<	f_col(dim,j) = f_col(dim,j) - ftmp
777	<	f_row(dim,i) = f_row(dim,i) + ftmp
775	>	f_Row(1,atom1) = f_Row(1,atom1) + fx
776	>	f_Row(2,atom1) = f_Row(2,atom1) + fy
777	>	f_Row(3,atom1) = f_Row(3,atom1) + fz
778	>
779	>	f_Col(1,atom2) = f_Col(1,atom2) - fx
780	>	f_Col(2,atom2) = f_Col(2,atom2) - fy
781	>	f_Col(3,atom2) = f_Col(3,atom2) - fz
782		#else
359	–	f(dim,j) = f(dim,j) - ftmp
360	–	f(dim,i) = f(dim,i) + ftmp
361	–	#endif
783
784	<	if (nmflag) then
785	<	idim = 3 * (i-1) + dim
786	<	jdim = 3 * (j-1) + dim
784	>	if(do_pot) then
785	>	pot = pot + phab
786	>	end if
787
788	<	do dim2 = 1, 3
788	>	f(1,atom1) = f(1,atom1) + fx
789	>	f(2,atom1) = f(2,atom1) + fy
790	>	f(3,atom1) = f(3,atom1) + fz
791	>
792	>	f(1,atom2) = f(1,atom2) - fx
793	>	f(2,atom2) = f(2,atom2) - fy
794	>	f(3,atom2) = f(3,atom2) - fz
795	>	#endif
796	>
797	>	if (nmflag) then
798
799	<	kt1 = d2 * efr(dim,j) * efr(dim2,j)/r/r
800	<	kt2 = - dudr * efr(dim,j) * efr(dim2,j)/r/r/r
799	>	drhoidr = drha
800	>	drhojdr = drhb
801	>	d2rhoidrdr = d2rha
802	>	d2rhojdrdr = d2rhb
803
804	<	if (dim.eq.dim2) then
805	<	kt3 = dudr / r
806	<	else
807	<	kt3 = 0.0E0_DP
808	<	endif
804	>	#ifdef IS_MPI
805	>	d2 = d2vpdrdr + &
806	>	d2rhoidrdr*dfrhodrho_col(atom2) + &
807	>	d2rhojdrdr*dfrhodrho_row(atom1) + &
808	>	drhoidr*drhoidr*d2frhodrhodrho_col(atom2) + &
809	>	drhojdr*drhojdr*d2frhodrhodrho_row(atom1)
810	>
811	>	#else
812
813	<	! The factor of 2 below is to compensate for
814	<	! overcounting.
815	<	! Mass weighting is done separately...
813	>	d2 = d2vpdrdr + &
814	>	d2rhoidrdr*dfrhodrho(atom2) + &
815	>	d2rhojdrdr*dfrhodrho(atom1) + &
816	>	drhoidr*drhoidr*d2frhodrhodrho(atom2) + &
817	>	drhojdr*drhojdr*d2frhodrhodrho(atom1)
818	>	#endif
819	>	end if
820
382	–	ktmp = (kt1+kt2+kt3)/2.0E0_DP
383	–	idim2 = 3 * (i-1) + dim2
384	–	jdim2 = 3 * (j-1) + dim2
821
822	<	d(idim,  idim2) = d(idim,idim2)  + ktmp
823	<	d(idim2, idim) = d(idim2,idim)   + ktmp
822	>
823	>
824	>	if (do_stress) then
825	>
826	>	#ifdef IS_MPI
827	>	id1 = tagRow(atom1)
828	>	id2 = tagColumn(atom2)
829	>	#else
830	>	id1 = atom1
831	>	id2 = atom2
832	>	#endif
833	>
834	>	if (molMembershipList(id1) .ne. molMembershipList(id2)) then
835
836	<	d(idim,  jdim2) = d(idim,jdim2)  - ktmp
837	<	d(idim2, jdim) = d(idim2,jdim)   - ktmp
838	<
839	<	d(jdim,  idim2) = d(jdim,idim2)  - ktmp
840	<	d(jdim2, idim) = d(jdim2,idim)   - ktmp
841	<
842	<	d(jdim,  jdim2) = d(jdim,jdim2)  + ktmp
843	<	d(jdim2, jdim) = d(jdim2,jdim)   + ktmp
844	<
398	<	enddo
399	<	endif
400	<	enddo
401	<
402	<	endif
403	<	enddo
404	<	endif
405	<
406	<	enddo
836	>	tau_Temp(1) = tau_Temp(1) - d(1) * fx
837	>	tau_Temp(2) = tau_Temp(2) - d(1) * fy
838	>	tau_Temp(3) = tau_Temp(3) - d(1) * fz
839	>	tau_Temp(4) = tau_Temp(4) - d(2) * fx
840	>	tau_Temp(5) = tau_Temp(5) - d(2) * fy
841	>	tau_Temp(6) = tau_Temp(6) - d(2) * fz
842	>	tau_Temp(7) = tau_Temp(7) - d(3) * fx
843	>	tau_Temp(8) = tau_Temp(8) - d(3) * fy
844	>	tau_Temp(9) = tau_Temp(9) - d(3) * fz
845
846	+	virial_Temp = virial_Temp + &
847	+	(tau_Temp(1) + tau_Temp(5) + tau_Temp(9))
848
849	+	endif
850	+	endif
851	+	endif
852
853	+
854	+	end subroutine do_eam_pair
855
411	–	end subroutine calc_eam_pair
856
857	<	subroutine calc_eam_rho(r, rho, drho, d2rho, atype)
857	>	subroutine eam_splint(nx, xa, ya, yppa, x, y, dy, d2y)
858
859	<	!  include 'headers/sizes.h'
859	>	integer :: atype, nx, j
860	>	real( kind = DP ), dimension(:) :: xa
861	>	real( kind = DP ), dimension(:) :: ya
862	>	real( kind = DP ), dimension(:) :: yppa
863	>	real( kind = DP ) :: x, y
864	>	real( kind = DP ) :: dy, d2y
865	>	real( kind = DP ) :: del, h, a, b, c, d
866	>	integer :: pp_arraySize
867
868	+
869	+	! this spline code assumes that the x points are equally spaced
870	+	! do not attempt to use this code if they are not.
871	+
872	+
873	+	! find the closest point with a value below our own:
874	+	j = FLOOR(real((nx-1),kind=dp) * (x - xa(1)) / (xa(nx) - xa(1))) + 1
875
876	<	integer atype, etype, number_r
877	<	real( kind = DP )  :: r, rho, drho, d2rho
878	<	integer :: i
876	>	! check to make sure we're inside the spline range:
877	>	if ((j.gt.nx).or.(j.lt.1)) then
878	>	write(errMSG,*) "EAM_splint: x is outside bounds of spline"
879	>	call handleError(routineName,errMSG)
880	>	endif
881	>	! check to make sure we haven't screwed up the calculation of j:
882	>	if ((x.lt.xa(j)).or.(x.gt.xa(j+1))) then
883	>	if (j.ne.nx) then
884	>	write(errMSG,*) "EAM_splint:",x," x is outside bounding range"
885	>	call handleError(routineName,errMSG)
886	>	endif
887	>	endif
888
889	+	del = xa(j+1) - x
890	+	h = xa(j+1) - xa(j)
891	+
892	+	a = del / h
893	+	b = 1.0E0_DP - a
894	+	c = a*(a*a - 1.0E0_DP)*h*h/6.0E0_DP
895	+	d = b*(b*b - 1.0E0_DP)*h*h/6.0E0_DP
896	+
897	+	y = a*ya(j) + b*ya(j+1) + c*yppa(j) + d*yppa(j+1)
898	+
899	+	dy = (ya(j+1)-ya(j))/h &
900	+	- (3.0E0_DP*a*a - 1.0E0_DP)*h*yppa(j)/6.0E0_DP &
901	+	+ (3.0E0_DP*b*b - 1.0E0_DP)*h*yppa(j+1)/6.0E0_DP
902	+
903	+
904	+	d2y = a*yppa(j) + b*yppa(j+1)
905	+
906
907	<	etype = eam_atype_map(atype)
907	>	end subroutine eam_splint
908
425	–	if (r.lt.eam_rcut(etype)) then
426	–	number_r = eam_nr(etype)
427	–	call eam_splint(etype, number_r, eam_rvals, eam_rho_r, &
428	–	eam_rho_r_pp, r, rho, drho, d2rho)
429	–	else
430	–	rho = 0.0E0_DP
431	–	drho = 0.0E0_DP
432	–	d2rho = 0.0E0_DP
433	–	endif
909
910	<	return
436	<	end subroutine calc_eam_rho
910	>	subroutine eam_spline(nx, xa, ya, yppa, yp1, ypn, boundary)
911
438	–	subroutine calc_eam_frho(dens, u, u1, u2, atype)
912
913	<	! include 'headers/sizes.h'
913	>	! yp1 and ypn are the first derivatives of y at the two endpoints
914	>	! if boundary is 'L' the lower derivative is used
915	>	! if boundary is 'U' the upper derivative is used
916	>	! if boundary is 'B' then both derivatives are used
917	>	! if boundary is anything else, then both derivatives are assumed to be 0
918	>
919	>	integer :: nx, i, k, max_array_size
920	>
921	>	real( kind = DP ), dimension(:)        :: xa
922	>	real( kind = DP ), dimension(:)        :: ya
923	>	real( kind = DP ), dimension(:)        :: yppa
924	>	real( kind = DP ), dimension(size(xa)) :: u
925	>	real( kind = DP ) :: yp1,ypn,un,qn,sig,p
926	>	character(len=*) :: boundary
927	>
928	>	! make sure the sizes match
929	>	if ((nx /= size(xa)) .or. (nx /= size(ya))) then
930	>	call handleWarning("EAM_SPLINE","Array size mismatch")
931	>	end if
932
933	<	integer atype, etype, number_rho
934	<	real( kind = DP ) :: dens, u, u1, u2
935	<	real( kind = DP ) :: rho_vals
933	>	if ((boundary.eq.'l').or.(boundary.eq.'L').or. &
934	>	(boundary.eq.'b').or.(boundary.eq.'B')) then
935	>	yppa(1) = -0.5E0_DP
936	>	u(1) = (3.0E0_DP/(xa(2)-xa(1)))*((ya(2)-&
937	>	ya(1))/(xa(2)-xa(1))-yp1)
938	>	else
939	>	yppa(1) = 0.0E0_DP
940	>	u(1)  = 0.0E0_DP
941	>	endif
942	>
943	>	do i = 2, nx - 1
944	>	sig = (xa(i) - xa(i-1)) / (xa(i+1) - xa(i-1))
945	>	p = sig * yppa(i-1) + 2.0E0_DP
946	>	yppa(i) = (sig - 1.0E0_DP) / p
947	>	u(i) = (6.0E0_DP*((ya(i+1)-ya(i))/(xa(i+1)-xa(i)) - &
948	>	(ya(i)-ya(i-1))/(xa(i)-xa(i-1)))/ &
949	>	(xa(i+1)-xa(i-1)) - sig * u(i-1))/p
950	>	enddo
951	>
952	>	if ((boundary.eq.'u').or.(boundary.eq.'U').or. &
953	>	(boundary.eq.'b').or.(boundary.eq.'B')) then
954	>	qn = 0.5E0_DP
955	>	un = (3.0E0_DP/(xa(nx)-xa(nx-1)))* &
956	>	(ypn-(ya(nx)-ya(nx-1))/(xa(nx)-xa(nx-1)))
957	>	else
958	>	qn = 0.0E0_DP
959	>	un = 0.0E0_DP
960	>	endif
961
962	<	etype = eam_atype_map(atype)
963	<	number_rho = eam_nrho(etype)
964	<	if (dens.lt.eam_rhovals(number_rho, etype)) then
965	<	call eam_splint(etype, number_rho, eam_rhovals, eam_f_rho, &
966	<	eam_f_rho_pp, dens, u, u1, u2)
451	<	else
452	<	rho_vals = eam_rhovals(number_rho,etype)
453	<	call eam_splint(etype, number_rho, eam_rhovals, eam_f_rho, &
454	<	eam_f_rho_pp, rho_vals, u, u1, u2)
455	<	endif
962	>	yppa(nx)=(un-qn*u(nx-1))/(qn*yppa(nx-1)+1.0E0_DP)
963	>
964	>	do k = nx-1, 1, -1
965	>	yppa(k)=yppa(k)*yppa(k+1)+u(k)
966	>	enddo
967
968	<	return
458	<	end subroutine calc_eam_frho
968	>	end subroutine eam_spline
969
460	–	subroutine calc_eam_phi(r, phi, dphi, d2phi, atype)
970
971
972
973	<
465	<	integer atype, etype, number_r
466	<	real( kind = DP ) :: r, phi, dphi, d2phi
467	<
468	<	etype = eam_atype_map(atype)
469	<
470	<	if (r.lt.eam_rcut(etype)) then
471	<	number_r = eam_nr(etype)
472	<	call eam_splint(etype, number_r, eam_rvals, eam_phi_r, &
473	<	eam_phi_r_pp, r, phi, dphi, d2phi)
474	<	else
475	<	phi = 0.0E0_DP
476	<	dphi = 0.0E0_DP
477	<	d2phi = 0.0E0_DP
478	<	endif
479	<
480	<	return
481	<	end subroutine calc_eam_phi
482	<
483	<
484	<	subroutine eam_splint(atype, nx, xa, ya, yppa, x, y, dy, d2y)
485	<
486	<	!  include 'headers/sizes.h'
487	<
488	<	real( kind = DP ), dimension(:,:) :: xa
489	<	real( kind = DP ), dimension(:,:) :: ya
490	<	real( kind = DP ), dimension(:,:) :: yppa
491	<	real( kind = DP ) :: x, y, dy, d2y
492	<	real( kind = DP ) :: del, h, a, b, c, d
493	<
494	<
495	<	integer atype, nx, j
496	<
497	<
498	<	! this spline code assumes that the x points are equally spaced
499	<	! do not attempt to use this code if they are not.
500	<
501	<
502	<	! find the closest point with a value below our own:
503	<	j = FLOOR(dble(nx-1) * (x - xa(1,atype)) / (xa(nx,atype) - xa(1,atype))) + 1
504	<
505	<	! check to make sure we're inside the spline range:
506	<	if ((j.gt.nx).or.(j.lt.1)) call error('eam_splint', &
507	<	'x is outside bounds of spline')
508	<
509	<	! check to make sure we haven't screwed up the calculation of j:
510	<	if ((x.lt.xa(j,atype)).or.(x.gt.xa(j+1,atype))) then
511	<	if (j.ne.nx) then
512	<	call error('eam_splint', &
513	<	'x is outside bounding range')
514	<	endif
515	<	endif
516	<
517	<	del = xa(j+1,atype) - x
518	<	h = xa(j+1,atype) - xa(j,atype)
519	<
520	<	a = del / h
521	<	b = 1.0E0_DP - a
522	<	c = a*(a*a - 1.0E0_DP)*h*h/6.0E0_DP
523	<	d = b*(b*b - 1.0E0_DP)*h*h/6.0E0_DP
524	<
525	<	y = a*ya(j,atype) + b*ya(j+1,atype) + c*yppa(j,atype) + d*yppa(j+1,atype)
526	<
527	<	dy = (ya(j+1,atype)-ya(j,atype))/h &
528	<	- (3.0E0_DP*a*a - 1.0E0_DP)*h*yppa(j,atype)/6.0E0_DP &
529	<	+ (3.0E0_DP*b*b - 1.0E0_DP)*h*yppa(j+1,atype)/6.0E0_DP
530	<
531	<	d2y = a*yppa(j,atype) + b*yppa(j+1,atype)
532	<
533	<	return
534	<	end subroutine eam_splint
535	<
536	<	subroutine eam_spline(atype, nx, xa, ya, yppa, yp1, ypn, boundary)
537	<
538	<	!  include 'headers/sizes.h'
539	<
540	<
541	<	! yp1 and ypn are the first derivatives of y at the two endpoints
542	<	! if boundary is 'L' the lower derivative is used
543	<	! if boundary is 'U' the upper derivative is used
544	<	! if boundary is 'B' then both derivatives are used
545	<	! if boundary is anything else, then both derivatives are assumed to be 0
546	<
547	<	integer nx, i, k, atype, max_array_size
548	<
549	<	real( kind = DP ), dimension(:,:) :: xa
550	<	real( kind = DP ), dimension(:,:) :: ya
551	<	real( kind = DP ), dimension(:,:) :: yppa
552	<	real( kind = DP ), allocatable, dimension(:) :: u
553	<	real( kind = DP ) :: yp1,ypn,un,qn,sig,p
554	<	character boundary
555	<
556	<	max_array_size = size(xa,1)
557	<	allocate(u(max_array_size))
558	<
559	<
560	<	if ((boundary.eq.'l').or.(boundary.eq.'L').or. &
561	<	(boundary.eq.'b').or.(boundary.eq.'B')) then
562	<	yppa(1, atype) = -0.5E0_DP
563	<	u(1) = (3.0E0_DP/(xa(2,atype)-xa(1,atype)))*((ya(2,atype)-&
564	<	ya(1,atype))/(xa(2,atype)-xa(1,atype))-yp1)
565	<	else
566	<	yppa(1,atype) = 0.0E0_DP
567	<	u(1)  = 0.0E0_DP
568	<	endif
569	<
570	<	do i = 2, nx - 1
571	<	sig = (xa(i,atype) - xa(i-1,atype)) / (xa(i+1,atype) - xa(i-1,atype))
572	<	p = sig * yppa(i-1,atype) + 2.0E0_DP
573	<	yppa(i,atype) = (sig - 1.0E0_DP) / p
574	<	u(i) = (6.0E0_DP*((ya(i+1,atype)-ya(i,atype))/(xa(i+1,atype)-xa(i,atype)) - &
575	<	(ya(i,atype)-ya(i-1,atype))/(xa(i,atype)-xa(i-1,atype)))/ &
576	<	(xa(i+1,atype)-xa(i-1,atype)) - sig * u(i-1))/p
577	<	enddo
578	<
579	<	if ((boundary.eq.'u').or.(boundary.eq.'U').or. &
580	<	(boundary.eq.'b').or.(boundary.eq.'B')) then
581	<	qn = 0.5E0_DP
582	<	un = (3.0E0_DP/(xa(nx,atype)-xa(nx-1,atype)))* &
583	<	(ypn-(ya(nx,atype)-ya(nx-1,atype))/(xa(nx,atype)-xa(nx-1,atype)))
584	<	else
585	<	qn = 0.0E0_DP
586	<	un = 0.0E0_DP
587	<	endif
588	<
589	<	yppa(nx,atype)=(un-qn*u(nx-1))/(qn*yppa(nx-1,atype)+1.0E0_DP)
590	<
591	<	do k = nx-1, 1, -1
592	<	yppa(k,atype)=yppa(k,atype)*yppa(k+1,atype)+u(k)
593	<	enddo
594	<
595	<	deallocate(u)
596	<	return
597	<	end subroutine eam_spline
598	<
599	<
600	<
601	<
602	<	end module calc_eam
973	>	end module eam

Diff Legend

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