1 |
!! |
2 |
!! Copyright (c) 2005, 2009 The University of Notre Dame. All Rights Reserved. |
3 |
!! |
4 |
!! The University of Notre Dame grants you ("Licensee") a |
5 |
!! non-exclusive, royalty free, license to use, modify and |
6 |
!! redistribute this software in source and binary code form, provided |
7 |
!! that the following conditions are met: |
8 |
!! |
9 |
!! 1. Redistributions of source code must retain the above copyright |
10 |
!! notice, this list of conditions and the following disclaimer. |
11 |
!! |
12 |
!! 2. Redistributions in binary form must reproduce the above copyright |
13 |
!! notice, this list of conditions and the following disclaimer in the |
14 |
!! documentation and/or other materials provided with the |
15 |
!! distribution. |
16 |
!! |
17 |
!! This software is provided "AS IS," without a warranty of any |
18 |
!! kind. All express or implied conditions, representations and |
19 |
!! warranties, including any implied warranty of merchantability, |
20 |
!! fitness for a particular purpose or non-infringement, are hereby |
21 |
!! excluded. The University of Notre Dame and its licensors shall not |
22 |
!! be liable for any damages suffered by licensee as a result of |
23 |
!! using, modifying or distributing the software or its |
24 |
!! derivatives. In no event will the University of Notre Dame or its |
25 |
!! licensors be liable for any lost revenue, profit or data, or for |
26 |
!! direct, indirect, special, consequential, incidental or punitive |
27 |
!! damages, however caused and regardless of the theory of liability, |
28 |
!! arising out of the use of or inability to use software, even if the |
29 |
!! University of Notre Dame has been advised of the possibility of |
30 |
!! such damages. |
31 |
!! |
32 |
!! SUPPORT OPEN SCIENCE! If you use OpenMD or its source code in your |
33 |
!! research, please cite the appropriate papers when you publish your |
34 |
!! work. Good starting points are: |
35 |
!! |
36 |
!! [1] Meineke, et al., J. Comp. Chem. 26, 252-271 (2005). |
37 |
!! [2] Fennell & Gezelter, J. Chem. Phys. 124, 234104 (2006). |
38 |
!! [3] Sun, Lin & Gezelter, J. Chem. Phys. 128, 24107 (2008). |
39 |
!! [4] Vardeman & Gezelter, in progress (2009). |
40 |
!! |
41 |
|
42 |
!! doForces.F90 |
43 |
!! module doForces |
44 |
!! Calculates Long Range forces. |
45 |
|
46 |
!! @author Charles F. Vardeman II |
47 |
!! @author Matthew Meineke |
48 |
!! @version $Id$, $Date$, $Name: not supported by cvs2svn $, $Revision$ |
49 |
|
50 |
|
51 |
module doForces |
52 |
use force_globals |
53 |
use fForceOptions |
54 |
use simulation |
55 |
use definitions |
56 |
use atype_module |
57 |
use switcheroo |
58 |
use neighborLists |
59 |
use electrostatic_module |
60 |
use shapes |
61 |
use vector_class |
62 |
use MetalNonMetal |
63 |
use status |
64 |
use ISO_C_BINDING |
65 |
|
66 |
#ifdef IS_MPI |
67 |
use mpiSimulation |
68 |
#endif |
69 |
|
70 |
implicit none |
71 |
PRIVATE |
72 |
|
73 |
real(kind=dp), external :: getSigma |
74 |
real(kind=dp), external :: getEpsilon |
75 |
real(kind=dp), external :: getEAMcut |
76 |
real(kind=dp), external :: getGayBerneCut |
77 |
real(kind=dp), external :: getStickyCut |
78 |
real(kind=dp), external :: getSCCut |
79 |
|
80 |
|
81 |
#define __FORTRAN90 |
82 |
#include "UseTheForce/fCutoffPolicy.h" |
83 |
#include "UseTheForce/DarkSide/fInteractionMap.h" |
84 |
#include "UseTheForce/DarkSide/fElectrostaticSummationMethod.h" |
85 |
|
86 |
INTEGER, PARAMETER:: PREPAIR_LOOP = 1 |
87 |
INTEGER, PARAMETER:: PAIR_LOOP = 2 |
88 |
|
89 |
logical, save :: haveNeighborList = .false. |
90 |
logical, save :: haveSIMvariables = .false. |
91 |
logical, save :: haveSaneForceField = .false. |
92 |
logical, save :: haveInteractionHash = .false. |
93 |
logical, save :: haveGtypeCutoffMap = .false. |
94 |
logical, save :: haveDefaultCutoffs = .false. |
95 |
logical, save :: haveSkinThickness = .false. |
96 |
logical, save :: haveElectrostaticSummationMethod = .false. |
97 |
logical, save :: haveCutoffPolicy = .false. |
98 |
logical, save :: VisitCutoffsAfterComputing = .false. |
99 |
logical, save :: do_box_dipole = .false. |
100 |
|
101 |
logical, save :: FF_uses_DirectionalAtoms |
102 |
logical, save :: FF_uses_Dipoles |
103 |
logical, save :: FF_uses_GayBerne |
104 |
logical, save :: FF_uses_EAM |
105 |
logical, save :: FF_uses_SC |
106 |
logical, save :: FF_uses_MNM |
107 |
|
108 |
|
109 |
logical, save :: SIM_uses_DirectionalAtoms |
110 |
logical, save :: SIM_uses_EAM |
111 |
logical, save :: SIM_uses_SC |
112 |
logical, save :: SIM_uses_MNM |
113 |
logical, save :: SIM_requires_postpair_calc |
114 |
logical, save :: SIM_requires_prepair_calc |
115 |
logical, save :: SIM_uses_PBC |
116 |
logical, save :: SIM_uses_AtomicVirial |
117 |
|
118 |
integer, save :: electrostaticSummationMethod |
119 |
integer, save :: cutoffPolicy = TRADITIONAL_CUTOFF_POLICY |
120 |
|
121 |
real(kind=dp), save :: defaultRcut, defaultRsw, largestRcut |
122 |
real(kind=dp), save :: skinThickness |
123 |
logical, save :: defaultDoShiftPot |
124 |
logical, save :: defaultDoShiftFrc |
125 |
|
126 |
public :: init_FF |
127 |
public :: setCutoffs |
128 |
public :: cWasLame |
129 |
public :: setElectrostaticMethod |
130 |
public :: setBoxDipole |
131 |
public :: getBoxDipole |
132 |
public :: setCutoffPolicy |
133 |
public :: setSkinThickness |
134 |
public :: do_force_loop |
135 |
|
136 |
#ifdef PROFILE |
137 |
public :: getforcetime |
138 |
real, save :: forceTime = 0 |
139 |
real :: forceTimeInitial, forceTimeFinal |
140 |
integer :: nLoops |
141 |
#endif |
142 |
|
143 |
!! Variables for cutoff mapping and interaction mapping |
144 |
! Bit hash to determine pair-pair interactions. |
145 |
integer, dimension(:,:), allocatable :: InteractionHash |
146 |
real(kind=dp), dimension(:), allocatable :: atypeMaxCutoff |
147 |
real(kind=dp), dimension(:), allocatable, target :: groupMaxCutoffRow |
148 |
real(kind=dp), dimension(:), pointer :: groupMaxCutoffCol |
149 |
|
150 |
integer, dimension(:), allocatable, target :: groupToGtypeRow |
151 |
integer, dimension(:), pointer :: groupToGtypeCol => null() |
152 |
|
153 |
real(kind=dp), dimension(:), allocatable,target :: gtypeMaxCutoffRow |
154 |
real(kind=dp), dimension(:), pointer :: gtypeMaxCutoffCol |
155 |
type ::gtypeCutoffs |
156 |
real(kind=dp) :: rcut |
157 |
real(kind=dp) :: rcutsq |
158 |
real(kind=dp) :: rlistsq |
159 |
end type gtypeCutoffs |
160 |
type(gtypeCutoffs), dimension(:,:), allocatable :: gtypeCutoffMap |
161 |
|
162 |
real(kind=dp), dimension(3) :: boxDipole |
163 |
|
164 |
contains |
165 |
|
166 |
subroutine createInteractionHash() |
167 |
integer :: nAtypes |
168 |
integer :: i |
169 |
integer :: j |
170 |
integer :: iHash |
171 |
!! Test Types |
172 |
logical :: i_is_LJ |
173 |
logical :: i_is_Elect |
174 |
logical :: i_is_Sticky |
175 |
logical :: i_is_StickyP |
176 |
logical :: i_is_GB |
177 |
logical :: i_is_EAM |
178 |
logical :: i_is_Shape |
179 |
logical :: i_is_SC |
180 |
logical :: j_is_LJ |
181 |
logical :: j_is_Elect |
182 |
logical :: j_is_Sticky |
183 |
logical :: j_is_StickyP |
184 |
logical :: j_is_GB |
185 |
logical :: j_is_EAM |
186 |
logical :: j_is_Shape |
187 |
logical :: j_is_SC |
188 |
real(kind=dp) :: myRcut |
189 |
|
190 |
if (.not. associated(atypes)) then |
191 |
call handleError("doForces", "atypes was not present before call of createInteractionHash!") |
192 |
return |
193 |
endif |
194 |
|
195 |
nAtypes = getSize(atypes) |
196 |
|
197 |
if (nAtypes == 0) then |
198 |
call handleError("doForces", "nAtypes was zero during call of createInteractionHash!") |
199 |
return |
200 |
end if |
201 |
|
202 |
if (.not. allocated(InteractionHash)) then |
203 |
allocate(InteractionHash(nAtypes,nAtypes)) |
204 |
else |
205 |
deallocate(InteractionHash) |
206 |
allocate(InteractionHash(nAtypes,nAtypes)) |
207 |
endif |
208 |
|
209 |
if (.not. allocated(atypeMaxCutoff)) then |
210 |
allocate(atypeMaxCutoff(nAtypes)) |
211 |
else |
212 |
deallocate(atypeMaxCutoff) |
213 |
allocate(atypeMaxCutoff(nAtypes)) |
214 |
endif |
215 |
|
216 |
do i = 1, nAtypes |
217 |
call getElementProperty(atypes, i, "is_LennardJones", i_is_LJ) |
218 |
call getElementProperty(atypes, i, "is_Electrostatic", i_is_Elect) |
219 |
call getElementProperty(atypes, i, "is_Sticky", i_is_Sticky) |
220 |
call getElementProperty(atypes, i, "is_StickyPower", i_is_StickyP) |
221 |
call getElementProperty(atypes, i, "is_GayBerne", i_is_GB) |
222 |
call getElementProperty(atypes, i, "is_EAM", i_is_EAM) |
223 |
call getElementProperty(atypes, i, "is_Shape", i_is_Shape) |
224 |
call getElementProperty(atypes, i, "is_SC", i_is_SC) |
225 |
|
226 |
do j = i, nAtypes |
227 |
|
228 |
iHash = 0 |
229 |
myRcut = 0.0_dp |
230 |
|
231 |
call getElementProperty(atypes, j, "is_LennardJones", j_is_LJ) |
232 |
call getElementProperty(atypes, j, "is_Electrostatic", j_is_Elect) |
233 |
call getElementProperty(atypes, j, "is_Sticky", j_is_Sticky) |
234 |
call getElementProperty(atypes, j, "is_StickyPower", j_is_StickyP) |
235 |
call getElementProperty(atypes, j, "is_GayBerne", j_is_GB) |
236 |
call getElementProperty(atypes, j, "is_EAM", j_is_EAM) |
237 |
call getElementProperty(atypes, j, "is_Shape", j_is_Shape) |
238 |
call getElementProperty(atypes, j, "is_SC", j_is_SC) |
239 |
|
240 |
if (i_is_LJ .and. j_is_LJ) then |
241 |
iHash = ior(iHash, LJ_PAIR) |
242 |
endif |
243 |
|
244 |
if (i_is_Elect .and. j_is_Elect) then |
245 |
iHash = ior(iHash, ELECTROSTATIC_PAIR) |
246 |
endif |
247 |
|
248 |
if (i_is_Sticky .and. j_is_Sticky) then |
249 |
iHash = ior(iHash, STICKY_PAIR) |
250 |
endif |
251 |
|
252 |
if (i_is_StickyP .and. j_is_StickyP) then |
253 |
iHash = ior(iHash, STICKYPOWER_PAIR) |
254 |
endif |
255 |
|
256 |
if (i_is_EAM .and. j_is_EAM) then |
257 |
iHash = ior(iHash, EAM_PAIR) |
258 |
endif |
259 |
|
260 |
if (i_is_SC .and. j_is_SC) then |
261 |
iHash = ior(iHash, SC_PAIR) |
262 |
endif |
263 |
|
264 |
if (i_is_GB .and. j_is_GB) iHash = ior(iHash, GAYBERNE_PAIR) |
265 |
if (i_is_GB .and. j_is_LJ) iHash = ior(iHash, GAYBERNE_LJ) |
266 |
if (i_is_LJ .and. j_is_GB) iHash = ior(iHash, GAYBERNE_LJ) |
267 |
|
268 |
if ((i_is_EAM.or.i_is_SC).and.(.not.(j_is_EAM.or.j_is_SC))) iHash = ior(iHash, MNM_PAIR) |
269 |
if ((j_is_EAM.or.j_is_SC).and.(.not.(i_is_EAM.or.i_is_SC))) iHash = ior(iHash, MNM_PAIR) |
270 |
|
271 |
if (i_is_Shape .and. j_is_Shape) iHash = ior(iHash, SHAPE_PAIR) |
272 |
if (i_is_Shape .and. j_is_LJ) iHash = ior(iHash, SHAPE_LJ) |
273 |
if (i_is_LJ .and. j_is_Shape) iHash = ior(iHash, SHAPE_LJ) |
274 |
|
275 |
|
276 |
InteractionHash(i,j) = iHash |
277 |
InteractionHash(j,i) = iHash |
278 |
|
279 |
end do |
280 |
|
281 |
end do |
282 |
|
283 |
haveInteractionHash = .true. |
284 |
end subroutine createInteractionHash |
285 |
|
286 |
subroutine createGtypeCutoffMap() |
287 |
|
288 |
logical :: i_is_LJ |
289 |
logical :: i_is_Elect |
290 |
logical :: i_is_Sticky |
291 |
logical :: i_is_StickyP |
292 |
logical :: i_is_GB |
293 |
logical :: i_is_EAM |
294 |
logical :: i_is_Shape |
295 |
logical :: i_is_SC |
296 |
logical :: GtypeFound |
297 |
|
298 |
integer :: myStatus, nAtypes, i, j, istart, iend, jstart, jend |
299 |
integer :: n_in_i, me_i, ia, g, atom1, ja, n_in_j,me_j |
300 |
integer :: nGroupsInRow |
301 |
integer :: nGroupsInCol |
302 |
integer :: nGroupTypesRow,nGroupTypesCol |
303 |
real(kind=dp):: thisSigma, bigSigma, thisRcut, tradRcut, tol |
304 |
real(kind=dp) :: biggestAtypeCutoff |
305 |
integer :: c_ident_i |
306 |
|
307 |
if (.not. haveInteractionHash) then |
308 |
call createInteractionHash() |
309 |
endif |
310 |
#ifdef IS_MPI |
311 |
nGroupsInRow = getNgroupsInRow(plan_group_row) |
312 |
nGroupsInCol = getNgroupsInCol(plan_group_col) |
313 |
#endif |
314 |
nAtypes = getSize(atypes) |
315 |
! Set all of the initial cutoffs to zero. |
316 |
atypeMaxCutoff = 0.0_dp |
317 |
biggestAtypeCutoff = 0.0_dp |
318 |
do i = 1, nAtypes |
319 |
if (SimHasAtype(i)) then |
320 |
call getElementProperty(atypes, i, "is_LennardJones", i_is_LJ) |
321 |
call getElementProperty(atypes, i, "is_Electrostatic", i_is_Elect) |
322 |
call getElementProperty(atypes, i, "is_Sticky", i_is_Sticky) |
323 |
call getElementProperty(atypes, i, "is_StickyPower", i_is_StickyP) |
324 |
call getElementProperty(atypes, i, "is_GayBerne", i_is_GB) |
325 |
call getElementProperty(atypes, i, "is_EAM", i_is_EAM) |
326 |
call getElementProperty(atypes, i, "is_Shape", i_is_Shape) |
327 |
call getElementProperty(atypes, i, "is_SC", i_is_SC) |
328 |
call getElementProperty(atypes, i, "c_ident", c_ident_i) |
329 |
|
330 |
if (haveDefaultCutoffs) then |
331 |
atypeMaxCutoff(i) = defaultRcut |
332 |
else |
333 |
if (i_is_LJ) then |
334 |
thisRcut = getSigma(c_ident_i) * 2.5_dp |
335 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
336 |
endif |
337 |
if (i_is_Elect) then |
338 |
thisRcut = defaultRcut |
339 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
340 |
endif |
341 |
if (i_is_Sticky) then |
342 |
thisRcut = getStickyCut(c_ident_i) |
343 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
344 |
endif |
345 |
if (i_is_StickyP) then |
346 |
thisRcut = getStickyCut(c_ident_i) |
347 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
348 |
endif |
349 |
if (i_is_GB) then |
350 |
thisRcut = getGayBerneCut(c_ident_i) |
351 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
352 |
endif |
353 |
if (i_is_EAM) then |
354 |
thisRcut = getEAMCut(c_ident_i) |
355 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
356 |
endif |
357 |
if (i_is_Shape) then |
358 |
thisRcut = getShapeCut(i) |
359 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
360 |
endif |
361 |
if (i_is_SC) then |
362 |
thisRcut = getSCCut(c_ident_i) |
363 |
if (thisRCut .gt. atypeMaxCutoff(i)) atypeMaxCutoff(i) = thisRCut |
364 |
endif |
365 |
endif |
366 |
|
367 |
if (atypeMaxCutoff(i).gt.biggestAtypeCutoff) then |
368 |
biggestAtypeCutoff = atypeMaxCutoff(i) |
369 |
endif |
370 |
|
371 |
endif |
372 |
enddo |
373 |
|
374 |
istart = 1 |
375 |
jstart = 1 |
376 |
#ifdef IS_MPI |
377 |
iend = nGroupsInRow |
378 |
jend = nGroupsInCol |
379 |
#else |
380 |
iend = nGroups |
381 |
jend = nGroups |
382 |
#endif |
383 |
|
384 |
!! allocate the groupToGtype and gtypeMaxCutoff here. |
385 |
if(.not.allocated(groupToGtypeRow)) then |
386 |
! allocate(groupToGtype(iend)) |
387 |
allocate(groupToGtypeRow(iend)) |
388 |
else |
389 |
deallocate(groupToGtypeRow) |
390 |
allocate(groupToGtypeRow(iend)) |
391 |
endif |
392 |
if(.not.allocated(groupMaxCutoffRow)) then |
393 |
allocate(groupMaxCutoffRow(iend)) |
394 |
else |
395 |
deallocate(groupMaxCutoffRow) |
396 |
allocate(groupMaxCutoffRow(iend)) |
397 |
end if |
398 |
|
399 |
if(.not.allocated(gtypeMaxCutoffRow)) then |
400 |
allocate(gtypeMaxCutoffRow(iend)) |
401 |
else |
402 |
deallocate(gtypeMaxCutoffRow) |
403 |
allocate(gtypeMaxCutoffRow(iend)) |
404 |
endif |
405 |
|
406 |
|
407 |
#ifdef IS_MPI |
408 |
! We only allocate new storage if we are in MPI because Ncol /= Nrow |
409 |
if(.not.associated(groupToGtypeCol)) then |
410 |
allocate(groupToGtypeCol(jend)) |
411 |
else |
412 |
deallocate(groupToGtypeCol) |
413 |
allocate(groupToGtypeCol(jend)) |
414 |
end if |
415 |
|
416 |
if(.not.associated(groupMaxCutoffCol)) then |
417 |
allocate(groupMaxCutoffCol(jend)) |
418 |
else |
419 |
deallocate(groupMaxCutoffCol) |
420 |
allocate(groupMaxCutoffCol(jend)) |
421 |
end if |
422 |
if(.not.associated(gtypeMaxCutoffCol)) then |
423 |
allocate(gtypeMaxCutoffCol(jend)) |
424 |
else |
425 |
deallocate(gtypeMaxCutoffCol) |
426 |
allocate(gtypeMaxCutoffCol(jend)) |
427 |
end if |
428 |
|
429 |
groupMaxCutoffCol = 0.0_dp |
430 |
gtypeMaxCutoffCol = 0.0_dp |
431 |
|
432 |
#endif |
433 |
groupMaxCutoffRow = 0.0_dp |
434 |
gtypeMaxCutoffRow = 0.0_dp |
435 |
|
436 |
|
437 |
!! first we do a single loop over the cutoff groups to find the |
438 |
!! largest cutoff for any atypes present in this group. We also |
439 |
!! create gtypes at this point. |
440 |
|
441 |
tol = 1.0e-6_dp |
442 |
nGroupTypesRow = 0 |
443 |
nGroupTypesCol = 0 |
444 |
do i = istart, iend |
445 |
n_in_i = groupStartRow(i+1) - groupStartRow(i) |
446 |
groupMaxCutoffRow(i) = 0.0_dp |
447 |
do ia = groupStartRow(i), groupStartRow(i+1)-1 |
448 |
atom1 = groupListRow(ia) |
449 |
#ifdef IS_MPI |
450 |
me_i = atid_row(atom1) |
451 |
#else |
452 |
me_i = atid(atom1) |
453 |
#endif |
454 |
if (atypeMaxCutoff(me_i).gt.groupMaxCutoffRow(i)) then |
455 |
groupMaxCutoffRow(i)=atypeMaxCutoff(me_i) |
456 |
endif |
457 |
enddo |
458 |
if (nGroupTypesRow.eq.0) then |
459 |
nGroupTypesRow = nGroupTypesRow + 1 |
460 |
gtypeMaxCutoffRow(nGroupTypesRow) = groupMaxCutoffRow(i) |
461 |
groupToGtypeRow(i) = nGroupTypesRow |
462 |
else |
463 |
GtypeFound = .false. |
464 |
do g = 1, nGroupTypesRow |
465 |
if ( abs(groupMaxCutoffRow(i) - gtypeMaxCutoffRow(g)).lt.tol) then |
466 |
groupToGtypeRow(i) = g |
467 |
GtypeFound = .true. |
468 |
endif |
469 |
enddo |
470 |
if (.not.GtypeFound) then |
471 |
nGroupTypesRow = nGroupTypesRow + 1 |
472 |
gtypeMaxCutoffRow(nGroupTypesRow) = groupMaxCutoffRow(i) |
473 |
groupToGtypeRow(i) = nGroupTypesRow |
474 |
endif |
475 |
endif |
476 |
enddo |
477 |
|
478 |
#ifdef IS_MPI |
479 |
do j = jstart, jend |
480 |
n_in_j = groupStartCol(j+1) - groupStartCol(j) |
481 |
groupMaxCutoffCol(j) = 0.0_dp |
482 |
do ja = groupStartCol(j), groupStartCol(j+1)-1 |
483 |
atom1 = groupListCol(ja) |
484 |
|
485 |
me_j = atid_col(atom1) |
486 |
|
487 |
if (atypeMaxCutoff(me_j).gt.groupMaxCutoffCol(j)) then |
488 |
groupMaxCutoffCol(j)=atypeMaxCutoff(me_j) |
489 |
endif |
490 |
enddo |
491 |
|
492 |
if (nGroupTypesCol.eq.0) then |
493 |
nGroupTypesCol = nGroupTypesCol + 1 |
494 |
gtypeMaxCutoffCol(nGroupTypesCol) = groupMaxCutoffCol(j) |
495 |
groupToGtypeCol(j) = nGroupTypesCol |
496 |
else |
497 |
GtypeFound = .false. |
498 |
do g = 1, nGroupTypesCol |
499 |
if ( abs(groupMaxCutoffCol(j) - gtypeMaxCutoffCol(g)).lt.tol) then |
500 |
groupToGtypeCol(j) = g |
501 |
GtypeFound = .true. |
502 |
endif |
503 |
enddo |
504 |
if (.not.GtypeFound) then |
505 |
nGroupTypesCol = nGroupTypesCol + 1 |
506 |
gtypeMaxCutoffCol(nGroupTypesCol) = groupMaxCutoffCol(j) |
507 |
groupToGtypeCol(j) = nGroupTypesCol |
508 |
endif |
509 |
endif |
510 |
enddo |
511 |
|
512 |
#else |
513 |
! Set pointers to information we just found |
514 |
nGroupTypesCol = nGroupTypesRow |
515 |
groupToGtypeCol => groupToGtypeRow |
516 |
gtypeMaxCutoffCol => gtypeMaxCutoffRow |
517 |
groupMaxCutoffCol => groupMaxCutoffRow |
518 |
#endif |
519 |
|
520 |
!! allocate the gtypeCutoffMap here. |
521 |
allocate(gtypeCutoffMap(nGroupTypesRow,nGroupTypesCol)) |
522 |
!! then we do a double loop over all the group TYPES to find the cutoff |
523 |
!! map between groups of two types |
524 |
tradRcut = max(maxval(gtypeMaxCutoffRow),maxval(gtypeMaxCutoffCol)) |
525 |
|
526 |
do i = 1, nGroupTypesRow |
527 |
do j = 1, nGroupTypesCol |
528 |
|
529 |
select case(cutoffPolicy) |
530 |
case(TRADITIONAL_CUTOFF_POLICY) |
531 |
thisRcut = tradRcut |
532 |
case(MIX_CUTOFF_POLICY) |
533 |
thisRcut = 0.5_dp * (gtypeMaxCutoffRow(i) + gtypeMaxCutoffCol(j)) |
534 |
case(MAX_CUTOFF_POLICY) |
535 |
thisRcut = max(gtypeMaxCutoffRow(i), gtypeMaxCutoffCol(j)) |
536 |
case default |
537 |
call handleError("createGtypeCutoffMap", "Unknown Cutoff Policy") |
538 |
return |
539 |
end select |
540 |
gtypeCutoffMap(i,j)%rcut = thisRcut |
541 |
|
542 |
if (thisRcut.gt.largestRcut) largestRcut = thisRcut |
543 |
|
544 |
gtypeCutoffMap(i,j)%rcutsq = thisRcut*thisRcut |
545 |
|
546 |
if (.not.haveSkinThickness) then |
547 |
skinThickness = 1.0_dp |
548 |
endif |
549 |
|
550 |
gtypeCutoffMap(i,j)%rlistsq = (thisRcut + skinThickness)**2 |
551 |
|
552 |
! sanity check |
553 |
|
554 |
if (haveDefaultCutoffs) then |
555 |
if (abs(gtypeCutoffMap(i,j)%rcut - defaultRcut).gt.0.0001) then |
556 |
call handleError("createGtypeCutoffMap", "user-specified rCut does not match computed group Cutoff") |
557 |
endif |
558 |
endif |
559 |
enddo |
560 |
enddo |
561 |
|
562 |
if(allocated(gtypeMaxCutoffRow)) deallocate(gtypeMaxCutoffRow) |
563 |
if(allocated(groupMaxCutoffRow)) deallocate(groupMaxCutoffRow) |
564 |
if(allocated(atypeMaxCutoff)) deallocate(atypeMaxCutoff) |
565 |
#ifdef IS_MPI |
566 |
if(associated(groupMaxCutoffCol)) deallocate(groupMaxCutoffCol) |
567 |
if(associated(gtypeMaxCutoffCol)) deallocate(gtypeMaxCutoffCol) |
568 |
#endif |
569 |
groupMaxCutoffCol => null() |
570 |
gtypeMaxCutoffCol => null() |
571 |
|
572 |
haveGtypeCutoffMap = .true. |
573 |
end subroutine createGtypeCutoffMap |
574 |
|
575 |
subroutine setCutoffs(defRcut, defRsw, defSP, defSF) |
576 |
|
577 |
real(kind=dp),intent(in) :: defRcut, defRsw |
578 |
integer, intent(in) :: defSP, defSF |
579 |
character(len = statusMsgSize) :: errMsg |
580 |
integer :: localError |
581 |
|
582 |
defaultRcut = defRcut |
583 |
defaultRsw = defRsw |
584 |
|
585 |
if (defSP .ne. 0) then |
586 |
defaultDoShiftPot = .true. |
587 |
else |
588 |
defaultDoShiftPot = .false. |
589 |
endif |
590 |
if (defSF .ne. 0) then |
591 |
defaultDoShiftFrc = .true. |
592 |
else |
593 |
defaultDoShiftFrc = .false. |
594 |
endif |
595 |
|
596 |
if (abs(defaultRcut-defaultRsw) .lt. 0.0001) then |
597 |
if (defaultDoShiftFrc) then |
598 |
write(errMsg, *) & |
599 |
'cutoffRadius and switchingRadius are set to the', newline & |
600 |
// tab, 'same value. OpenMD will use shifted force', newline & |
601 |
// tab, 'potentials instead of switching functions.' |
602 |
|
603 |
call handleInfo("setCutoffs", errMsg) |
604 |
else |
605 |
write(errMsg, *) & |
606 |
'cutoffRadius and switchingRadius are set to the', newline & |
607 |
// tab, 'same value. OpenMD will use shifted', newline & |
608 |
// tab, 'potentials instead of switching functions.' |
609 |
|
610 |
call handleInfo("setCutoffs", errMsg) |
611 |
|
612 |
defaultDoShiftPot = .true. |
613 |
endif |
614 |
|
615 |
endif |
616 |
|
617 |
localError = 0 |
618 |
call setLJDefaultCutoff( defaultRcut, defaultDoShiftPot, & |
619 |
defaultDoShiftFrc ) |
620 |
call setElectrostaticCutoffRadius( defaultRcut, defaultRsw ) |
621 |
call setCutoffEAM( defaultRcut ) |
622 |
call setCutoffSC( defaultRcut ) |
623 |
call setMnMDefaultCutoff( defaultRcut, defaultDoShiftPot, & |
624 |
defaultDoShiftFrc ) |
625 |
call set_switch(defaultRsw, defaultRcut) |
626 |
call setHmatDangerousRcutValue(defaultRcut) |
627 |
|
628 |
haveDefaultCutoffs = .true. |
629 |
haveGtypeCutoffMap = .false. |
630 |
|
631 |
end subroutine setCutoffs |
632 |
|
633 |
subroutine cWasLame() |
634 |
|
635 |
VisitCutoffsAfterComputing = .true. |
636 |
return |
637 |
|
638 |
end subroutine cWasLame |
639 |
|
640 |
subroutine setCutoffPolicy(cutPolicy) |
641 |
|
642 |
integer, intent(in) :: cutPolicy |
643 |
|
644 |
cutoffPolicy = cutPolicy |
645 |
haveCutoffPolicy = .true. |
646 |
haveGtypeCutoffMap = .false. |
647 |
|
648 |
end subroutine setCutoffPolicy |
649 |
|
650 |
subroutine setBoxDipole() |
651 |
|
652 |
do_box_dipole = .true. |
653 |
|
654 |
end subroutine setBoxDipole |
655 |
|
656 |
subroutine getBoxDipole( box_dipole ) |
657 |
|
658 |
real(kind=dp), intent(inout), dimension(3) :: box_dipole |
659 |
|
660 |
box_dipole = boxDipole |
661 |
|
662 |
end subroutine getBoxDipole |
663 |
|
664 |
subroutine setElectrostaticMethod( thisESM ) |
665 |
|
666 |
integer, intent(in) :: thisESM |
667 |
|
668 |
electrostaticSummationMethod = thisESM |
669 |
haveElectrostaticSummationMethod = .true. |
670 |
|
671 |
end subroutine setElectrostaticMethod |
672 |
|
673 |
subroutine setSkinThickness( thisSkin ) |
674 |
|
675 |
real(kind=dp), intent(in) :: thisSkin |
676 |
|
677 |
skinThickness = thisSkin |
678 |
haveSkinThickness = .true. |
679 |
haveGtypeCutoffMap = .false. |
680 |
|
681 |
end subroutine setSkinThickness |
682 |
|
683 |
subroutine setSimVariables() |
684 |
SIM_uses_DirectionalAtoms = SimUsesDirectionalAtoms() |
685 |
SIM_uses_EAM = SimUsesEAM() |
686 |
SIM_requires_postpair_calc = SimRequiresPostpairCalc() |
687 |
SIM_requires_prepair_calc = SimRequiresPrepairCalc() |
688 |
SIM_uses_PBC = SimUsesPBC() |
689 |
SIM_uses_SC = SimUsesSC() |
690 |
SIM_uses_AtomicVirial = SimUsesAtomicVirial() |
691 |
|
692 |
haveSIMvariables = .true. |
693 |
|
694 |
return |
695 |
end subroutine setSimVariables |
696 |
|
697 |
subroutine doReadyCheck(error) |
698 |
integer, intent(out) :: error |
699 |
integer :: myStatus |
700 |
|
701 |
error = 0 |
702 |
|
703 |
if (.not. haveInteractionHash) then |
704 |
call createInteractionHash() |
705 |
endif |
706 |
|
707 |
if (.not. haveGtypeCutoffMap) then |
708 |
call createGtypeCutoffMap() |
709 |
endif |
710 |
|
711 |
if (VisitCutoffsAfterComputing) then |
712 |
call set_switch(largestRcut, largestRcut) |
713 |
call setHmatDangerousRcutValue(largestRcut) |
714 |
call setCutoffEAM(largestRcut) |
715 |
call setCutoffSC(largestRcut) |
716 |
VisitCutoffsAfterComputing = .false. |
717 |
endif |
718 |
|
719 |
if (.not. haveSIMvariables) then |
720 |
call setSimVariables() |
721 |
endif |
722 |
|
723 |
if (.not. haveNeighborList) then |
724 |
write(default_error, *) 'neighbor list has not been initialized in doForces!' |
725 |
error = -1 |
726 |
return |
727 |
end if |
728 |
|
729 |
if (.not. haveSaneForceField) then |
730 |
write(default_error, *) 'Force Field is not sane in doForces!' |
731 |
error = -1 |
732 |
return |
733 |
end if |
734 |
|
735 |
#ifdef IS_MPI |
736 |
if (.not. isMPISimSet()) then |
737 |
write(default_error,*) "ERROR: mpiSimulation has not been initialized!" |
738 |
error = -1 |
739 |
return |
740 |
endif |
741 |
#endif |
742 |
return |
743 |
end subroutine doReadyCheck |
744 |
|
745 |
|
746 |
subroutine init_FF(thisStat) |
747 |
|
748 |
integer, intent(out) :: thisStat |
749 |
integer :: my_status, nMatches |
750 |
integer, pointer :: MatchList(:) => null() |
751 |
|
752 |
!! assume things are copacetic, unless they aren't |
753 |
thisStat = 0 |
754 |
|
755 |
!! init_FF is called *after* all of the atom types have been |
756 |
!! defined in atype_module using the new_atype subroutine. |
757 |
!! |
758 |
!! this will scan through the known atypes and figure out what |
759 |
!! interactions are used by the force field. |
760 |
|
761 |
FF_uses_DirectionalAtoms = .false. |
762 |
FF_uses_Dipoles = .false. |
763 |
FF_uses_GayBerne = .false. |
764 |
FF_uses_EAM = .false. |
765 |
FF_uses_SC = .false. |
766 |
|
767 |
call getMatchingElementList(atypes, "is_Directional", .true., & |
768 |
nMatches, MatchList) |
769 |
if (nMatches .gt. 0) FF_uses_DirectionalAtoms = .true. |
770 |
|
771 |
call getMatchingElementList(atypes, "is_Dipole", .true., & |
772 |
nMatches, MatchList) |
773 |
if (nMatches .gt. 0) FF_uses_Dipoles = .true. |
774 |
|
775 |
call getMatchingElementList(atypes, "is_GayBerne", .true., & |
776 |
nMatches, MatchList) |
777 |
if (nMatches .gt. 0) FF_uses_GayBerne = .true. |
778 |
|
779 |
call getMatchingElementList(atypes, "is_EAM", .true., nMatches, MatchList) |
780 |
if (nMatches .gt. 0) FF_uses_EAM = .true. |
781 |
|
782 |
call getMatchingElementList(atypes, "is_SC", .true., nMatches, MatchList) |
783 |
if (nMatches .gt. 0) FF_uses_SC = .true. |
784 |
|
785 |
|
786 |
haveSaneForceField = .true. |
787 |
|
788 |
|
789 |
if (.not. haveNeighborList) then |
790 |
!! Create neighbor lists |
791 |
call expandNeighborList(nLocal, my_status) |
792 |
if (my_Status /= 0) then |
793 |
write(default_error,*) "SimSetup: ExpandNeighborList returned error." |
794 |
thisStat = -1 |
795 |
return |
796 |
endif |
797 |
haveNeighborList = .true. |
798 |
endif |
799 |
|
800 |
end subroutine init_FF |
801 |
|
802 |
|
803 |
!! Does force loop over i,j pairs. Calls do_pair to calculates forces. |
804 |
!-------------------------------------------------------------> |
805 |
subroutine do_force_loop(q, q_group, A, eFrame, f, t, tau, pot, particle_pot, & |
806 |
error) |
807 |
!! Position array provided by C, dimensioned by getNlocal |
808 |
real ( kind = dp ), dimension(3, nLocal) :: q |
809 |
!! molecular center-of-mass position array |
810 |
real ( kind = dp ), dimension(3, nGroups) :: q_group |
811 |
!! Rotation Matrix for each long range particle in simulation. |
812 |
real( kind = dp), dimension(9, nLocal) :: A |
813 |
!! Unit vectors for dipoles (lab frame) |
814 |
real( kind = dp ), dimension(9,nLocal) :: eFrame |
815 |
!! Force array provided by C, dimensioned by getNlocal |
816 |
real ( kind = dp ), dimension(3,nLocal) :: f |
817 |
!! Torsion array provided by C, dimensioned by getNlocal |
818 |
real( kind = dp ), dimension(3,nLocal) :: t |
819 |
|
820 |
!! Stress Tensor |
821 |
real( kind = dp), dimension(9) :: tau |
822 |
real ( kind = dp ),dimension(LR_POT_TYPES) :: pot |
823 |
real( kind = dp ), dimension(nLocal) :: particle_pot |
824 |
|
825 |
logical :: in_switching_region |
826 |
#ifdef IS_MPI |
827 |
real( kind = DP ), dimension(LR_POT_TYPES) :: pot_local |
828 |
integer :: nAtomsInRow |
829 |
integer :: nAtomsInCol |
830 |
integer :: nprocs |
831 |
integer :: nGroupsInRow |
832 |
integer :: nGroupsInCol |
833 |
#endif |
834 |
integer :: natoms |
835 |
logical :: update_nlist |
836 |
integer :: i, j, jstart, jend, jnab |
837 |
integer :: istart, iend |
838 |
integer :: ia, jb, atom1, atom2 |
839 |
integer :: nlist |
840 |
real( kind = DP ) :: ratmsq, rgrpsq, rgrp, rag, vpair, vij |
841 |
real( kind = DP ) :: sw, dswdr, swderiv, mf |
842 |
real( kind = DP ) :: rVal |
843 |
real(kind=dp),dimension(3) :: d_atm, d_grp, fpair, fij, fg, dag |
844 |
real(kind=dp) :: rfpot, mu_i |
845 |
real(kind=dp):: rCut |
846 |
integer :: me_i, me_j, n_in_i, n_in_j, iG, j1 |
847 |
logical :: is_dp_i |
848 |
integer :: neighborListSize |
849 |
integer :: listerror, error |
850 |
integer :: localError |
851 |
integer :: propPack_i, propPack_j |
852 |
integer :: loopStart, loopEnd, loop |
853 |
integer :: iHash, jHash |
854 |
integer :: i1, topoDist |
855 |
|
856 |
!! the variables for the box dipole moment |
857 |
#ifdef IS_MPI |
858 |
integer :: pChgCount_local |
859 |
integer :: nChgCount_local |
860 |
real(kind=dp) :: pChg_local |
861 |
real(kind=dp) :: nChg_local |
862 |
real(kind=dp), dimension(3) :: pChgPos_local |
863 |
real(kind=dp), dimension(3) :: nChgPos_local |
864 |
real(kind=dp), dimension(3) :: dipVec_local |
865 |
#endif |
866 |
integer :: pChgCount |
867 |
integer :: nChgCount |
868 |
real(kind=dp) :: pChg |
869 |
real(kind=dp) :: nChg |
870 |
real(kind=dp) :: chg_value |
871 |
real(kind=dp), dimension(3) :: pChgPos |
872 |
real(kind=dp), dimension(3) :: nChgPos |
873 |
real(kind=dp), dimension(3) :: dipVec |
874 |
real(kind=dp), dimension(3) :: chgVec |
875 |
real(kind=dp) :: skch |
876 |
|
877 |
!! initialize box dipole variables |
878 |
if (do_box_dipole) then |
879 |
#ifdef IS_MPI |
880 |
pChg_local = 0.0_dp |
881 |
nChg_local = 0.0_dp |
882 |
pChgCount_local = 0 |
883 |
nChgCount_local = 0 |
884 |
do i=1, 3 |
885 |
pChgPos_local = 0.0_dp |
886 |
nChgPos_local = 0.0_dp |
887 |
dipVec_local = 0.0_dp |
888 |
enddo |
889 |
#endif |
890 |
pChg = 0.0_dp |
891 |
nChg = 0.0_dp |
892 |
pChgCount = 0 |
893 |
nChgCount = 0 |
894 |
chg_value = 0.0_dp |
895 |
|
896 |
do i=1, 3 |
897 |
pChgPos(i) = 0.0_dp |
898 |
nChgPos(i) = 0.0_dp |
899 |
dipVec(i) = 0.0_dp |
900 |
chgVec(i) = 0.0_dp |
901 |
boxDipole(i) = 0.0_dp |
902 |
enddo |
903 |
endif |
904 |
|
905 |
!! initialize local variables |
906 |
|
907 |
#ifdef IS_MPI |
908 |
pot_local = 0.0_dp |
909 |
nAtomsInRow = getNatomsInRow(plan_atom_row) |
910 |
nAtomsInCol = getNatomsInCol(plan_atom_col) |
911 |
nGroupsInRow = getNgroupsInRow(plan_group_row) |
912 |
nGroupsInCol = getNgroupsInCol(plan_group_col) |
913 |
#else |
914 |
natoms = nlocal |
915 |
#endif |
916 |
|
917 |
call doReadyCheck(localError) |
918 |
if ( localError .ne. 0 ) then |
919 |
call handleError("do_force_loop", "Not Initialized") |
920 |
error = -1 |
921 |
return |
922 |
end if |
923 |
call zero_work_arrays() |
924 |
|
925 |
! Gather all information needed by all force loops: |
926 |
|
927 |
#ifdef IS_MPI |
928 |
|
929 |
call gather(q, q_Row, plan_atom_row_3d) |
930 |
call gather(q, q_Col, plan_atom_col_3d) |
931 |
|
932 |
call gather(q_group, q_group_Row, plan_group_row_3d) |
933 |
call gather(q_group, q_group_Col, plan_group_col_3d) |
934 |
|
935 |
if (FF_UsesDirectionalAtoms() .and. SIM_uses_DirectionalAtoms) then |
936 |
call gather(eFrame, eFrame_Row, plan_atom_row_rotation) |
937 |
call gather(eFrame, eFrame_Col, plan_atom_col_rotation) |
938 |
|
939 |
call gather(A, A_Row, plan_atom_row_rotation) |
940 |
call gather(A, A_Col, plan_atom_col_rotation) |
941 |
endif |
942 |
|
943 |
#endif |
944 |
|
945 |
!! Begin force loop timing: |
946 |
#ifdef PROFILE |
947 |
call cpu_time(forceTimeInitial) |
948 |
nloops = nloops + 1 |
949 |
#endif |
950 |
|
951 |
loopEnd = PAIR_LOOP |
952 |
if (FF_RequiresPrepairCalc() .and. SIM_requires_prepair_calc) then |
953 |
loopStart = PREPAIR_LOOP |
954 |
else |
955 |
loopStart = PAIR_LOOP |
956 |
endif |
957 |
|
958 |
do loop = loopStart, loopEnd |
959 |
|
960 |
! See if we need to update neighbor lists |
961 |
! (but only on the first time through): |
962 |
if (loop .eq. loopStart) then |
963 |
#ifdef IS_MPI |
964 |
call checkNeighborList(nGroupsInRow, q_group_row, skinThickness, & |
965 |
update_nlist) |
966 |
#else |
967 |
call checkNeighborList(nGroups, q_group, skinThickness, & |
968 |
update_nlist) |
969 |
#endif |
970 |
endif |
971 |
|
972 |
if (update_nlist) then |
973 |
!! save current configuration and construct neighbor list |
974 |
#ifdef IS_MPI |
975 |
call saveNeighborList(nGroupsInRow, q_group_row) |
976 |
#else |
977 |
call saveNeighborList(nGroups, q_group) |
978 |
#endif |
979 |
neighborListSize = size(list) |
980 |
nlist = 0 |
981 |
endif |
982 |
|
983 |
istart = 1 |
984 |
#ifdef IS_MPI |
985 |
iend = nGroupsInRow |
986 |
#else |
987 |
iend = nGroups - 1 |
988 |
#endif |
989 |
outer: do i = istart, iend |
990 |
|
991 |
if (update_nlist) point(i) = nlist + 1 |
992 |
|
993 |
n_in_i = groupStartRow(i+1) - groupStartRow(i) |
994 |
|
995 |
if (update_nlist) then |
996 |
#ifdef IS_MPI |
997 |
jstart = 1 |
998 |
jend = nGroupsInCol |
999 |
#else |
1000 |
jstart = i+1 |
1001 |
jend = nGroups |
1002 |
#endif |
1003 |
else |
1004 |
jstart = point(i) |
1005 |
jend = point(i+1) - 1 |
1006 |
! make sure group i has neighbors |
1007 |
if (jstart .gt. jend) cycle outer |
1008 |
endif |
1009 |
|
1010 |
do jnab = jstart, jend |
1011 |
if (update_nlist) then |
1012 |
j = jnab |
1013 |
else |
1014 |
j = list(jnab) |
1015 |
endif |
1016 |
|
1017 |
#ifdef IS_MPI |
1018 |
me_j = atid_col(j) |
1019 |
call get_interatomic_vector(q_group_Row(:,i), & |
1020 |
q_group_Col(:,j), d_grp, rgrpsq) |
1021 |
#else |
1022 |
me_j = atid(j) |
1023 |
call get_interatomic_vector(q_group(:,i), & |
1024 |
q_group(:,j), d_grp, rgrpsq) |
1025 |
#endif |
1026 |
|
1027 |
if (rgrpsq < gtypeCutoffMap(groupToGtypeRow(i),groupToGtypeCol(j))%rListsq) then |
1028 |
if (update_nlist) then |
1029 |
nlist = nlist + 1 |
1030 |
|
1031 |
if (nlist > neighborListSize) then |
1032 |
#ifdef IS_MPI |
1033 |
call expandNeighborList(nGroupsInRow, listerror) |
1034 |
#else |
1035 |
call expandNeighborList(nGroups, listerror) |
1036 |
#endif |
1037 |
if (listerror /= 0) then |
1038 |
error = -1 |
1039 |
write(DEFAULT_ERROR,*) "ERROR: nlist > list size and max allocations exceeded." |
1040 |
return |
1041 |
end if |
1042 |
neighborListSize = size(list) |
1043 |
endif |
1044 |
|
1045 |
list(nlist) = j |
1046 |
endif |
1047 |
|
1048 |
if (rgrpsq < gtypeCutoffMap(groupToGtypeRow(i),groupToGtypeCol(j))%rCutsq) then |
1049 |
|
1050 |
rCut = gtypeCutoffMap(groupToGtypeRow(i),groupToGtypeCol(j))%rCut |
1051 |
if (loop .eq. PAIR_LOOP) then |
1052 |
vij = 0.0_dp |
1053 |
fij(1) = 0.0_dp |
1054 |
fij(2) = 0.0_dp |
1055 |
fij(3) = 0.0_dp |
1056 |
endif |
1057 |
|
1058 |
call get_switch(rgrpsq, sw, dswdr,rgrp, in_switching_region) |
1059 |
|
1060 |
n_in_j = groupStartCol(j+1) - groupStartCol(j) |
1061 |
|
1062 |
do ia = groupStartRow(i), groupStartRow(i+1)-1 |
1063 |
|
1064 |
atom1 = groupListRow(ia) |
1065 |
|
1066 |
inner: do jb = groupStartCol(j), groupStartCol(j+1)-1 |
1067 |
|
1068 |
atom2 = groupListCol(jb) |
1069 |
|
1070 |
if (skipThisPair(atom1, atom2)) cycle inner |
1071 |
|
1072 |
if ((n_in_i .eq. 1).and.(n_in_j .eq. 1)) then |
1073 |
d_atm(1) = d_grp(1) |
1074 |
d_atm(2) = d_grp(2) |
1075 |
d_atm(3) = d_grp(3) |
1076 |
ratmsq = rgrpsq |
1077 |
else |
1078 |
#ifdef IS_MPI |
1079 |
call get_interatomic_vector(q_Row(:,atom1), & |
1080 |
q_Col(:,atom2), d_atm, ratmsq) |
1081 |
#else |
1082 |
call get_interatomic_vector(q(:,atom1), & |
1083 |
q(:,atom2), d_atm, ratmsq) |
1084 |
#endif |
1085 |
endif |
1086 |
|
1087 |
topoDist = getTopoDistance(atom1, atom2) |
1088 |
|
1089 |
if (loop .eq. PREPAIR_LOOP) then |
1090 |
#ifdef IS_MPI |
1091 |
call do_prepair(atom1, atom2, ratmsq, d_atm, sw, & |
1092 |
rgrpsq, d_grp, rCut, & |
1093 |
eFrame, A, f, t, pot_local) |
1094 |
#else |
1095 |
call do_prepair(atom1, atom2, ratmsq, d_atm, sw, & |
1096 |
rgrpsq, d_grp, rCut, & |
1097 |
eFrame, A, f, t, pot) |
1098 |
#endif |
1099 |
else |
1100 |
#ifdef IS_MPI |
1101 |
call do_pair(atom1, atom2, ratmsq, d_atm, sw, & |
1102 |
eFrame, A, f, t, pot_local, particle_pot, vpair, & |
1103 |
fpair, d_grp, rgrp, rCut, topoDist) |
1104 |
! particle_pot will be accumulated from row & column |
1105 |
! arrays later |
1106 |
#else |
1107 |
call do_pair(atom1, atom2, ratmsq, d_atm, sw, & |
1108 |
eFrame, A, f, t, pot, particle_pot, vpair, & |
1109 |
fpair, d_grp, rgrp, rCut, topoDist) |
1110 |
#endif |
1111 |
vij = vij + vpair |
1112 |
fij(1) = fij(1) + fpair(1) |
1113 |
fij(2) = fij(2) + fpair(2) |
1114 |
fij(3) = fij(3) + fpair(3) |
1115 |
call add_stress_tensor(d_atm, fpair, tau) |
1116 |
endif |
1117 |
enddo inner |
1118 |
enddo |
1119 |
|
1120 |
if (loop .eq. PAIR_LOOP) then |
1121 |
if (in_switching_region) then |
1122 |
swderiv = vij*dswdr/rgrp |
1123 |
fg = swderiv*d_grp |
1124 |
|
1125 |
fij(1) = fij(1) + fg(1) |
1126 |
fij(2) = fij(2) + fg(2) |
1127 |
fij(3) = fij(3) + fg(3) |
1128 |
|
1129 |
if ((n_in_i .eq. 1).and.(n_in_j .eq. 1)) then |
1130 |
call add_stress_tensor(d_atm, fg, tau) |
1131 |
endif |
1132 |
|
1133 |
do ia=groupStartRow(i), groupStartRow(i+1)-1 |
1134 |
atom1=groupListRow(ia) |
1135 |
mf = mfactRow(atom1) |
1136 |
! fg is the force on atom ia due to cutoff group's |
1137 |
! presence in switching region |
1138 |
fg = swderiv*d_grp*mf |
1139 |
#ifdef IS_MPI |
1140 |
f_Row(1,atom1) = f_Row(1,atom1) + fg(1) |
1141 |
f_Row(2,atom1) = f_Row(2,atom1) + fg(2) |
1142 |
f_Row(3,atom1) = f_Row(3,atom1) + fg(3) |
1143 |
#else |
1144 |
f(1,atom1) = f(1,atom1) + fg(1) |
1145 |
f(2,atom1) = f(2,atom1) + fg(2) |
1146 |
f(3,atom1) = f(3,atom1) + fg(3) |
1147 |
#endif |
1148 |
if (n_in_i .gt. 1) then |
1149 |
if (SIM_uses_AtomicVirial) then |
1150 |
! find the distance between the atom |
1151 |
! and the center of the cutoff group: |
1152 |
#ifdef IS_MPI |
1153 |
call get_interatomic_vector(q_Row(:,atom1), & |
1154 |
q_group_Row(:,i), dag, rag) |
1155 |
#else |
1156 |
call get_interatomic_vector(q(:,atom1), & |
1157 |
q_group(:,i), dag, rag) |
1158 |
#endif |
1159 |
call add_stress_tensor(dag,fg,tau) |
1160 |
endif |
1161 |
endif |
1162 |
enddo |
1163 |
|
1164 |
do jb=groupStartCol(j), groupStartCol(j+1)-1 |
1165 |
atom2=groupListCol(jb) |
1166 |
mf = mfactCol(atom2) |
1167 |
! fg is the force on atom jb due to cutoff group's |
1168 |
! presence in switching region |
1169 |
fg = -swderiv*d_grp*mf |
1170 |
#ifdef IS_MPI |
1171 |
f_Col(1,atom2) = f_Col(1,atom2) + fg(1) |
1172 |
f_Col(2,atom2) = f_Col(2,atom2) + fg(2) |
1173 |
f_Col(3,atom2) = f_Col(3,atom2) + fg(3) |
1174 |
#else |
1175 |
f(1,atom2) = f(1,atom2) + fg(1) |
1176 |
f(2,atom2) = f(2,atom2) + fg(2) |
1177 |
f(3,atom2) = f(3,atom2) + fg(3) |
1178 |
#endif |
1179 |
if (n_in_j .gt. 1) then |
1180 |
if (SIM_uses_AtomicVirial) then |
1181 |
! find the distance between the atom |
1182 |
! and the center of the cutoff group: |
1183 |
#ifdef IS_MPI |
1184 |
call get_interatomic_vector(q_Col(:,atom2), & |
1185 |
q_group_Col(:,j), dag, rag) |
1186 |
#else |
1187 |
call get_interatomic_vector(q(:,atom2), & |
1188 |
q_group(:,j), dag, rag) |
1189 |
#endif |
1190 |
call add_stress_tensor(dag,fg,tau) |
1191 |
endif |
1192 |
endif |
1193 |
enddo |
1194 |
endif |
1195 |
!if (.not.SIM_uses_AtomicVirial) then |
1196 |
! call add_stress_tensor(d_grp, fij, tau) |
1197 |
!endif |
1198 |
endif |
1199 |
endif |
1200 |
endif |
1201 |
enddo |
1202 |
|
1203 |
enddo outer |
1204 |
|
1205 |
if (update_nlist) then |
1206 |
#ifdef IS_MPI |
1207 |
point(nGroupsInRow + 1) = nlist + 1 |
1208 |
#else |
1209 |
point(nGroups) = nlist + 1 |
1210 |
#endif |
1211 |
if (loop .eq. PREPAIR_LOOP) then |
1212 |
! we just did the neighbor list update on the first |
1213 |
! pass, so we don't need to do it |
1214 |
! again on the second pass |
1215 |
update_nlist = .false. |
1216 |
endif |
1217 |
endif |
1218 |
|
1219 |
if (loop .eq. PREPAIR_LOOP) then |
1220 |
#ifdef IS_MPI |
1221 |
call do_preforce(nlocal, pot_local, particle_pot) |
1222 |
#else |
1223 |
call do_preforce(nlocal, pot, particle_pot) |
1224 |
#endif |
1225 |
endif |
1226 |
|
1227 |
enddo |
1228 |
|
1229 |
!! Do timing |
1230 |
#ifdef PROFILE |
1231 |
call cpu_time(forceTimeFinal) |
1232 |
forceTime = forceTime + forceTimeFinal - forceTimeInitial |
1233 |
#endif |
1234 |
|
1235 |
#ifdef IS_MPI |
1236 |
!!distribute forces |
1237 |
|
1238 |
f_temp = 0.0_dp |
1239 |
call scatter(f_Row,f_temp,plan_atom_row_3d) |
1240 |
do i = 1,nlocal |
1241 |
f(1:3,i) = f(1:3,i) + f_temp(1:3,i) |
1242 |
end do |
1243 |
|
1244 |
f_temp = 0.0_dp |
1245 |
call scatter(f_Col,f_temp,plan_atom_col_3d) |
1246 |
do i = 1,nlocal |
1247 |
f(1:3,i) = f(1:3,i) + f_temp(1:3,i) |
1248 |
end do |
1249 |
|
1250 |
if (FF_UsesDirectionalAtoms() .and. SIM_uses_DirectionalAtoms) then |
1251 |
t_temp = 0.0_dp |
1252 |
call scatter(t_Row,t_temp,plan_atom_row_3d) |
1253 |
do i = 1,nlocal |
1254 |
t(1:3,i) = t(1:3,i) + t_temp(1:3,i) |
1255 |
end do |
1256 |
t_temp = 0.0_dp |
1257 |
call scatter(t_Col,t_temp,plan_atom_col_3d) |
1258 |
|
1259 |
do i = 1,nlocal |
1260 |
t(1:3,i) = t(1:3,i) + t_temp(1:3,i) |
1261 |
end do |
1262 |
endif |
1263 |
|
1264 |
! scatter/gather pot_row into the members of my column |
1265 |
do i = 1,LR_POT_TYPES |
1266 |
call scatter(pot_Row(i,:), pot_Temp(i,:), plan_atom_row) |
1267 |
end do |
1268 |
! scatter/gather pot_local into all other procs |
1269 |
! add resultant to get total pot |
1270 |
do i = 1, nlocal |
1271 |
pot_local(1:LR_POT_TYPES) = pot_local(1:LR_POT_TYPES) & |
1272 |
+ pot_Temp(1:LR_POT_TYPES,i) |
1273 |
enddo |
1274 |
|
1275 |
do i = 1,LR_POT_TYPES |
1276 |
particle_pot(1:nlocal) = particle_pot(1:nlocal) + pot_Temp(i,1:nlocal) |
1277 |
enddo |
1278 |
|
1279 |
pot_Temp = 0.0_DP |
1280 |
|
1281 |
do i = 1,LR_POT_TYPES |
1282 |
call scatter(pot_Col(i,:), pot_Temp(i,:), plan_atom_col) |
1283 |
end do |
1284 |
|
1285 |
do i = 1, nlocal |
1286 |
pot_local(1:LR_POT_TYPES) = pot_local(1:LR_POT_TYPES)& |
1287 |
+ pot_Temp(1:LR_POT_TYPES,i) |
1288 |
enddo |
1289 |
|
1290 |
do i = 1,LR_POT_TYPES |
1291 |
particle_pot(1:nlocal) = particle_pot(1:nlocal) + pot_Temp(i,1:nlocal) |
1292 |
enddo |
1293 |
|
1294 |
ppot_Temp = 0.0_DP |
1295 |
|
1296 |
call scatter(ppot_Row(:), ppot_Temp(:), plan_atom_row) |
1297 |
do i = 1, nlocal |
1298 |
particle_pot(i) = particle_pot(i) + ppot_Temp(i) |
1299 |
enddo |
1300 |
|
1301 |
ppot_Temp = 0.0_DP |
1302 |
|
1303 |
call scatter(ppot_Col(:), ppot_Temp(:), plan_atom_col) |
1304 |
do i = 1, nlocal |
1305 |
particle_pot(i) = particle_pot(i) + ppot_Temp(i) |
1306 |
enddo |
1307 |
|
1308 |
#endif |
1309 |
|
1310 |
if (SIM_requires_postpair_calc) then |
1311 |
do i = 1, nlocal |
1312 |
|
1313 |
! we loop only over the local atoms, so we don't need row and column |
1314 |
! lookups for the types |
1315 |
|
1316 |
me_i = atid(i) |
1317 |
|
1318 |
! is the atom electrostatic? See if it would have an |
1319 |
! electrostatic interaction with itself |
1320 |
iHash = InteractionHash(me_i,me_i) |
1321 |
|
1322 |
if ( iand(iHash, ELECTROSTATIC_PAIR).ne.0 ) then |
1323 |
|
1324 |
! loop over the excludes to accumulate charge in the |
1325 |
! cutoff sphere that we've left out of the normal pair loop |
1326 |
skch = 0.0_dp |
1327 |
|
1328 |
do i1 = 1, nSkipsForLocalAtom(i) |
1329 |
j = skipsForLocalAtom(i, i1) |
1330 |
me_j = atid(j) |
1331 |
jHash = InteractionHash(me_i,me_j) |
1332 |
if ( iand(jHash, ELECTROSTATIC_PAIR).ne.0 ) then |
1333 |
skch = skch + getCharge(me_j) |
1334 |
endif |
1335 |
enddo |
1336 |
|
1337 |
#ifdef IS_MPI |
1338 |
call self_self(i, eFrame, skch, pot_local(ELECTROSTATIC_POT), t) |
1339 |
#else |
1340 |
call self_self(i, eFrame, skch, pot(ELECTROSTATIC_POT), t) |
1341 |
#endif |
1342 |
endif |
1343 |
|
1344 |
|
1345 |
if (electrostaticSummationMethod.eq.REACTION_FIELD) then |
1346 |
|
1347 |
! loop over the excludes to accumulate RF stuff we've |
1348 |
! left out of the normal pair loop |
1349 |
|
1350 |
do i1 = 1, nSkipsForLocalAtom(i) |
1351 |
j = skipsForLocalAtom(i, i1) |
1352 |
|
1353 |
! prevent overcounting of the skips |
1354 |
if (i.lt.j) then |
1355 |
call get_interatomic_vector(q(:,i), q(:,j), d_atm, ratmsq) |
1356 |
rVal = sqrt(ratmsq) |
1357 |
call get_switch(ratmsq, sw, dswdr, rVal,in_switching_region) |
1358 |
#ifdef IS_MPI |
1359 |
call rf_self_excludes(i, j, sw, 1.0_dp, eFrame, d_atm, rVal, & |
1360 |
vpair, pot_local(ELECTROSTATIC_POT), f, t) |
1361 |
#else |
1362 |
call rf_self_excludes(i, j, sw, 1.0_dp, eFrame, d_atm, rVal, & |
1363 |
vpair, pot(ELECTROSTATIC_POT), f, t) |
1364 |
#endif |
1365 |
endif |
1366 |
enddo |
1367 |
endif |
1368 |
|
1369 |
if (do_box_dipole) then |
1370 |
#ifdef IS_MPI |
1371 |
call accumulate_box_dipole(i, eFrame, q(:,i), pChg_local, & |
1372 |
nChg_local, pChgPos_local, nChgPos_local, dipVec_local, & |
1373 |
pChgCount_local, nChgCount_local) |
1374 |
#else |
1375 |
call accumulate_box_dipole(i, eFrame, q(:,i), pChg, nChg, & |
1376 |
pChgPos, nChgPos, dipVec, pChgCount, nChgCount) |
1377 |
#endif |
1378 |
endif |
1379 |
enddo |
1380 |
endif |
1381 |
|
1382 |
#ifdef IS_MPI |
1383 |
#ifdef SINGLE_PRECISION |
1384 |
call mpi_allreduce(pot_local, pot, LR_POT_TYPES,mpi_real,mpi_sum, & |
1385 |
mpi_comm_world,mpi_err) |
1386 |
#else |
1387 |
call mpi_allreduce(pot_local, pot, LR_POT_TYPES,mpi_double_precision, & |
1388 |
mpi_sum, mpi_comm_world,mpi_err) |
1389 |
#endif |
1390 |
|
1391 |
if (do_box_dipole) then |
1392 |
|
1393 |
#ifdef SINGLE_PRECISION |
1394 |
call mpi_allreduce(pChg_local, pChg, 1, mpi_real, mpi_sum, & |
1395 |
mpi_comm_world, mpi_err) |
1396 |
call mpi_allreduce(nChg_local, nChg, 1, mpi_real, mpi_sum, & |
1397 |
mpi_comm_world, mpi_err) |
1398 |
call mpi_allreduce(pChgCount_local, pChgCount, 1, mpi_integer, mpi_sum,& |
1399 |
mpi_comm_world, mpi_err) |
1400 |
call mpi_allreduce(nChgCount_local, nChgCount, 1, mpi_integer, mpi_sum,& |
1401 |
mpi_comm_world, mpi_err) |
1402 |
call mpi_allreduce(pChgPos_local, pChgPos, 3, mpi_real, mpi_sum, & |
1403 |
mpi_comm_world, mpi_err) |
1404 |
call mpi_allreduce(nChgPos_local, nChgPos, 3, mpi_real, mpi_sum, & |
1405 |
mpi_comm_world, mpi_err) |
1406 |
call mpi_allreduce(dipVec_local, dipVec, 3, mpi_real, mpi_sum, & |
1407 |
mpi_comm_world, mpi_err) |
1408 |
#else |
1409 |
call mpi_allreduce(pChg_local, pChg, 1, mpi_double_precision, mpi_sum, & |
1410 |
mpi_comm_world, mpi_err) |
1411 |
call mpi_allreduce(nChg_local, nChg, 1, mpi_double_precision, mpi_sum, & |
1412 |
mpi_comm_world, mpi_err) |
1413 |
call mpi_allreduce(pChgCount_local, pChgCount, 1, mpi_integer,& |
1414 |
mpi_sum, mpi_comm_world, mpi_err) |
1415 |
call mpi_allreduce(nChgCount_local, nChgCount, 1, mpi_integer,& |
1416 |
mpi_sum, mpi_comm_world, mpi_err) |
1417 |
call mpi_allreduce(pChgPos_local, pChgPos, 3, mpi_double_precision, & |
1418 |
mpi_sum, mpi_comm_world, mpi_err) |
1419 |
call mpi_allreduce(nChgPos_local, nChgPos, 3, mpi_double_precision, & |
1420 |
mpi_sum, mpi_comm_world, mpi_err) |
1421 |
call mpi_allreduce(dipVec_local, dipVec, 3, mpi_double_precision, & |
1422 |
mpi_sum, mpi_comm_world, mpi_err) |
1423 |
#endif |
1424 |
|
1425 |
endif |
1426 |
|
1427 |
#endif |
1428 |
|
1429 |
if (do_box_dipole) then |
1430 |
! first load the accumulated dipole moment (if dipoles were present) |
1431 |
boxDipole(1) = dipVec(1) |
1432 |
boxDipole(2) = dipVec(2) |
1433 |
boxDipole(3) = dipVec(3) |
1434 |
|
1435 |
! now include the dipole moment due to charges |
1436 |
! use the lesser of the positive and negative charge totals |
1437 |
if (nChg .le. pChg) then |
1438 |
chg_value = nChg |
1439 |
else |
1440 |
chg_value = pChg |
1441 |
endif |
1442 |
|
1443 |
! find the average positions |
1444 |
if (pChgCount .gt. 0 .and. nChgCount .gt. 0) then |
1445 |
pChgPos = pChgPos / pChgCount |
1446 |
nChgPos = nChgPos / nChgCount |
1447 |
endif |
1448 |
|
1449 |
! dipole is from the negative to the positive (physics notation) |
1450 |
chgVec(1) = pChgPos(1) - nChgPos(1) |
1451 |
chgVec(2) = pChgPos(2) - nChgPos(2) |
1452 |
chgVec(3) = pChgPos(3) - nChgPos(3) |
1453 |
|
1454 |
boxDipole(1) = boxDipole(1) + chgVec(1) * chg_value |
1455 |
boxDipole(2) = boxDipole(2) + chgVec(2) * chg_value |
1456 |
boxDipole(3) = boxDipole(3) + chgVec(3) * chg_value |
1457 |
|
1458 |
endif |
1459 |
|
1460 |
end subroutine do_force_loop |
1461 |
|
1462 |
subroutine do_pair(i, j, rijsq, d, sw, & |
1463 |
eFrame, A, f, t, pot, particle_pot, vpair, & |
1464 |
fpair, d_grp, r_grp, rCut, topoDist) |
1465 |
|
1466 |
real( kind = dp ) :: vpair, sw |
1467 |
real( kind = dp ), dimension(LR_POT_TYPES) :: pot |
1468 |
real( kind = dp ), dimension(nLocal) :: particle_pot |
1469 |
real( kind = dp ), dimension(3) :: fpair |
1470 |
real( kind = dp ), dimension(nLocal) :: mfact |
1471 |
real( kind = dp ), dimension(9,nLocal) :: eFrame |
1472 |
real( kind = dp ), dimension(9,nLocal) :: A |
1473 |
real( kind = dp ), dimension(3,nLocal) :: f |
1474 |
real( kind = dp ), dimension(3,nLocal) :: t |
1475 |
|
1476 |
integer, intent(in) :: i, j |
1477 |
real ( kind = dp ), intent(inout) :: rijsq |
1478 |
real ( kind = dp ), intent(inout) :: r_grp |
1479 |
real ( kind = dp ), intent(inout) :: d(3) |
1480 |
real ( kind = dp ), intent(inout) :: d_grp(3) |
1481 |
real ( kind = dp ), intent(inout) :: rCut |
1482 |
integer, intent(inout) :: topoDist |
1483 |
real ( kind = dp ) :: r, pair_pot, vdwMult, electroMult |
1484 |
real ( kind = dp ) :: a_k, b_k, c_k, d_k, dx |
1485 |
|
1486 |
real( kind = dp), dimension(3) :: f1, t1, t2 |
1487 |
real( kind = dp), dimension(9) :: A1, A2, eF1, eF2 |
1488 |
real( kind = dp) :: dfrhodrho_i, dfrhodrho_j |
1489 |
real( kind = dp) :: rho_i, rho_j |
1490 |
real( kind = dp) :: fshift_i, fshift_j |
1491 |
real( kind = dp) :: p_vdw, p_elect, p_hb, p_met |
1492 |
integer :: atid_i, atid_j, id1, id2, idx |
1493 |
integer :: k |
1494 |
integer :: c_ident_i, c_ident_j |
1495 |
|
1496 |
integer :: iHash |
1497 |
|
1498 |
r = sqrt(rijsq) |
1499 |
|
1500 |
vpair = 0.0_dp |
1501 |
fpair(1:3) = 0.0_dp |
1502 |
|
1503 |
p_vdw = 0.0 |
1504 |
p_elect = 0.0 |
1505 |
p_hb = 0.0 |
1506 |
p_met = 0.0 |
1507 |
|
1508 |
f1(1:3) = 0.0 |
1509 |
t1(1:3) = 0.0 |
1510 |
t2(1:3) = 0.0 |
1511 |
|
1512 |
#ifdef IS_MPI |
1513 |
atid_i = atid_row(i) |
1514 |
atid_j = atid_col(j) |
1515 |
c_ident_i = c_idents_row(i) |
1516 |
c_ident_j = c_idents_col(j) |
1517 |
|
1518 |
do idx = 1, 9 |
1519 |
A1(idx) = A_Row(idx, i) |
1520 |
A2(idx) = A_Col(idx, j) |
1521 |
eF1(idx) = eFrame_Row(idx, i) |
1522 |
eF2(idx) = eFrame_Col(idx, j) |
1523 |
enddo |
1524 |
|
1525 |
#else |
1526 |
atid_i = atid(i) |
1527 |
atid_j = atid(j) |
1528 |
c_ident_i = c_idents_local(i) |
1529 |
c_ident_j = c_idents_local(j) |
1530 |
|
1531 |
do idx = 1, 9 |
1532 |
A1(idx) = A(idx, i) |
1533 |
A2(idx) = A(idx, j) |
1534 |
eF1(idx) = eFrame(idx, i) |
1535 |
eF2(idx) = eFrame(idx, j) |
1536 |
enddo |
1537 |
|
1538 |
#endif |
1539 |
|
1540 |
|
1541 |
iHash = InteractionHash(atid_i, atid_j) |
1542 |
|
1543 |
!! For the metallic potentials, we need to pass dF[rho]/drho since |
1544 |
!! the pair calculation routines no longer are aware of parallel. |
1545 |
|
1546 |
if ( (iand(iHash, EAM_PAIR).ne.0) .or. (iand(iHash, SC_PAIR).ne.0) ) then |
1547 |
#ifdef IS_MPI |
1548 |
dfrhodrho_i = dfrhodrho_row(i) |
1549 |
dfrhodrho_j = dfrhodrho_col(j) |
1550 |
rho_i = rho_row(i) |
1551 |
rho_j = rho_col(j) |
1552 |
#else |
1553 |
dfrhodrho_i = dfrhodrho(i) |
1554 |
dfrhodrho_j = dfrhodrho(j) |
1555 |
rho_i = rho(i) |
1556 |
rho_j = rho(j) |
1557 |
#endif |
1558 |
end if |
1559 |
|
1560 |
vdwMult = vdwScale(topoDist) |
1561 |
electroMult = electrostaticScale(topoDist) |
1562 |
|
1563 |
if ( iand(iHash, LJ_PAIR).ne.0 ) then |
1564 |
! this now calls a c routine so use c_idents instead of atids |
1565 |
call do_lj_pair(c_ident_i, c_ident_j, d, r, rijsq, rcut, sw, vdwMult, & |
1566 |
vpair, p_vdw, f1) |
1567 |
endif |
1568 |
|
1569 |
if ( iand(iHash, ELECTROSTATIC_PAIR).ne.0 ) then |
1570 |
call doElectrostaticPair(atid_i, atid_j, d, r, rijsq, rcut, sw, electroMult, & |
1571 |
vpair, p_elect, eF1, eF2, f1, t1, t2) |
1572 |
endif |
1573 |
|
1574 |
if ( iand(iHash, STICKY_PAIR).ne.0 ) then |
1575 |
call do_sticky_pair(c_ident_i, c_ident_j, d, r, rijsq, sw, vpair, & |
1576 |
p_hb, A1, A2, f1, t1, t2) |
1577 |
endif |
1578 |
|
1579 |
if ( iand(iHash, STICKYPOWER_PAIR).ne.0 ) then |
1580 |
! C++ sticky module now handles all sticky and sticky power interactions |
1581 |
call do_sticky_pair(c_ident_i, c_ident_j, d, r, rijsq, sw, vpair, & |
1582 |
p_hb, A1, A2, f1, t1, t2) |
1583 |
endif |
1584 |
|
1585 |
if ( iand(iHash, GAYBERNE_PAIR).ne.0 ) then |
1586 |
call do_gb_pair(c_ident_i, c_ident_j, d, r, rijsq, sw, vdwMult, vpair, & |
1587 |
p_vdw, A1, A2, f1, t1, t2) |
1588 |
endif |
1589 |
|
1590 |
if ( iand(iHash, GAYBERNE_LJ).ne.0 ) then |
1591 |
call do_gb_pair(c_ident_i, c_ident_j, d, r, rijsq, sw, vdwMult, vpair, & |
1592 |
p_vdw, A1, A2, f1, t1, t2) |
1593 |
endif |
1594 |
|
1595 |
if ( iand(iHash, SHAPE_PAIR).ne.0 ) then |
1596 |
call do_shape_pair(atid_i, atid_j, d, r, rijsq, sw, vpair, & |
1597 |
p_vdw, A1, A2, f1, t1, t2) |
1598 |
endif |
1599 |
|
1600 |
if ( iand(iHash, SHAPE_LJ).ne.0 ) then |
1601 |
call do_shape_pair(atid_i, atid_j, d, r, rijsq, sw, vpair, & |
1602 |
p_vdw, A1, A2, f1, t1, t2) |
1603 |
endif |
1604 |
|
1605 |
if ( iand(iHash, EAM_PAIR).ne.0 ) then |
1606 |
call do_eam_pair(c_ident_i, c_ident_j, d, r, rijsq, sw, vpair, p_met, & |
1607 |
f1, rho_i, rho_j, dfrhodrho_i, dfrhodrho_j, fshift_i, fshift_j) |
1608 |
endif |
1609 |
|
1610 |
if ( iand(iHash, SC_PAIR).ne.0 ) then |
1611 |
call do_SC_pair(c_ident_i, c_ident_j, d, r, rijsq, sw, vpair, p_met, & |
1612 |
f1, rho_i, rho_j, dfrhodrho_i, dfrhodrho_j, fshift_i, fshift_j) |
1613 |
endif |
1614 |
|
1615 |
if ( iand(iHash, MNM_PAIR).ne.0 ) then |
1616 |
call do_mnm_pair(atid_i, atid_j, d, r, rijsq, rcut, sw, vdwMult, vpair, & |
1617 |
p_vdw, A1, A2, f1, t1, t2) |
1618 |
endif |
1619 |
|
1620 |
|
1621 |
#ifdef IS_MPI |
1622 |
id1 = AtomRowToGlobal(i) |
1623 |
id2 = AtomColToGlobal(j) |
1624 |
|
1625 |
pot_row(VDW_POT,i) = pot_row(VDW_POT,i) + 0.5*p_vdw |
1626 |
pot_col(VDW_POT,j) = pot_col(VDW_POT,j) + 0.5*p_vdw |
1627 |
pot_row(ELECTROSTATIC_POT,i) = pot_row(ELECTROSTATIC_POT,i) + 0.5*p_elect |
1628 |
pot_col(ELECTROSTATIC_POT,j) = pot_col(ELECTROSTATIC_POT,j) + 0.5*p_elect |
1629 |
pot_row(HB_POT,i) = pot_row(HB_POT,i) + 0.5*p_hb |
1630 |
pot_col(HB_POT,j) = pot_col(HB_POT,j) + 0.5*p_hb |
1631 |
pot_Row(METALLIC_POT,i) = pot_Row(METALLIC_POT,i) + 0.5*p_met |
1632 |
pot_Col(METALLIC_POT,j) = pot_Col(METALLIC_POT,j) + 0.5*p_met |
1633 |
|
1634 |
do idx = 1, 3 |
1635 |
f_Row(idx,i) = f_Row(idx,i) + f1(idx) |
1636 |
f_Col(idx,j) = f_Col(idx,j) - f1(idx) |
1637 |
|
1638 |
t_Row(idx,i) = t_Row(idx,i) + t1(idx) |
1639 |
t_Col(idx,j) = t_Col(idx,j) + t2(idx) |
1640 |
enddo |
1641 |
! particle_pot is the difference between the full potential |
1642 |
! and the full potential without the presence of a particular |
1643 |
! particle (atom1). |
1644 |
! |
1645 |
! This reduces the density at other particle locations, so |
1646 |
! we need to recompute the density at atom2 assuming atom1 |
1647 |
! didn't contribute. This then requires recomputing the |
1648 |
! density functional for atom2 as well. |
1649 |
! |
1650 |
! Most of the particle_pot heavy lifting comes from the |
1651 |
! pair interaction, and will be handled by vpair. Parallel version. |
1652 |
|
1653 |
if ( (iand(iHash, EAM_PAIR).ne.0) .or. (iand(iHash, SC_PAIR).ne.0) ) then |
1654 |
ppot_row(i) = ppot_row(i) - frho_row(j) + fshift_j |
1655 |
ppot_col(j) = ppot_col(j) - frho_col(i) + fshift_i |
1656 |
end if |
1657 |
|
1658 |
#else |
1659 |
id1 = i |
1660 |
id2 = j |
1661 |
|
1662 |
pot(VDW_POT) = pot(VDW_POT) + p_vdw |
1663 |
pot(ELECTROSTATIC_POT) = pot(ELECTROSTATIC_POT) + p_elect |
1664 |
pot(HB_POT) = pot(HB_POT) + p_hb |
1665 |
pot(METALLIC_POT) = pot(METALLIC_POT) + p_met |
1666 |
|
1667 |
do idx = 1, 3 |
1668 |
f(idx,i) = f(idx,i) + f1(idx) |
1669 |
f(idx,j) = f(idx,j) - f1(idx) |
1670 |
|
1671 |
t(idx,i) = t(idx,i) + t1(idx) |
1672 |
t(idx,j) = t(idx,j) + t2(idx) |
1673 |
enddo |
1674 |
! particle_pot is the difference between the full potential |
1675 |
! and the full potential without the presence of a particular |
1676 |
! particle (atom1). |
1677 |
! |
1678 |
! This reduces the density at other particle locations, so |
1679 |
! we need to recompute the density at atom2 assuming atom1 |
1680 |
! didn't contribute. This then requires recomputing the |
1681 |
! density functional for atom2 as well. |
1682 |
! |
1683 |
! Most of the particle_pot heavy lifting comes from the |
1684 |
! pair interaction, and will be handled by vpair. NonParallel version. |
1685 |
|
1686 |
if ( (iand(iHash, EAM_PAIR).ne.0) .or. (iand(iHash, SC_PAIR).ne.0) ) then |
1687 |
particle_pot(i) = particle_pot(i) - frho(j) + fshift_j |
1688 |
particle_pot(j) = particle_pot(j) - frho(i) + fshift_i |
1689 |
end if |
1690 |
|
1691 |
|
1692 |
#endif |
1693 |
|
1694 |
if (molMembershipList(id1) .ne. molMembershipList(id2)) then |
1695 |
|
1696 |
fpair(1) = fpair(1) + f1(1) |
1697 |
fpair(2) = fpair(2) + f1(2) |
1698 |
fpair(3) = fpair(3) + f1(3) |
1699 |
|
1700 |
endif |
1701 |
|
1702 |
|
1703 |
!!$ |
1704 |
!!$ particle_pot(i) = particle_pot(i) + vpair*sw |
1705 |
!!$ particle_pot(j) = particle_pot(j) + vpair*sw |
1706 |
|
1707 |
end subroutine do_pair |
1708 |
|
1709 |
subroutine do_prepair(i, j, rijsq, d, sw, rcijsq, dc, rCut, & |
1710 |
eFrame, A, f, t, pot) |
1711 |
|
1712 |
real( kind = dp ) :: sw |
1713 |
real( kind = dp ), dimension(LR_POT_TYPES) :: pot |
1714 |
real( kind = dp ), dimension(9,nLocal) :: eFrame |
1715 |
real (kind=dp), dimension(9,nLocal) :: A |
1716 |
real (kind=dp), dimension(3,nLocal) :: f |
1717 |
real (kind=dp), dimension(3,nLocal) :: t |
1718 |
|
1719 |
integer, intent(in) :: i, j |
1720 |
real ( kind = dp ), intent(inout) :: rijsq, rcijsq, rCut |
1721 |
real ( kind = dp ) :: r, rc |
1722 |
real ( kind = dp ), intent(inout) :: d(3), dc(3) |
1723 |
real ( kind = dp ) :: rho_i_at_j, rho_j_at_i |
1724 |
integer :: atid_i, atid_j, iHash, c_ident_i, c_ident_j |
1725 |
|
1726 |
r = sqrt(rijsq) |
1727 |
|
1728 |
#ifdef IS_MPI |
1729 |
atid_i = atid_row(i) |
1730 |
atid_j = atid_col(j) |
1731 |
c_ident_i = c_idents_row(i) |
1732 |
c_ident_j = c_idents_col(j) |
1733 |
#else |
1734 |
atid_i = atid(i) |
1735 |
atid_j = atid(j) |
1736 |
c_ident_i = c_idents_local(i) |
1737 |
c_ident_j = c_idents_local(j) |
1738 |
#endif |
1739 |
rho_i_at_j = 0.0_dp |
1740 |
rho_j_at_i = 0.0_dp |
1741 |
|
1742 |
iHash = InteractionHash(atid_i, atid_j) |
1743 |
|
1744 |
if ( iand(iHash, EAM_PAIR).ne.0 ) then |
1745 |
call calc_EAM_prepair_rho(c_ident_i, c_ident_j, r, rho_i_at_j, rho_j_at_i) |
1746 |
endif |
1747 |
|
1748 |
if ( iand(iHash, SC_PAIR).ne.0 ) then |
1749 |
call calc_SC_prepair_rho(c_ident_i, c_ident_j, r, rho_i_at_j, rho_j_at_i) |
1750 |
endif |
1751 |
|
1752 |
if ( iand(iHash, EAM_PAIR).ne.0 .or. iand(iHash, SC_PAIR).ne.0 ) then |
1753 |
#ifdef IS_MPI |
1754 |
rho_col(j) = rho_col(j) + rho_i_at_j |
1755 |
rho_row(i) = rho_row(i) + rho_j_at_i |
1756 |
#else |
1757 |
rho(j) = rho(j) + rho_i_at_j |
1758 |
rho(i) = rho(i) + rho_j_at_i |
1759 |
#endif |
1760 |
endif |
1761 |
|
1762 |
end subroutine do_prepair |
1763 |
|
1764 |
|
1765 |
subroutine do_preforce(nlocal, pot, particle_pot) |
1766 |
integer :: nlocal |
1767 |
real( kind = dp ),dimension(LR_POT_TYPES) :: pot |
1768 |
real( kind = dp ),dimension(nlocal) :: particle_pot |
1769 |
integer :: sc_err = 0 |
1770 |
integer :: atid1, atom, c_ident1 |
1771 |
|
1772 |
if ((FF_uses_EAM .and. SIM_uses_EAM) .or. (FF_uses_SC .and. SIM_uses_SC)) then |
1773 |
|
1774 |
#ifdef IS_MPI |
1775 |
call scatter(rho_row,rho,plan_atom_row,sc_err) |
1776 |
if (sc_err /= 0 ) then |
1777 |
call handleError("do_preforce()", "Error scattering rho_row into rho") |
1778 |
endif |
1779 |
call scatter(rho_col,rho_tmp,plan_atom_col,sc_err) |
1780 |
if (sc_err /= 0 ) then |
1781 |
call handleError("do_preforce()", "Error scattering rho_col into rho") |
1782 |
endif |
1783 |
rho(1:nlocal) = rho(1:nlocal) + rho_tmp(1:nlocal) |
1784 |
#endif |
1785 |
|
1786 |
|
1787 |
do atom = 1, nlocal |
1788 |
c_ident1 = c_idents_local(atom) |
1789 |
|
1790 |
if (FF_uses_EAM .and. SIM_uses_EAM) then |
1791 |
|
1792 |
call calc_EAM_preforce_Frho(c_ident1, rho(atom), frho(atom), dfrhodrho(atom)) |
1793 |
|
1794 |
else if (FF_uses_SC .and. SIM_uses_SC) then |
1795 |
|
1796 |
call calc_SC_preforce_Frho(c_ident1, rho(atom), frho(atom), dfrhodrho(atom)) |
1797 |
endif |
1798 |
|
1799 |
pot(METALLIC_POT) = pot(METALLIC_POT) + frho(atom) |
1800 |
particle_pot(atom) = particle_pot(atom) + frho(atom) |
1801 |
end do |
1802 |
|
1803 |
#ifdef IS_MPI |
1804 |
!! communicate f(rho) and derivatives back into row and column arrays |
1805 |
call gather(frho,frho_row,plan_atom_row, sc_err) |
1806 |
if (sc_err /= 0) then |
1807 |
call handleError("do_preforce()","MPI gather frho_row failure") |
1808 |
endif |
1809 |
call gather(dfrhodrho,dfrhodrho_row,plan_atom_row, sc_err) |
1810 |
if (sc_err /= 0) then |
1811 |
call handleError("do_preforce()","MPI gather dfrhodrho_row failure") |
1812 |
endif |
1813 |
call gather(frho,frho_col,plan_atom_col, sc_err) |
1814 |
if (sc_err /= 0) then |
1815 |
call handleError("do_preforce()","MPI gather frho_col failure") |
1816 |
endif |
1817 |
call gather(dfrhodrho,dfrhodrho_col,plan_atom_col, sc_err) |
1818 |
if (sc_err /= 0) then |
1819 |
call handleError("do_preforce()","MPI gather dfrhodrho_col failure") |
1820 |
endif |
1821 |
#endif |
1822 |
|
1823 |
end if |
1824 |
end subroutine do_preforce |
1825 |
|
1826 |
|
1827 |
subroutine get_interatomic_vector(q_i, q_j, d, r_sq) |
1828 |
|
1829 |
real (kind = dp), dimension(3) :: q_i |
1830 |
real (kind = dp), dimension(3) :: q_j |
1831 |
real ( kind = dp ), intent(out) :: r_sq |
1832 |
real( kind = dp ) :: d(3), scaled(3) |
1833 |
integer i |
1834 |
|
1835 |
d(1) = q_j(1) - q_i(1) |
1836 |
d(2) = q_j(2) - q_i(2) |
1837 |
d(3) = q_j(3) - q_i(3) |
1838 |
|
1839 |
! Wrap back into periodic box if necessary |
1840 |
if ( SIM_uses_PBC ) then |
1841 |
|
1842 |
if( .not.boxIsOrthorhombic ) then |
1843 |
! calc the scaled coordinates. |
1844 |
! scaled = matmul(HmatInv, d) |
1845 |
|
1846 |
scaled(1) = HmatInv(1,1)*d(1) + HmatInv(1,2)*d(2) + HmatInv(1,3)*d(3) |
1847 |
scaled(2) = HmatInv(2,1)*d(1) + HmatInv(2,2)*d(2) + HmatInv(2,3)*d(3) |
1848 |
scaled(3) = HmatInv(3,1)*d(1) + HmatInv(3,2)*d(2) + HmatInv(3,3)*d(3) |
1849 |
|
1850 |
! wrap the scaled coordinates |
1851 |
|
1852 |
scaled(1) = scaled(1) - anint(scaled(1), kind=dp) |
1853 |
scaled(2) = scaled(2) - anint(scaled(2), kind=dp) |
1854 |
scaled(3) = scaled(3) - anint(scaled(3), kind=dp) |
1855 |
|
1856 |
! calc the wrapped real coordinates from the wrapped scaled |
1857 |
! coordinates |
1858 |
! d = matmul(Hmat,scaled) |
1859 |
d(1)= Hmat(1,1)*scaled(1) + Hmat(1,2)*scaled(2) + Hmat(1,3)*scaled(3) |
1860 |
d(2)= Hmat(2,1)*scaled(1) + Hmat(2,2)*scaled(2) + Hmat(2,3)*scaled(3) |
1861 |
d(3)= Hmat(3,1)*scaled(1) + Hmat(3,2)*scaled(2) + Hmat(3,3)*scaled(3) |
1862 |
|
1863 |
else |
1864 |
! calc the scaled coordinates. |
1865 |
|
1866 |
scaled(1) = d(1) * HmatInv(1,1) |
1867 |
scaled(2) = d(2) * HmatInv(2,2) |
1868 |
scaled(3) = d(3) * HmatInv(3,3) |
1869 |
|
1870 |
! wrap the scaled coordinates |
1871 |
|
1872 |
scaled(1) = scaled(1) - anint(scaled(1), kind=dp) |
1873 |
scaled(2) = scaled(2) - anint(scaled(2), kind=dp) |
1874 |
scaled(3) = scaled(3) - anint(scaled(3), kind=dp) |
1875 |
|
1876 |
! calc the wrapped real coordinates from the wrapped scaled |
1877 |
! coordinates |
1878 |
|
1879 |
d(1) = scaled(1)*Hmat(1,1) |
1880 |
d(2) = scaled(2)*Hmat(2,2) |
1881 |
d(3) = scaled(3)*Hmat(3,3) |
1882 |
|
1883 |
endif |
1884 |
|
1885 |
endif |
1886 |
|
1887 |
r_sq = d(1)*d(1) + d(2)*d(2) + d(3)*d(3) |
1888 |
|
1889 |
end subroutine get_interatomic_vector |
1890 |
|
1891 |
subroutine zero_work_arrays() |
1892 |
|
1893 |
#ifdef IS_MPI |
1894 |
|
1895 |
q_Row = 0.0_dp |
1896 |
q_Col = 0.0_dp |
1897 |
|
1898 |
q_group_Row = 0.0_dp |
1899 |
q_group_Col = 0.0_dp |
1900 |
|
1901 |
eFrame_Row = 0.0_dp |
1902 |
eFrame_Col = 0.0_dp |
1903 |
|
1904 |
A_Row = 0.0_dp |
1905 |
A_Col = 0.0_dp |
1906 |
|
1907 |
f_Row = 0.0_dp |
1908 |
f_Col = 0.0_dp |
1909 |
f_Temp = 0.0_dp |
1910 |
|
1911 |
t_Row = 0.0_dp |
1912 |
t_Col = 0.0_dp |
1913 |
t_Temp = 0.0_dp |
1914 |
|
1915 |
pot_Row = 0.0_dp |
1916 |
pot_Col = 0.0_dp |
1917 |
pot_Temp = 0.0_dp |
1918 |
ppot_Temp = 0.0_dp |
1919 |
|
1920 |
frho_row = 0.0_dp |
1921 |
frho_col = 0.0_dp |
1922 |
rho_row = 0.0_dp |
1923 |
rho_col = 0.0_dp |
1924 |
rho_tmp = 0.0_dp |
1925 |
dfrhodrho_row = 0.0_dp |
1926 |
dfrhodrho_col = 0.0_dp |
1927 |
|
1928 |
#endif |
1929 |
rho = 0.0_dp |
1930 |
frho = 0.0_dp |
1931 |
dfrhodrho = 0.0_dp |
1932 |
|
1933 |
end subroutine zero_work_arrays |
1934 |
|
1935 |
function skipThisPair(atom1, atom2) result(skip_it) |
1936 |
integer, intent(in) :: atom1 |
1937 |
integer, intent(in), optional :: atom2 |
1938 |
logical :: skip_it |
1939 |
integer :: unique_id_1, unique_id_2 |
1940 |
integer :: me_i,me_j |
1941 |
integer :: i |
1942 |
|
1943 |
skip_it = .false. |
1944 |
|
1945 |
!! there are a number of reasons to skip a pair or a particle |
1946 |
!! mostly we do this to exclude atoms who are involved in short |
1947 |
!! range interactions (bonds, bends, torsions), but we also need |
1948 |
!! to exclude some overcounted interactions that result from |
1949 |
!! the parallel decomposition |
1950 |
|
1951 |
#ifdef IS_MPI |
1952 |
!! in MPI, we have to look up the unique IDs for each atom |
1953 |
unique_id_1 = AtomRowToGlobal(atom1) |
1954 |
unique_id_2 = AtomColToGlobal(atom2) |
1955 |
!! this situation should only arise in MPI simulations |
1956 |
if (unique_id_1 == unique_id_2) then |
1957 |
skip_it = .true. |
1958 |
return |
1959 |
end if |
1960 |
|
1961 |
!! this prevents us from doing the pair on multiple processors |
1962 |
if (unique_id_1 < unique_id_2) then |
1963 |
if (mod(unique_id_1 + unique_id_2,2) == 0) then |
1964 |
skip_it = .true. |
1965 |
return |
1966 |
endif |
1967 |
else |
1968 |
if (mod(unique_id_1 + unique_id_2,2) == 1) then |
1969 |
skip_it = .true. |
1970 |
return |
1971 |
endif |
1972 |
endif |
1973 |
#else |
1974 |
!! in the normal loop, the atom numbers are unique |
1975 |
unique_id_1 = atom1 |
1976 |
unique_id_2 = atom2 |
1977 |
#endif |
1978 |
|
1979 |
#ifdef IS_MPI |
1980 |
do i = 1, nSkipsForRowAtom(atom1) |
1981 |
if (skipsForRowAtom(atom1, i) .eq. unique_id_2) then |
1982 |
skip_it = .true. |
1983 |
return |
1984 |
endif |
1985 |
end do |
1986 |
#else |
1987 |
do i = 1, nSkipsForLocalAtom(atom1) |
1988 |
if (skipsForLocalAtom(atom1, i) .eq. unique_id_2) then |
1989 |
skip_it = .true. |
1990 |
return |
1991 |
endif |
1992 |
end do |
1993 |
#endif |
1994 |
|
1995 |
return |
1996 |
end function skipThisPair |
1997 |
|
1998 |
function getTopoDistance(atom1, atom2) result(topoDist) |
1999 |
integer, intent(in) :: atom1 |
2000 |
integer, intent(in) :: atom2 |
2001 |
integer :: topoDist |
2002 |
integer :: unique_id_2 |
2003 |
integer :: i |
2004 |
|
2005 |
#ifdef IS_MPI |
2006 |
unique_id_2 = AtomColToGlobal(atom2) |
2007 |
#else |
2008 |
unique_id_2 = atom2 |
2009 |
#endif |
2010 |
|
2011 |
! zero is default for unconnected (i.e. normal) pair interactions |
2012 |
|
2013 |
topoDist = 0 |
2014 |
|
2015 |
do i = 1, nTopoPairsForAtom(atom1) |
2016 |
if (toposForAtom(atom1, i) .eq. unique_id_2) then |
2017 |
topoDist = topoDistance(atom1, i) |
2018 |
return |
2019 |
endif |
2020 |
end do |
2021 |
|
2022 |
return |
2023 |
end function getTopoDistance |
2024 |
|
2025 |
function FF_UsesDirectionalAtoms() result(doesit) |
2026 |
logical :: doesit |
2027 |
doesit = FF_uses_DirectionalAtoms |
2028 |
end function FF_UsesDirectionalAtoms |
2029 |
|
2030 |
function FF_RequiresPrepairCalc() result(doesit) |
2031 |
logical :: doesit |
2032 |
doesit = FF_uses_EAM .or. FF_uses_SC |
2033 |
end function FF_RequiresPrepairCalc |
2034 |
|
2035 |
#ifdef PROFILE |
2036 |
function getforcetime() result(totalforcetime) |
2037 |
real(kind=dp) :: totalforcetime |
2038 |
totalforcetime = forcetime |
2039 |
end function getforcetime |
2040 |
#endif |
2041 |
|
2042 |
!! This cleans componets of force arrays belonging only to fortran |
2043 |
|
2044 |
subroutine add_stress_tensor(dpair, fpair, tau) |
2045 |
|
2046 |
real( kind = dp ), dimension(3), intent(in) :: dpair, fpair |
2047 |
real( kind = dp ), dimension(9), intent(inout) :: tau |
2048 |
|
2049 |
! because the d vector is the rj - ri vector, and |
2050 |
! because fx, fy, fz are the force on atom i, we need a |
2051 |
! negative sign here: |
2052 |
|
2053 |
tau(1) = tau(1) - dpair(1) * fpair(1) |
2054 |
tau(2) = tau(2) - dpair(1) * fpair(2) |
2055 |
tau(3) = tau(3) - dpair(1) * fpair(3) |
2056 |
tau(4) = tau(4) - dpair(2) * fpair(1) |
2057 |
tau(5) = tau(5) - dpair(2) * fpair(2) |
2058 |
tau(6) = tau(6) - dpair(2) * fpair(3) |
2059 |
tau(7) = tau(7) - dpair(3) * fpair(1) |
2060 |
tau(8) = tau(8) - dpair(3) * fpair(2) |
2061 |
tau(9) = tau(9) - dpair(3) * fpair(3) |
2062 |
|
2063 |
end subroutine add_stress_tensor |
2064 |
|
2065 |
end module doForces |