Line data Source code
1 : !--------------------------------------------------------------------------------------------------!
2 : ! CP2K: A general program to perform molecular dynamics simulations !
3 : ! Copyright 2000-2024 CP2K developers group <https://cp2k.org> !
4 : ! !
5 : ! SPDX-License-Identifier: GPL-2.0-or-later !
6 : !--------------------------------------------------------------------------------------------------!
7 :
8 : MODULE qs_tddfpt2_fhxc
9 : USE admm_types, ONLY: admm_type
10 : USE cp_control_types, ONLY: dft_control_type,&
11 : stda_control_type
12 : USE cp_dbcsr_api, ONLY: &
13 : dbcsr_add, dbcsr_copy, dbcsr_create, dbcsr_deallocate_matrix, dbcsr_get_info, &
14 : dbcsr_p_type, dbcsr_release, dbcsr_set, dbcsr_type, dbcsr_type_symmetric
15 : USE cp_dbcsr_cp2k_link, ONLY: cp_dbcsr_alloc_block_from_nbl
16 : USE cp_dbcsr_operations, ONLY: copy_fm_to_dbcsr,&
17 : cp_dbcsr_plus_fm_fm_t,&
18 : cp_dbcsr_sm_fm_multiply
19 : USE cp_fm_types, ONLY: cp_fm_create,&
20 : cp_fm_get_info,&
21 : cp_fm_release,&
22 : cp_fm_type
23 : USE input_constants, ONLY: do_admm_aux_exch_func_none
24 : USE kinds, ONLY: default_string_length,&
25 : dp
26 : USE lri_environment_types, ONLY: lri_kind_type
27 : USE message_passing, ONLY: mp_para_env_type
28 : USE parallel_gemm_api, ONLY: parallel_gemm
29 : USE pw_env_types, ONLY: pw_env_get
30 : USE pw_methods, ONLY: pw_axpy,&
31 : pw_scale,&
32 : pw_zero
33 : USE pw_pool_types, ONLY: pw_pool_type
34 : USE pw_types, ONLY: pw_c1d_gs_type,&
35 : pw_r3d_rs_type
36 : USE qs_environment_types, ONLY: get_qs_env,&
37 : qs_environment_type
38 : USE qs_gapw_densities, ONLY: prepare_gapw_den
39 : USE qs_integrate_potential, ONLY: integrate_v_rspace,&
40 : integrate_v_rspace_one_center
41 : USE qs_kernel_types, ONLY: full_kernel_env_type
42 : USE qs_ks_atom, ONLY: update_ks_atom
43 : USE qs_rho_atom_types, ONLY: rho_atom_type
44 : USE qs_rho_methods, ONLY: qs_rho_update_rho,&
45 : qs_rho_update_tddfpt
46 : USE qs_rho_types, ONLY: qs_rho_get
47 : USE qs_tddfpt2_densities, ONLY: tddfpt_construct_aux_fit_density
48 : USE qs_tddfpt2_lri_utils, ONLY: tddfpt2_lri_Amat
49 : USE qs_tddfpt2_operators, ONLY: tddfpt_apply_coulomb,&
50 : tddfpt_apply_xc,&
51 : tddfpt_apply_xc_potential
52 : USE qs_tddfpt2_stda_types, ONLY: stda_env_type
53 : USE qs_tddfpt2_stda_utils, ONLY: stda_calculate_kernel
54 : USE qs_tddfpt2_subgroups, ONLY: tddfpt_subgroup_env_type
55 : USE qs_tddfpt2_types, ONLY: tddfpt_work_matrices
56 : USE qs_vxc_atom, ONLY: calculate_xc_2nd_deriv_atom
57 : USE task_list_types, ONLY: task_list_type
58 : #include "./base/base_uses.f90"
59 :
60 : IMPLICIT NONE
61 :
62 : PRIVATE
63 :
64 : CHARACTER(len=*), PARAMETER, PRIVATE :: moduleN = 'qs_tddfpt2_fhxc'
65 :
66 : INTEGER, PARAMETER, PRIVATE :: maxspins = 2
67 :
68 : PUBLIC :: fhxc_kernel, stda_kernel
69 :
70 : ! **************************************************************************************************
71 :
72 : CONTAINS
73 :
74 : ! **************************************************************************************************
75 : !> \brief Compute action matrix-vector products with the FHxc Kernel
76 : !> \param Aop_evects action of TDDFPT operator on trial vectors (modified on exit)
77 : !> \param evects TDDFPT trial vectors
78 : !> \param is_rks_triplets indicates that a triplet excited states calculation using
79 : !> spin-unpolarised molecular orbitals has been requested
80 : !> \param do_hfx flag that activates computation of exact-exchange terms
81 : !> \param do_admm ...
82 : !> \param qs_env Quickstep environment
83 : !> \param kernel_env kernel environment
84 : !> \param kernel_env_admm_aux kernel environment for ADMM correction
85 : !> \param sub_env parallel (sub)group environment
86 : !> \param work_matrices collection of work matrices (modified on exit)
87 : !> \param admm_symm use symmetric definition of ADMM kernel correction
88 : !> \param admm_xc_correction use ADMM XC kernel correction
89 : !> \param do_lrigpw ...
90 : !> \par History
91 : !> * 06.2016 created [Sergey Chulkov]
92 : !> * 03.2017 refactored [Sergey Chulkov]
93 : !> * 04.2019 refactored [JHU]
94 : ! **************************************************************************************************
95 3318 : SUBROUTINE fhxc_kernel(Aop_evects, evects, is_rks_triplets, &
96 : do_hfx, do_admm, qs_env, kernel_env, kernel_env_admm_aux, &
97 : sub_env, work_matrices, admm_symm, admm_xc_correction, do_lrigpw)
98 : TYPE(cp_fm_type), DIMENSION(:, :), INTENT(IN) :: Aop_evects, evects
99 : LOGICAL, INTENT(in) :: is_rks_triplets, do_hfx, do_admm
100 : TYPE(qs_environment_type), POINTER :: qs_env
101 : TYPE(full_kernel_env_type), POINTER :: kernel_env, kernel_env_admm_aux
102 : TYPE(tddfpt_subgroup_env_type), INTENT(in) :: sub_env
103 : TYPE(tddfpt_work_matrices), INTENT(inout) :: work_matrices
104 : LOGICAL, INTENT(in) :: admm_symm, admm_xc_correction, do_lrigpw
105 :
106 : CHARACTER(LEN=*), PARAMETER :: routineN = 'fhxc_kernel'
107 :
108 : CHARACTER(LEN=default_string_length) :: basis_type
109 : INTEGER :: handle, ikind, ispin, ivect, nao, &
110 : nao_aux, nkind, nspins, nvects
111 3318 : INTEGER, DIMENSION(:), POINTER :: blk_sizes
112 : INTEGER, DIMENSION(maxspins) :: nactive
113 : LOGICAL :: gapw, gapw_xc
114 : TYPE(admm_type), POINTER :: admm_env
115 : TYPE(cp_fm_type) :: work_aux_orb, work_orb_orb
116 3318 : TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: A_xc_munu_sub, rho_ia_ao, &
117 3318 : rho_ia_ao_aux_fit
118 : TYPE(dbcsr_type), POINTER :: dbwork
119 : TYPE(dft_control_type), POINTER :: dft_control
120 3318 : TYPE(lri_kind_type), DIMENSION(:), POINTER :: lri_v_int
121 : TYPE(mp_para_env_type), POINTER :: para_env
122 3318 : TYPE(pw_c1d_gs_type), DIMENSION(:), POINTER :: rho_ia_g, rho_ia_g_aux_fit
123 : TYPE(pw_pool_type), POINTER :: auxbas_pw_pool
124 3318 : TYPE(pw_r3d_rs_type), ALLOCATABLE, DIMENSION(:) :: V_rspace_sub
125 3318 : TYPE(pw_r3d_rs_type), DIMENSION(:), POINTER :: rho_ia_r, rho_ia_r_aux_fit
126 3318 : TYPE(rho_atom_type), DIMENSION(:), POINTER :: rho1_atom_set, rho_atom_set
127 : TYPE(task_list_type), POINTER :: task_list
128 :
129 3318 : CALL timeset(routineN, handle)
130 :
131 3318 : nspins = SIZE(evects, 1)
132 3318 : nvects = SIZE(evects, 2)
133 3318 : IF (do_admm) THEN
134 716 : CPASSERT(do_hfx)
135 716 : CPASSERT(ASSOCIATED(sub_env%admm_A))
136 : END IF
137 3318 : CALL get_qs_env(qs_env, dft_control=dft_control)
138 :
139 3318 : gapw = dft_control%qs_control%gapw
140 3318 : gapw_xc = dft_control%qs_control%gapw_xc
141 :
142 3318 : CALL cp_fm_get_info(evects(1, 1), nrow_global=nao)
143 7214 : DO ispin = 1, nspins
144 7214 : CALL cp_fm_get_info(evects(ispin, 1), ncol_global=nactive(ispin))
145 : END DO
146 :
147 : CALL qs_rho_get(work_matrices%rho_orb_struct_sub, rho_ao=rho_ia_ao, &
148 3318 : rho_g=rho_ia_g, rho_r=rho_ia_r)
149 3318 : IF (do_hfx .AND. do_admm) THEN
150 716 : CALL get_qs_env(qs_env, admm_env=admm_env)
151 : CALL qs_rho_get(work_matrices%rho_aux_fit_struct_sub, &
152 : rho_ao=rho_ia_ao_aux_fit, rho_g=rho_ia_g_aux_fit, &
153 716 : rho_r=rho_ia_r_aux_fit)
154 : END IF
155 :
156 10852 : DO ivect = 1, nvects
157 7534 : IF (ALLOCATED(work_matrices%evects_sub)) THEN
158 20 : IF (ASSOCIATED(work_matrices%evects_sub(1, ivect)%matrix_struct)) THEN
159 20 : DO ispin = 1, nspins
160 10 : CALL dbcsr_set(rho_ia_ao(ispin)%matrix, 0.0_dp)
161 : CALL cp_dbcsr_plus_fm_fm_t(rho_ia_ao(ispin)%matrix, &
162 : matrix_v=sub_env%mos_occ(ispin), &
163 : matrix_g=work_matrices%evects_sub(ispin, ivect), &
164 20 : ncol=nactive(ispin), symmetry_mode=1)
165 : END DO
166 : ELSE
167 : ! skip trial vectors which are assigned to different parallel groups
168 : CYCLE
169 : END IF
170 : ELSE
171 16500 : DO ispin = 1, nspins
172 8986 : CALL dbcsr_set(rho_ia_ao(ispin)%matrix, 0.0_dp)
173 : CALL cp_dbcsr_plus_fm_fm_t(rho_ia_ao(ispin)%matrix, &
174 : matrix_v=sub_env%mos_occ(ispin), &
175 : matrix_g=evects(ispin, ivect), &
176 16500 : ncol=nactive(ispin), symmetry_mode=1)
177 : END DO
178 : END IF
179 :
180 7524 : IF (do_lrigpw) THEN
181 : CALL qs_rho_update_tddfpt(work_matrices%rho_orb_struct_sub, qs_env, &
182 : pw_env_external=sub_env%pw_env, &
183 : task_list_external=sub_env%task_list_orb, &
184 : para_env_external=sub_env%para_env, &
185 : tddfpt_lri_env=kernel_env%lri_env, &
186 204 : tddfpt_lri_density=kernel_env%lri_density)
187 7320 : ELSEIF (dft_control%qs_control%lrigpw .OR. &
188 : dft_control%qs_control%rigpw) THEN
189 : CALL qs_rho_update_tddfpt(work_matrices%rho_orb_struct_sub, qs_env, &
190 : pw_env_external=sub_env%pw_env, &
191 : task_list_external=sub_env%task_list_orb, &
192 0 : para_env_external=sub_env%para_env)
193 : ELSE
194 7320 : IF (gapw) THEN
195 : CALL qs_rho_update_rho(work_matrices%rho_orb_struct_sub, qs_env, &
196 : local_rho_set=work_matrices%local_rho_set, &
197 : pw_env_external=sub_env%pw_env, &
198 : task_list_external=sub_env%task_list_orb_soft, &
199 1184 : para_env_external=sub_env%para_env)
200 : CALL prepare_gapw_den(qs_env, work_matrices%local_rho_set, &
201 1184 : do_rho0=(.NOT. is_rks_triplets))
202 6136 : ELSEIF (gapw_xc) THEN
203 : CALL qs_rho_update_rho(work_matrices%rho_orb_struct_sub, qs_env, &
204 : rho_xc_external=work_matrices%rho_xc_struct_sub, &
205 : local_rho_set=work_matrices%local_rho_set, &
206 : pw_env_external=sub_env%pw_env, &
207 : task_list_external=sub_env%task_list_orb, &
208 : task_list_external_soft=sub_env%task_list_orb_soft, &
209 222 : para_env_external=sub_env%para_env)
210 222 : CALL prepare_gapw_den(qs_env, work_matrices%local_rho_set, do_rho0=.FALSE.)
211 : ELSE
212 : CALL qs_rho_update_rho(work_matrices%rho_orb_struct_sub, qs_env, &
213 : pw_env_external=sub_env%pw_env, &
214 : task_list_external=sub_env%task_list_orb, &
215 5914 : para_env_external=sub_env%para_env)
216 : END IF
217 : END IF
218 :
219 16520 : DO ispin = 1, nspins
220 16520 : CALL dbcsr_set(work_matrices%A_ia_munu_sub(ispin)%matrix, 0.0_dp)
221 : END DO
222 :
223 : ! electron-hole exchange-correlation interaction
224 16520 : DO ispin = 1, nspins
225 16520 : CALL pw_zero(work_matrices%A_ia_rspace_sub(ispin))
226 : END DO
227 :
228 : ! C_x d^{2}E_{x}^{DFT}[\rho] / d\rho^2
229 : ! + C_{HF} d^{2}E_{x, ADMM}^{DFT}[\rho] / d\rho^2 in case of ADMM calculation
230 7524 : IF (gapw_xc) THEN
231 222 : IF (kernel_env%do_exck) THEN
232 0 : CPABORT("NYA")
233 : ELSE
234 : CALL tddfpt_apply_xc(A_ia_rspace=work_matrices%A_ia_rspace_sub, kernel_env=kernel_env, &
235 : rho_ia_struct=work_matrices%rho_xc_struct_sub, &
236 : is_rks_triplets=is_rks_triplets, pw_env=sub_env%pw_env, &
237 : work_v_xc=work_matrices%wpw_rspace_sub, &
238 222 : work_v_xc_tau=work_matrices%wpw_tau_rspace_sub)
239 : END IF
240 444 : DO ispin = 1, nspins
241 : CALL pw_scale(work_matrices%A_ia_rspace_sub(ispin), &
242 222 : work_matrices%A_ia_rspace_sub(ispin)%pw_grid%dvol)
243 : CALL integrate_v_rspace(v_rspace=work_matrices%A_ia_rspace_sub(ispin), &
244 : hmat=work_matrices%A_ia_munu_sub(ispin), &
245 : qs_env=qs_env, calculate_forces=.FALSE., gapw=gapw_xc, &
246 : pw_env_external=sub_env%pw_env, &
247 222 : task_list_external=sub_env%task_list_orb_soft)
248 444 : CALL pw_zero(work_matrices%A_ia_rspace_sub(ispin))
249 : END DO
250 : ELSE
251 7302 : IF (kernel_env%do_exck) THEN
252 : CALL tddfpt_apply_xc_potential(work_matrices%A_ia_rspace_sub, work_matrices%fxc_rspace_sub, &
253 186 : work_matrices%rho_orb_struct_sub, is_rks_triplets)
254 : ELSE
255 : CALL tddfpt_apply_xc(A_ia_rspace=work_matrices%A_ia_rspace_sub, kernel_env=kernel_env, &
256 : rho_ia_struct=work_matrices%rho_orb_struct_sub, &
257 : is_rks_triplets=is_rks_triplets, pw_env=sub_env%pw_env, &
258 : work_v_xc=work_matrices%wpw_rspace_sub, &
259 7116 : work_v_xc_tau=work_matrices%wpw_tau_rspace_sub)
260 : END IF
261 : END IF
262 7524 : IF (gapw .OR. gapw_xc) THEN
263 1406 : rho_atom_set => sub_env%local_rho_set%rho_atom_set
264 1406 : rho1_atom_set => work_matrices%local_rho_set%rho_atom_set
265 : CALL calculate_xc_2nd_deriv_atom(rho_atom_set, rho1_atom_set, qs_env, kernel_env%xc_section, &
266 1406 : sub_env%para_env, do_tddfpt2=.TRUE., do_triplet=is_rks_triplets)
267 : END IF
268 :
269 : ! ADMM correction
270 7524 : IF (do_admm .AND. admm_xc_correction) THEN
271 1256 : IF (dft_control%admm_control%aux_exch_func /= do_admm_aux_exch_func_none) THEN
272 : CALL tddfpt_construct_aux_fit_density(rho_orb_struct=work_matrices%rho_orb_struct_sub, &
273 : rho_aux_fit_struct=work_matrices%rho_aux_fit_struct_sub, &
274 : local_rho_set=work_matrices%local_rho_set_admm, &
275 : qs_env=qs_env, sub_env=sub_env, &
276 : wfm_rho_orb=work_matrices%rho_ao_orb_fm_sub, &
277 : wfm_rho_aux_fit=work_matrices%rho_ao_aux_fit_fm_sub, &
278 894 : wfm_aux_orb=work_matrices%wfm_aux_orb_sub)
279 : ! - C_{HF} d^{2}E_{x, ADMM}^{DFT}[\hat{\rho}] / d\hat{\rho}^2
280 894 : IF (admm_symm) THEN
281 894 : CALL dbcsr_get_info(rho_ia_ao_aux_fit(1)%matrix, row_blk_size=blk_sizes)
282 3576 : ALLOCATE (A_xc_munu_sub(nspins))
283 1788 : DO ispin = 1, nspins
284 894 : ALLOCATE (A_xc_munu_sub(ispin)%matrix)
285 : CALL dbcsr_create(matrix=A_xc_munu_sub(ispin)%matrix, name="ADMM_XC", &
286 : dist=sub_env%dbcsr_dist, matrix_type=dbcsr_type_symmetric, &
287 894 : row_blk_size=blk_sizes, col_blk_size=blk_sizes, nze=0)
288 894 : CALL cp_dbcsr_alloc_block_from_nbl(A_xc_munu_sub(ispin)%matrix, sub_env%sab_aux_fit)
289 1788 : CALL dbcsr_set(A_xc_munu_sub(ispin)%matrix, 0.0_dp)
290 : END DO
291 :
292 894 : CALL pw_env_get(sub_env%pw_env, auxbas_pw_pool=auxbas_pw_pool)
293 3576 : ALLOCATE (V_rspace_sub(nspins))
294 1788 : DO ispin = 1, nspins
295 894 : CALL auxbas_pw_pool%create_pw(V_rspace_sub(ispin))
296 1788 : CALL pw_zero(V_rspace_sub(ispin))
297 : END DO
298 :
299 894 : IF (admm_env%do_gapw) THEN
300 116 : basis_type = "AUX_FIT_SOFT"
301 116 : task_list => sub_env%task_list_aux_fit_soft
302 : ELSE
303 778 : basis_type = "AUX_FIT"
304 778 : task_list => sub_env%task_list_aux_fit
305 : END IF
306 :
307 : CALL tddfpt_apply_xc(A_ia_rspace=V_rspace_sub, &
308 : kernel_env=kernel_env_admm_aux, &
309 : rho_ia_struct=work_matrices%rho_aux_fit_struct_sub, &
310 : is_rks_triplets=is_rks_triplets, pw_env=sub_env%pw_env, &
311 : work_v_xc=work_matrices%wpw_rspace_sub, &
312 894 : work_v_xc_tau=work_matrices%wpw_tau_rspace_sub)
313 1788 : DO ispin = 1, nspins
314 894 : CALL pw_scale(V_rspace_sub(ispin), V_rspace_sub(ispin)%pw_grid%dvol)
315 : CALL integrate_v_rspace(v_rspace=V_rspace_sub(ispin), &
316 : hmat=A_xc_munu_sub(ispin), &
317 : qs_env=qs_env, calculate_forces=.FALSE., &
318 : pw_env_external=sub_env%pw_env, &
319 : basis_type=basis_type, &
320 1788 : task_list_external=task_list)
321 : END DO
322 894 : IF (admm_env%do_gapw) THEN
323 116 : rho_atom_set => sub_env%local_rho_set_admm%rho_atom_set
324 116 : rho1_atom_set => work_matrices%local_rho_set_admm%rho_atom_set
325 : CALL calculate_xc_2nd_deriv_atom(rho_atom_set, rho1_atom_set, qs_env, &
326 : kernel_env_admm_aux%xc_section, &
327 : sub_env%para_env, do_tddfpt2=.TRUE., &
328 : do_triplet=is_rks_triplets, &
329 116 : kind_set_external=admm_env%admm_gapw_env%admm_kind_set)
330 : CALL update_ks_atom(qs_env, A_xc_munu_sub, rho_ia_ao_aux_fit, forces=.FALSE., tddft=.TRUE., &
331 : rho_atom_external=rho1_atom_set, &
332 : kind_set_external=admm_env%admm_gapw_env%admm_kind_set, &
333 : oce_external=admm_env%admm_gapw_env%oce, &
334 116 : sab_external=sub_env%sab_aux_fit)
335 : END IF
336 894 : ALLOCATE (dbwork)
337 894 : CALL dbcsr_create(dbwork, template=work_matrices%A_ia_munu_sub(1)%matrix)
338 : CALL cp_fm_create(work_aux_orb, &
339 894 : matrix_struct=work_matrices%wfm_aux_orb_sub%matrix_struct)
340 : CALL cp_fm_create(work_orb_orb, &
341 894 : matrix_struct=work_matrices%rho_ao_orb_fm_sub%matrix_struct)
342 894 : CALL cp_fm_get_info(work_aux_orb, nrow_global=nao_aux, ncol_global=nao)
343 1788 : DO ispin = 1, nspins
344 : CALL cp_dbcsr_sm_fm_multiply(A_xc_munu_sub(ispin)%matrix, sub_env%admm_A, &
345 894 : work_aux_orb, nao)
346 : CALL parallel_gemm('T', 'N', nao, nao, nao_aux, 1.0_dp, sub_env%admm_A, &
347 894 : work_aux_orb, 0.0_dp, work_orb_orb)
348 894 : CALL dbcsr_copy(dbwork, work_matrices%A_ia_munu_sub(1)%matrix)
349 894 : CALL dbcsr_set(dbwork, 0.0_dp)
350 894 : CALL copy_fm_to_dbcsr(work_orb_orb, dbwork, keep_sparsity=.TRUE.)
351 1788 : CALL dbcsr_add(work_matrices%A_ia_munu_sub(ispin)%matrix, dbwork, 1.0_dp, 1.0_dp)
352 : END DO
353 894 : CALL dbcsr_release(dbwork)
354 894 : DEALLOCATE (dbwork)
355 1788 : DO ispin = 1, nspins
356 1788 : CALL auxbas_pw_pool%give_back_pw(V_rspace_sub(ispin))
357 : END DO
358 894 : DEALLOCATE (V_rspace_sub)
359 894 : CALL cp_fm_release(work_aux_orb)
360 894 : CALL cp_fm_release(work_orb_orb)
361 1788 : DO ispin = 1, nspins
362 1788 : CALL dbcsr_deallocate_matrix(A_xc_munu_sub(ispin)%matrix)
363 : END DO
364 1788 : DEALLOCATE (A_xc_munu_sub)
365 : ELSE
366 : CALL tddfpt_apply_xc(A_ia_rspace=work_matrices%A_ia_rspace_sub, &
367 : kernel_env=kernel_env_admm_aux, &
368 : rho_ia_struct=work_matrices%rho_aux_fit_struct_sub, &
369 : is_rks_triplets=is_rks_triplets, pw_env=sub_env%pw_env, &
370 : work_v_xc=work_matrices%wpw_rspace_sub, &
371 0 : work_v_xc_tau=work_matrices%wpw_tau_rspace_sub)
372 0 : IF (admm_env%do_gapw) THEN
373 0 : CPWARN("GAPW/ADMM needs symmetric ADMM kernel")
374 0 : CPABORT("GAPW/ADMM@TDDFT")
375 : END IF
376 : END IF
377 : END IF
378 : END IF
379 :
380 : ! electron-hole Coulomb interaction
381 7524 : IF (.NOT. is_rks_triplets) THEN
382 : ! a sum J_i{alpha}a{alpha}_munu + J_i{beta}a{beta}_munu can be computed by solving
383 : ! the Poisson equation for combined density (rho_{ia,alpha} + rho_{ia,beta}) .
384 : ! The following action will destroy reciprocal-space grid in spin-unrestricted case.
385 8326 : DO ispin = 2, nspins
386 8326 : CALL pw_axpy(rho_ia_g(ispin), rho_ia_g(1))
387 : END DO
388 : CALL tddfpt_apply_coulomb(A_ia_rspace=work_matrices%A_ia_rspace_sub, &
389 : rho_ia_g=rho_ia_g(1), &
390 : local_rho_set=work_matrices%local_rho_set, &
391 : hartree_local=work_matrices%hartree_local, &
392 : qs_env=qs_env, sub_env=sub_env, gapw=gapw, &
393 : work_v_gspace=work_matrices%wpw_gspace_sub(1), &
394 6854 : work_v_rspace=work_matrices%wpw_rspace_sub(1))
395 : END IF
396 :
397 : ! convert from the plane-wave representation into the Gaussian basis set representation
398 16520 : DO ispin = 1, nspins
399 16520 : IF (.NOT. do_lrigpw) THEN
400 : CALL pw_scale(work_matrices%A_ia_rspace_sub(ispin), &
401 8792 : work_matrices%A_ia_rspace_sub(ispin)%pw_grid%dvol)
402 :
403 8792 : IF (gapw) THEN
404 : CALL integrate_v_rspace(v_rspace=work_matrices%A_ia_rspace_sub(ispin), &
405 : hmat=work_matrices%A_ia_munu_sub(ispin), &
406 : qs_env=qs_env, calculate_forces=.FALSE., gapw=gapw, &
407 : pw_env_external=sub_env%pw_env, &
408 1184 : task_list_external=sub_env%task_list_orb_soft)
409 7608 : ELSEIF (gapw_xc) THEN
410 222 : IF (.NOT. is_rks_triplets) THEN
411 : CALL integrate_v_rspace(v_rspace=work_matrices%A_ia_rspace_sub(ispin), &
412 : hmat=work_matrices%A_ia_munu_sub(ispin), &
413 : qs_env=qs_env, calculate_forces=.FALSE., gapw=.FALSE., &
414 222 : pw_env_external=sub_env%pw_env, task_list_external=sub_env%task_list_orb)
415 : END IF
416 : ELSE
417 : CALL integrate_v_rspace(v_rspace=work_matrices%A_ia_rspace_sub(ispin), &
418 : hmat=work_matrices%A_ia_munu_sub(ispin), &
419 : qs_env=qs_env, calculate_forces=.FALSE., gapw=.FALSE., &
420 7386 : pw_env_external=sub_env%pw_env, task_list_external=sub_env%task_list_orb)
421 : END IF
422 : ELSE ! for full kernel using lri
423 : CALL pw_scale(work_matrices%A_ia_rspace_sub(ispin), &
424 204 : work_matrices%A_ia_rspace_sub(ispin)%pw_grid%dvol)
425 204 : lri_v_int => kernel_env%lri_density%lri_coefs(ispin)%lri_kinds
426 204 : CALL get_qs_env(qs_env, nkind=nkind, para_env=para_env)
427 612 : DO ikind = 1, nkind
428 122316 : lri_v_int(ikind)%v_int = 0.0_dp
429 : END DO
430 : CALL integrate_v_rspace_one_center(work_matrices%A_ia_rspace_sub(ispin), &
431 204 : qs_env, lri_v_int, .FALSE., "P_LRI_AUX")
432 612 : DO ikind = 1, nkind
433 244020 : CALL para_env%sum(lri_v_int(ikind)%v_int)
434 : END DO
435 : END IF ! for full kernel using lri
436 : END DO
437 :
438 : ! local atom contributions
439 7524 : IF (.NOT. do_lrigpw) THEN
440 7320 : IF (gapw .OR. gapw_xc) THEN
441 : ! rho_ia_ao will not be touched
442 : CALL update_ks_atom(qs_env, work_matrices%A_ia_munu_sub, rho_ia_ao, forces=.FALSE., &
443 : rho_atom_external=work_matrices%local_rho_set%rho_atom_set, &
444 1406 : tddft=.TRUE.)
445 : END IF
446 : END IF
447 :
448 : ! calculate Coulomb contribution to response vector for lrigpw !
449 : ! this is restricting lri to Coulomb only at the moment !
450 7524 : IF (do_lrigpw .AND. (.NOT. is_rks_triplets)) THEN !
451 204 : CALL tddfpt2_lri_Amat(qs_env, sub_env, kernel_env%lri_env, lri_v_int, work_matrices%A_ia_munu_sub)
452 : END IF
453 :
454 10842 : IF (ALLOCATED(work_matrices%evects_sub)) THEN
455 20 : DO ispin = 1, nspins
456 : CALL cp_dbcsr_sm_fm_multiply(work_matrices%A_ia_munu_sub(ispin)%matrix, &
457 : sub_env%mos_occ(ispin), &
458 : work_matrices%Aop_evects_sub(ispin, ivect), &
459 20 : ncol=nactive(ispin), alpha=1.0_dp, beta=0.0_dp)
460 : END DO
461 : ELSE
462 16500 : DO ispin = 1, nspins
463 : CALL cp_dbcsr_sm_fm_multiply(work_matrices%A_ia_munu_sub(ispin)%matrix, &
464 : sub_env%mos_occ(ispin), &
465 : Aop_evects(ispin, ivect), &
466 16510 : ncol=nactive(ispin), alpha=1.0_dp, beta=0.0_dp)
467 : END DO
468 : END IF
469 : END DO
470 :
471 3318 : CALL timestop(handle)
472 :
473 6636 : END SUBROUTINE fhxc_kernel
474 :
475 : ! **************************************************************************************************
476 : !> \brief Compute action matrix-vector products with the sTDA Kernel
477 : !> \param Aop_evects action of TDDFPT operator on trial vectors (modified on exit)
478 : !> \param evects TDDFPT trial vectors
479 : !> \param is_rks_triplets indicates that a triplet excited states calculation using
480 : !> spin-unpolarised molecular orbitals has been requested
481 : !> \param qs_env Quickstep environment
482 : !> \param stda_control control parameters for sTDA kernel
483 : !> \param stda_env ...
484 : !> \param sub_env parallel (sub)group environment
485 : !> \param work_matrices collection of work matrices (modified on exit)
486 : !> \par History
487 : !> * 04.2019 initial version [JHU]
488 : ! **************************************************************************************************
489 2514 : SUBROUTINE stda_kernel(Aop_evects, evects, is_rks_triplets, &
490 : qs_env, stda_control, stda_env, &
491 : sub_env, work_matrices)
492 :
493 : TYPE(cp_fm_type), DIMENSION(:, :), INTENT(IN) :: Aop_evects, evects
494 : LOGICAL, INTENT(in) :: is_rks_triplets
495 : TYPE(qs_environment_type), POINTER :: qs_env
496 : TYPE(stda_control_type) :: stda_control
497 : TYPE(stda_env_type) :: stda_env
498 : TYPE(tddfpt_subgroup_env_type) :: sub_env
499 : TYPE(tddfpt_work_matrices), INTENT(inout) :: work_matrices
500 :
501 : CHARACTER(LEN=*), PARAMETER :: routineN = 'stda_kernel'
502 :
503 : INTEGER :: handle, ivect, nvects
504 :
505 2514 : CALL timeset(routineN, handle)
506 :
507 2514 : nvects = SIZE(evects, 2)
508 :
509 9602 : DO ivect = 1, nvects
510 9602 : IF (ALLOCATED(work_matrices%evects_sub)) THEN
511 0 : IF (ASSOCIATED(work_matrices%evects_sub(1, ivect)%matrix_struct)) THEN
512 : CALL stda_calculate_kernel(qs_env, stda_control, stda_env, sub_env, work_matrices, &
513 : is_rks_triplets, work_matrices%evects_sub(:, ivect), &
514 0 : work_matrices%Aop_evects_sub(:, ivect))
515 : ELSE
516 : ! skip trial vectors which are assigned to different parallel groups
517 : CYCLE
518 : END IF
519 : ELSE
520 : CALL stda_calculate_kernel(qs_env, stda_control, stda_env, sub_env, work_matrices, &
521 7088 : is_rks_triplets, evects(:, ivect), Aop_evects(:, ivect))
522 : END IF
523 : END DO
524 :
525 2514 : CALL timestop(handle)
526 :
527 2514 : END SUBROUTINE stda_kernel
528 :
529 : ! **************************************************************************************************
530 :
531 : END MODULE qs_tddfpt2_fhxc
|
