Line data Source code
1 : !--------------------------------------------------------------------------------------------------!
2 : ! CP2K: A general program to perform molecular dynamics simulations !
3 : ! Copyright 2000-2026 CP2K developers group <https://cp2k.org> !
4 : ! !
5 : ! SPDX-License-Identifier: GPL-2.0-or-later !
6 : !--------------------------------------------------------------------------------------------------!
7 :
8 : MODULE qs_tddfpt2_properties
9 : USE atomic_kind_types, ONLY: atomic_kind_type
10 : USE bibliography, ONLY: Martin2003,&
11 : cite_reference
12 : USE bse_print, ONLY: print_exciton_descriptors
13 : USE bse_properties, ONLY: exciton_descr_type,&
14 : get_exciton_descriptors
15 : USE bse_util, ONLY: get_multipoles_mo
16 : USE cell_types, ONLY: cell_type
17 : USE cp_blacs_env, ONLY: cp_blacs_env_type
18 : USE cp_cfm_basic_linalg, ONLY: cp_cfm_solve
19 : USE cp_cfm_types, ONLY: cp_cfm_create,&
20 : cp_cfm_release,&
21 : cp_cfm_set_all,&
22 : cp_cfm_to_fm,&
23 : cp_cfm_type,&
24 : cp_fm_to_cfm
25 : USE cp_control_types, ONLY: dft_control_type,&
26 : tddfpt2_control_type
27 : USE cp_dbcsr_api, ONLY: &
28 : dbcsr_copy, dbcsr_get_block_p, dbcsr_get_info, dbcsr_init_p, dbcsr_iterator_blocks_left, &
29 : dbcsr_iterator_next_block, dbcsr_iterator_start, dbcsr_iterator_stop, dbcsr_iterator_type, &
30 : dbcsr_p_type, dbcsr_set, dbcsr_type
31 : USE cp_dbcsr_operations, ONLY: copy_dbcsr_to_fm,&
32 : copy_fm_to_dbcsr,&
33 : cp_dbcsr_sm_fm_multiply,&
34 : dbcsr_allocate_matrix_set,&
35 : dbcsr_deallocate_matrix_set
36 : USE cp_fm_basic_linalg, ONLY: cp_fm_scale,&
37 : cp_fm_scale_and_add,&
38 : cp_fm_trace
39 : USE cp_fm_diag, ONLY: choose_eigv_solver,&
40 : cp_fm_geeig
41 : USE cp_fm_struct, ONLY: cp_fm_struct_create,&
42 : cp_fm_struct_release,&
43 : cp_fm_struct_type
44 : USE cp_fm_types, ONLY: cp_fm_create,&
45 : cp_fm_get_info,&
46 : cp_fm_release,&
47 : cp_fm_set_all,&
48 : cp_fm_to_fm,&
49 : cp_fm_to_fm_submat_general,&
50 : cp_fm_type,&
51 : cp_fm_vectorsnorm
52 : USE cp_log_handling, ONLY: cp_get_default_logger,&
53 : cp_logger_get_default_io_unit,&
54 : cp_logger_type
55 : USE cp_output_handling, ONLY: cp_p_file,&
56 : cp_print_key_finished_output,&
57 : cp_print_key_should_output,&
58 : cp_print_key_unit_nr
59 : USE cp_realspace_grid_cube, ONLY: cp_pw_to_cube
60 : USE input_constants, ONLY: no_sf_tddfpt,&
61 : tddfpt_dipole_berry,&
62 : tddfpt_dipole_length,&
63 : tddfpt_dipole_scf_moment,&
64 : tddfpt_dipole_velocity,&
65 : tddfpt_dipole_velocity_old
66 : USE input_section_types, ONLY: section_vals_get_subs_vals,&
67 : section_vals_type,&
68 : section_vals_val_get
69 : USE kinds, ONLY: default_path_length,&
70 : dp,&
71 : int_8
72 : USE mathconstants, ONLY: twopi,&
73 : z_one,&
74 : z_zero
75 : USE message_passing, ONLY: mp_comm_type,&
76 : mp_para_env_type,&
77 : mp_request_type
78 : USE molden_utils, ONLY: write_mos_molden
79 : USE moments_utils, ONLY: get_reference_point
80 : USE parallel_gemm_api, ONLY: parallel_gemm
81 : USE particle_list_types, ONLY: particle_list_type
82 : USE particle_types, ONLY: particle_type
83 : USE physcon, ONLY: evolt
84 : USE pw_env_types, ONLY: pw_env_get,&
85 : pw_env_type
86 : USE pw_poisson_types, ONLY: pw_poisson_type
87 : USE pw_pool_types, ONLY: pw_pool_p_type,&
88 : pw_pool_type
89 : USE pw_types, ONLY: pw_c1d_gs_type,&
90 : pw_r3d_rs_type
91 : USE qs_collocate_density, ONLY: calculate_wavefunction
92 : USE qs_environment_types, ONLY: get_qs_env,&
93 : qs_environment_type
94 : USE qs_kind_types, ONLY: qs_kind_type
95 : USE qs_ks_types, ONLY: qs_ks_env_type
96 : USE qs_mo_types, ONLY: allocate_mo_set,&
97 : deallocate_mo_set,&
98 : get_mo_set,&
99 : init_mo_set,&
100 : mo_set_type,&
101 : set_mo_set
102 : USE qs_moments, ONLY: build_berry_moment_matrix
103 : USE qs_neighbor_list_types, ONLY: neighbor_list_set_p_type
104 : USE qs_operators_ao, ONLY: rRc_xyz_ao
105 : USE qs_overlap, ONLY: build_overlap_matrix
106 : USE qs_subsys_types, ONLY: qs_subsys_get,&
107 : qs_subsys_type
108 : USE qs_tddfpt2_types, ONLY: tddfpt_ground_state_mos
109 : USE string_utilities, ONLY: integer_to_string
110 : USE util, ONLY: sort
111 : #include "./base/base_uses.f90"
112 :
113 : IMPLICIT NONE
114 :
115 : PRIVATE
116 :
117 : CHARACTER(len=*), PARAMETER, PRIVATE :: moduleN = 'qs_tddfpt2_properties'
118 :
119 : ! number of first derivative components (3: d/dx, d/dy, d/dz)
120 : INTEGER, PARAMETER, PRIVATE :: nderivs = 3
121 : INTEGER, PARAMETER, PRIVATE :: maxspins = 2
122 :
123 : PUBLIC :: tddfpt_dipole_operator, tddfpt_print_summary, tddfpt_print_excitation_analysis, &
124 : tddfpt_print_nto_analysis, tddfpt_print_exciton_descriptors
125 :
126 : ! **************************************************************************************************
127 :
128 : CONTAINS
129 :
130 : ! **************************************************************************************************
131 : !> \brief Compute the action of the dipole operator on the ground state wave function.
132 : !> \param dipole_op_mos_occ 2-D array [x,y,z ; spin] of matrices where to put the computed quantity
133 : !> (allocated and initialised on exit)
134 : !> \param tddfpt_control TDDFPT control parameters
135 : !> \param gs_mos molecular orbitals optimised for the ground state
136 : !> \param qs_env Quickstep environment
137 : !> \par History
138 : !> * 05.2016 created as 'tddfpt_print_summary' [Sergey Chulkov]
139 : !> * 06.2018 dipole operator based on the Berry-phase formula [Sergey Chulkov]
140 : !> * 08.2018 splited of from 'tddfpt_print_summary' and merged with code from 'tddfpt'
141 : !> [Sergey Chulkov]
142 : !> \note \parblock
143 : !> Adapted version of the subroutine find_contributions() which was originally created
144 : !> by Thomas Chassaing on 02.2005.
145 : !>
146 : !> The relation between dipole integrals in velocity and length forms are the following:
147 : !> \f[<\psi_i|\nabla|\psi_a> = <\psi_i|\vec{r}|\hat{H}\psi_a> - <\hat{H}\psi_i|\vec{r}|\psi_a>
148 : !> = (\epsilon_a - \epsilon_i) <\psi_i|\vec{r}|\psi_a> .\f],
149 : !> due to the commutation identity:
150 : !> \f[\vec{r}\hat{H} - \hat{H}\vec{r} = [\vec{r},\hat{H}] = [\vec{r},-1/2 \nabla^2] = \nabla\f] .
151 : !> \endparblock
152 : ! **************************************************************************************************
153 1416 : SUBROUTINE tddfpt_dipole_operator(dipole_op_mos_occ, tddfpt_control, gs_mos, qs_env)
154 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:, :), &
155 : INTENT(inout) :: dipole_op_mos_occ
156 : TYPE(tddfpt2_control_type), POINTER :: tddfpt_control
157 : TYPE(tddfpt_ground_state_mos), DIMENSION(:), &
158 : INTENT(in) :: gs_mos
159 : TYPE(qs_environment_type), POINTER :: qs_env
160 :
161 : CHARACTER(LEN=*), PARAMETER :: routineN = 'tddfpt_dipole_operator'
162 :
163 : INTEGER :: handle, i_cos_sin, icol, ideriv, irow, &
164 : ispin, jderiv, nao, ncols_local, &
165 : ndim_periodic, nrows_local, nspins
166 1416 : INTEGER, DIMENSION(:), POINTER :: col_indices, row_indices
167 : INTEGER, DIMENSION(maxspins) :: nmo_occ, nmo_virt
168 : REAL(kind=dp) :: eval_occ
169 : REAL(kind=dp), CONTIGUOUS, DIMENSION(:, :), &
170 1416 : POINTER :: local_data_ediff, local_data_wfm
171 : REAL(kind=dp), DIMENSION(3) :: kvec, reference_point
172 : TYPE(cell_type), POINTER :: cell
173 : TYPE(cp_blacs_env_type), POINTER :: blacs_env
174 1416 : TYPE(cp_cfm_type), ALLOCATABLE, DIMENSION(:) :: gamma_00, gamma_inv_00
175 : TYPE(cp_fm_struct_type), POINTER :: fm_struct
176 : TYPE(cp_fm_type) :: ediff_inv, wfm_ao_ao, wfm_mo_virt_mo_occ
177 1416 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:) :: S_mos_virt
178 1416 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:, :) :: dBerry_mos_occ, gamma_real_imag, opvec
179 1416 : TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: berry_cossin_xyz, matrix_s, rRc_xyz, scrm
180 : TYPE(dft_control_type), POINTER :: dft_control
181 : TYPE(neighbor_list_set_p_type), DIMENSION(:), &
182 1416 : POINTER :: sab_orb
183 : TYPE(pw_env_type), POINTER :: pw_env
184 : TYPE(pw_poisson_type), POINTER :: poisson_env
185 : TYPE(qs_ks_env_type), POINTER :: ks_env
186 :
187 1416 : CALL timeset(routineN, handle)
188 :
189 1416 : NULLIFY (blacs_env, cell, matrix_s, pw_env)
190 1416 : CALL get_qs_env(qs_env, blacs_env=blacs_env, cell=cell, matrix_s=matrix_s, pw_env=pw_env)
191 :
192 1416 : nspins = SIZE(gs_mos)
193 1416 : CALL dbcsr_get_info(matrix_s(1)%matrix, nfullrows_total=nao)
194 3026 : DO ispin = 1, nspins
195 1610 : nmo_occ(ispin) = SIZE(gs_mos(ispin)%evals_occ)
196 3026 : nmo_virt(ispin) = SIZE(gs_mos(ispin)%evals_virt)
197 : END DO
198 :
199 : ! +++ allocate dipole operator matrices (must be deallocated elsewhere)
200 10688 : ALLOCATE (dipole_op_mos_occ(nderivs, nspins))
201 3026 : DO ispin = 1, nspins
202 1610 : CALL cp_fm_get_info(gs_mos(ispin)%mos_occ, matrix_struct=fm_struct)
203 :
204 7856 : DO ideriv = 1, nderivs
205 6440 : CALL cp_fm_create(dipole_op_mos_occ(ideriv, ispin), fm_struct)
206 : END DO
207 : END DO
208 :
209 : ! +++ allocate work matrices
210 5858 : ALLOCATE (S_mos_virt(nspins))
211 3026 : DO ispin = 1, nspins
212 1610 : CALL cp_fm_get_info(gs_mos(ispin)%mos_virt, matrix_struct=fm_struct)
213 1610 : CALL cp_fm_create(S_mos_virt(ispin), fm_struct)
214 : CALL cp_dbcsr_sm_fm_multiply(matrix_s(1)%matrix, &
215 : gs_mos(ispin)%mos_virt, &
216 : S_mos_virt(ispin), &
217 3026 : ncol=nmo_virt(ispin), alpha=1.0_dp, beta=0.0_dp)
218 : END DO
219 :
220 : ! check that the chosen dipole operator is consistent with the periodic boundary conditions used
221 1416 : CALL pw_env_get(pw_env, poisson_env=poisson_env)
222 5664 : ndim_periodic = COUNT(poisson_env%parameters%periodic == 1)
223 :
224 : ! select default for dipole form
225 1416 : IF (tddfpt_control%dipole_form == 0) THEN
226 644 : CALL get_qs_env(qs_env, dft_control=dft_control)
227 644 : IF (dft_control%qs_control%xtb) THEN
228 44 : IF (ndim_periodic == 0) THEN
229 0 : tddfpt_control%dipole_form = tddfpt_dipole_length
230 : ELSE
231 44 : tddfpt_control%dipole_form = tddfpt_dipole_velocity
232 : END IF
233 : ELSE
234 600 : tddfpt_control%dipole_form = tddfpt_dipole_velocity
235 : END IF
236 : END IF
237 :
238 1420 : SELECT CASE (tddfpt_control%dipole_form)
239 : CASE (tddfpt_dipole_berry)
240 4 : IF (ndim_periodic /= 3) THEN
241 : CALL cp_warn(__LOCATION__, &
242 : "Fully periodic Poisson solver (PERIODIC xyz) "// &
243 : "or a large supercell in non-periodic directions is needed "// &
244 0 : "for oscillator strengths based on the Berry phase formula")
245 : END IF
246 :
247 4 : NULLIFY (berry_cossin_xyz)
248 : ! index: 1 = Re[exp(-i * G_t * t)],
249 : ! 2 = Im[exp(-i * G_t * t)];
250 : ! t = x,y,z
251 4 : CALL dbcsr_allocate_matrix_set(berry_cossin_xyz, 2)
252 :
253 12 : DO i_cos_sin = 1, 2
254 8 : CALL dbcsr_init_p(berry_cossin_xyz(i_cos_sin)%matrix)
255 12 : CALL dbcsr_copy(berry_cossin_xyz(i_cos_sin)%matrix, matrix_s(1)%matrix)
256 : END DO
257 :
258 : ! +++ allocate berry-phase-related work matrices
259 72 : ALLOCATE (gamma_00(nspins), gamma_inv_00(nspins), gamma_real_imag(2, nspins), opvec(2, nspins))
260 32 : ALLOCATE (dBerry_mos_occ(nderivs, nspins))
261 10 : DO ispin = 1, nspins
262 6 : NULLIFY (fm_struct)
263 : CALL cp_fm_struct_create(fm_struct, nrow_global=nmo_occ(ispin), &
264 6 : ncol_global=nmo_occ(ispin), context=blacs_env)
265 :
266 6 : CALL cp_cfm_create(gamma_00(ispin), fm_struct)
267 6 : CALL cp_cfm_create(gamma_inv_00(ispin), fm_struct)
268 :
269 18 : DO i_cos_sin = 1, 2
270 18 : CALL cp_fm_create(gamma_real_imag(i_cos_sin, ispin), fm_struct)
271 : END DO
272 6 : CALL cp_fm_struct_release(fm_struct)
273 :
274 : ! G_real C_0, G_imag C_0
275 6 : CALL cp_fm_get_info(gs_mos(ispin)%mos_occ, matrix_struct=fm_struct)
276 18 : DO i_cos_sin = 1, 2
277 18 : CALL cp_fm_create(opvec(i_cos_sin, ispin), fm_struct)
278 : END DO
279 :
280 : ! dBerry * C_0
281 28 : DO ideriv = 1, nderivs
282 18 : CALL cp_fm_create(dBerry_mos_occ(ideriv, ispin), fm_struct)
283 24 : CALL cp_fm_set_all(dBerry_mos_occ(ideriv, ispin), 0.0_dp)
284 : END DO
285 : END DO
286 :
287 16 : DO ideriv = 1, nderivs
288 48 : kvec(:) = twopi*cell%h_inv(ideriv, :)
289 36 : DO i_cos_sin = 1, 2
290 36 : CALL dbcsr_set(berry_cossin_xyz(i_cos_sin)%matrix, 0.0_dp)
291 : END DO
292 : CALL build_berry_moment_matrix(qs_env, berry_cossin_xyz(1)%matrix, &
293 12 : berry_cossin_xyz(2)%matrix, kvec)
294 :
295 34 : DO ispin = 1, nspins
296 : ! i_cos_sin = 1: cos (real) component; opvec(1) = gamma_real C_0
297 : ! i_cos_sin = 2: sin (imaginary) component; opvec(2) = gamma_imag C_0
298 54 : DO i_cos_sin = 1, 2
299 : CALL cp_dbcsr_sm_fm_multiply(berry_cossin_xyz(i_cos_sin)%matrix, &
300 : gs_mos(ispin)%mos_occ, &
301 : opvec(i_cos_sin, ispin), &
302 54 : ncol=nmo_occ(ispin), alpha=1.0_dp, beta=0.0_dp)
303 : END DO
304 :
305 : CALL parallel_gemm('T', 'N', nmo_occ(ispin), nmo_occ(ispin), nao, &
306 : 1.0_dp, gs_mos(ispin)%mos_occ, opvec(1, ispin), &
307 18 : 0.0_dp, gamma_real_imag(1, ispin))
308 :
309 : CALL parallel_gemm('T', 'N', nmo_occ(ispin), nmo_occ(ispin), nao, &
310 : -1.0_dp, gs_mos(ispin)%mos_occ, opvec(2, ispin), &
311 18 : 0.0_dp, gamma_real_imag(2, ispin))
312 :
313 : CALL cp_fm_to_cfm(msourcer=gamma_real_imag(1, ispin), &
314 : msourcei=gamma_real_imag(2, ispin), &
315 18 : mtarget=gamma_00(ispin))
316 :
317 : ! gamma_inv_00 = Q = [C_0^T (gamma_real - i gamma_imag) C_0] ^ {-1}
318 18 : CALL cp_cfm_set_all(gamma_inv_00(ispin), z_zero, z_one)
319 18 : CALL cp_cfm_solve(gamma_00(ispin), gamma_inv_00(ispin))
320 :
321 : CALL cp_cfm_to_fm(msource=gamma_inv_00(ispin), &
322 : mtargetr=gamma_real_imag(1, ispin), &
323 18 : mtargeti=gamma_real_imag(2, ispin))
324 :
325 : ! dBerry_mos_occ is identical to dBerry_psi0 from qs_linres_op % polar_operators()
326 : CALL parallel_gemm("N", "N", nao, nmo_occ(ispin), nmo_occ(ispin), &
327 : 1.0_dp, opvec(1, ispin), gamma_real_imag(2, ispin), &
328 18 : 0.0_dp, dipole_op_mos_occ(1, ispin))
329 : CALL parallel_gemm("N", "N", nao, nmo_occ(ispin), nmo_occ(ispin), &
330 : -1.0_dp, opvec(2, ispin), gamma_real_imag(1, ispin), &
331 18 : 1.0_dp, dipole_op_mos_occ(1, ispin))
332 :
333 84 : DO jderiv = 1, nderivs
334 : CALL cp_fm_scale_and_add(1.0_dp, dBerry_mos_occ(jderiv, ispin), &
335 72 : cell%hmat(jderiv, ideriv), dipole_op_mos_occ(1, ispin))
336 : END DO
337 : END DO
338 : END DO
339 :
340 : ! --- release berry-phase-related work matrices
341 4 : CALL cp_fm_release(opvec)
342 4 : CALL cp_fm_release(gamma_real_imag)
343 10 : DO ispin = nspins, 1, -1
344 6 : CALL cp_cfm_release(gamma_inv_00(ispin))
345 10 : CALL cp_cfm_release(gamma_00(ispin))
346 : END DO
347 4 : DEALLOCATE (gamma_00, gamma_inv_00)
348 4 : CALL dbcsr_deallocate_matrix_set(berry_cossin_xyz)
349 :
350 : ! trans_dipole = 2|e|/|G_mu| * Tr Imag(evects^T * (gamma_real - i gamma_imag) * C_0 * gamma_inv_00) +
351 : ! 2|e|/|G_mu| * Tr Imag(C_0^T * (gamma_real - i gamma_imag) * evects * gamma_inv_00) ,
352 : !
353 : ! Taking into account the symmetry of the matrices 'gamma_real' and 'gamma_imag' and the fact
354 : ! that the response wave-function is a real-valued function, the above expression can be simplified as
355 : ! trans_dipole = 4|e|/|G_mu| * Tr Imag(evects^T * (gamma_real - i gamma_imag) * C_0 * gamma_inv_00)
356 : !
357 : ! 1/|G_mu| = |lattice_vector_mu| / (2*pi) .
358 10 : DO ispin = 1, nspins
359 :
360 28 : DO ideriv = 1, nderivs
361 24 : CALL cp_fm_to_fm(dBerry_mos_occ(ideriv, ispin), dipole_op_mos_occ(ideriv, ispin))
362 : END DO
363 : END DO
364 :
365 4 : CALL cp_fm_release(wfm_ao_ao)
366 4 : CALL cp_fm_release(dBerry_mos_occ)
367 :
368 : CASE (tddfpt_dipole_length)
369 20 : IF (ndim_periodic /= 0) THEN
370 : CALL cp_warn(__LOCATION__, &
371 : "Non-periodic Poisson solver (PERIODIC none) "// &
372 : "or a large supercell approach is needed "// &
373 4 : "for oscillator strengths based on the length operator")
374 : END IF
375 :
376 : ! compute components of the dipole operator in the length form
377 20 : NULLIFY (rRc_xyz)
378 20 : CALL dbcsr_allocate_matrix_set(rRc_xyz, nderivs)
379 :
380 80 : DO ideriv = 1, nderivs
381 60 : CALL dbcsr_init_p(rRc_xyz(ideriv)%matrix)
382 80 : CALL dbcsr_copy(rRc_xyz(ideriv)%matrix, matrix_s(1)%matrix)
383 : END DO
384 :
385 : CALL get_reference_point(reference_point, qs_env=qs_env, &
386 : reference=tddfpt_control%dipole_reference, &
387 20 : ref_point=tddfpt_control%dipole_ref_point)
388 :
389 : CALL rRc_xyz_ao(op=rRc_xyz, qs_env=qs_env, rc=reference_point, order=1, &
390 20 : minimum_image=.FALSE., soft=.FALSE.)
391 :
392 42 : DO ispin = 1, nspins
393 :
394 108 : DO ideriv = 1, nderivs
395 : CALL cp_dbcsr_sm_fm_multiply(rRc_xyz(ideriv)%matrix, &
396 : gs_mos(ispin)%mos_occ, &
397 : dipole_op_mos_occ(ideriv, ispin), &
398 88 : ncol=nmo_occ(ispin), alpha=1.0_dp, beta=0.0_dp)
399 : END DO
400 :
401 : END DO
402 :
403 20 : CALL dbcsr_deallocate_matrix_set(rRc_xyz)
404 :
405 : CASE (tddfpt_dipole_velocity)
406 : ! generate overlap derivatives
407 1388 : CALL get_qs_env(qs_env, ks_env=ks_env, sab_orb=sab_orb)
408 1388 : NULLIFY (scrm)
409 : CALL build_overlap_matrix(ks_env, matrix_s=scrm, nderivative=1, &
410 : basis_type_a="ORB", basis_type_b="ORB", &
411 1388 : sab_nl=sab_orb)
412 :
413 2964 : DO ispin = 1, nspins
414 6304 : DO ideriv = 1, nderivs
415 : CALL cp_dbcsr_sm_fm_multiply(scrm(ideriv + 1)%matrix, &
416 : gs_mos(ispin)%mos_occ, &
417 : dipole_op_mos_occ(ideriv, ispin), &
418 6304 : ncol=nmo_occ(ispin), alpha=1.0_dp, beta=0.0_dp)
419 : END DO
420 :
421 2964 : CALL cp_fm_release(wfm_mo_virt_mo_occ)
422 : END DO
423 1388 : CALL dbcsr_deallocate_matrix_set(scrm)
424 :
425 : CASE (tddfpt_dipole_velocity_old)
426 : ! generate overlap derivatives
427 4 : CALL get_qs_env(qs_env, ks_env=ks_env, sab_orb=sab_orb)
428 4 : NULLIFY (scrm)
429 : CALL build_overlap_matrix(ks_env, matrix_s=scrm, nderivative=1, &
430 : basis_type_a="ORB", basis_type_b="ORB", &
431 4 : sab_nl=sab_orb)
432 :
433 10 : DO ispin = 1, nspins
434 6 : NULLIFY (fm_struct)
435 : CALL cp_fm_struct_create(fm_struct, nrow_global=nmo_virt(ispin), &
436 6 : ncol_global=nmo_occ(ispin), context=blacs_env)
437 6 : CALL cp_fm_create(ediff_inv, fm_struct)
438 6 : CALL cp_fm_create(wfm_mo_virt_mo_occ, fm_struct)
439 6 : CALL cp_fm_struct_release(fm_struct)
440 :
441 : CALL cp_fm_get_info(ediff_inv, nrow_local=nrows_local, ncol_local=ncols_local, &
442 6 : row_indices=row_indices, col_indices=col_indices, local_data=local_data_ediff)
443 6 : CALL cp_fm_get_info(wfm_mo_virt_mo_occ, local_data=local_data_wfm)
444 :
445 : !$OMP PARALLEL DO DEFAULT(NONE), &
446 : !$OMP PRIVATE(eval_occ, icol, irow), &
447 6 : !$OMP SHARED(col_indices, gs_mos, ispin, local_data_ediff, ncols_local, nrows_local, row_indices)
448 : DO icol = 1, ncols_local
449 : ! E_occ_i ; imo_occ = col_indices(icol)
450 : eval_occ = gs_mos(ispin)%evals_occ(col_indices(icol))
451 :
452 : DO irow = 1, nrows_local
453 : ! ediff_inv_weights(a, i) = 1.0 / (E_virt_a - E_occ_i)
454 : ! imo_virt = row_indices(irow)
455 : local_data_ediff(irow, icol) = 1.0_dp/(gs_mos(ispin)%evals_virt(row_indices(irow)) - eval_occ)
456 : END DO
457 : END DO
458 : !$OMP END PARALLEL DO
459 :
460 24 : DO ideriv = 1, nderivs
461 : CALL cp_dbcsr_sm_fm_multiply(scrm(ideriv + 1)%matrix, &
462 : gs_mos(ispin)%mos_occ, &
463 : dipole_op_mos_occ(ideriv, ispin), &
464 18 : ncol=nmo_occ(ispin), alpha=1.0_dp, beta=0.0_dp)
465 :
466 : CALL parallel_gemm('T', 'N', nmo_virt(ispin), nmo_occ(ispin), nao, &
467 : 1.0_dp, gs_mos(ispin)%mos_virt, dipole_op_mos_occ(ideriv, ispin), &
468 18 : 0.0_dp, wfm_mo_virt_mo_occ)
469 :
470 : ! in-place element-wise (Schur) product;
471 : ! avoid allocation of a temporary [nmo_virt x nmo_occ] matrix which is needed
472 : ! for cp_fm_schur_product() subroutine call
473 :
474 : !$OMP PARALLEL DO DEFAULT(NONE), &
475 : !$OMP PRIVATE(icol, irow), &
476 18 : !$OMP SHARED(ispin, local_data_ediff, local_data_wfm, ncols_local, nrows_local)
477 : DO icol = 1, ncols_local
478 : DO irow = 1, nrows_local
479 : local_data_wfm(irow, icol) = local_data_wfm(irow, icol)*local_data_ediff(irow, icol)
480 : END DO
481 : END DO
482 : !$OMP END PARALLEL DO
483 :
484 : CALL parallel_gemm('N', 'N', nao, nmo_occ(ispin), nmo_virt(ispin), &
485 : 1.0_dp, S_mos_virt(ispin), wfm_mo_virt_mo_occ, &
486 24 : 0.0_dp, dipole_op_mos_occ(ideriv, ispin))
487 : END DO
488 :
489 6 : CALL cp_fm_release(wfm_mo_virt_mo_occ)
490 22 : CALL cp_fm_release(ediff_inv)
491 : END DO
492 4 : CALL dbcsr_deallocate_matrix_set(scrm)
493 :
494 : CASE DEFAULT
495 1416 : CPABORT("Unimplemented form of the dipole operator")
496 : END SELECT
497 :
498 : ! --- release work matrices
499 1416 : CALL cp_fm_release(S_mos_virt)
500 :
501 1416 : CALL timestop(handle)
502 4248 : END SUBROUTINE tddfpt_dipole_operator
503 :
504 : ! **************************************************************************************************
505 : !> \brief Print final TDDFPT excitation energies and oscillator strengths.
506 : !> \param log_unit output unit
507 : !> \param evects TDDFPT trial vectors (SIZE(evects,1) -- number of spins;
508 : !> SIZE(evects,2) -- number of excited states to print)
509 : !> \param evals TDDFPT eigenvalues
510 : !> \param gs_mos ...
511 : !> \param ostrength TDDFPT oscillator strength
512 : !> \param mult multiplicity
513 : !> \param dipole_op_mos_occ action of the dipole operator on the ground state wave function
514 : !> [x,y,z ; spin]
515 : !> \param dipole_form ...
516 : !> \par History
517 : !> * 05.2016 created [Sergey Chulkov]
518 : !> * 06.2016 transition dipole moments and oscillator strengths [Sergey Chulkov]
519 : !> * 07.2016 spin-unpolarised electron density [Sergey Chulkov]
520 : !> * 08.2018 compute 'dipole_op_mos_occ' in a separate subroutine [Sergey Chulkov]
521 : !> \note \parblock
522 : !> Adapted version of the subroutine find_contributions() which was originally created
523 : !> by Thomas Chassaing on 02.2005.
524 : !>
525 : !> Transition dipole moment along direction 'd' is computed as following:
526 : !> \f[ t_d(spin) = Tr[evects^T dipole\_op\_mos\_occ(d, spin)] .\f]
527 : !> \endparblock
528 : ! **************************************************************************************************
529 2852 : SUBROUTINE tddfpt_print_summary(log_unit, evects, evals, gs_mos, ostrength, mult, &
530 1426 : dipole_op_mos_occ, dipole_form)
531 : INTEGER, INTENT(in) :: log_unit
532 : TYPE(cp_fm_type), DIMENSION(:, :), INTENT(in) :: evects
533 : REAL(kind=dp), DIMENSION(:), INTENT(in) :: evals
534 : TYPE(tddfpt_ground_state_mos), DIMENSION(:), &
535 : POINTER :: gs_mos
536 : REAL(kind=dp), DIMENSION(:), INTENT(inout) :: ostrength
537 : INTEGER, INTENT(in) :: mult
538 : TYPE(cp_fm_type), DIMENSION(:, :), INTENT(in) :: dipole_op_mos_occ
539 : INTEGER, INTENT(in) :: dipole_form
540 :
541 : CHARACTER(LEN=*), PARAMETER :: routineN = 'tddfpt_print_summary'
542 :
543 : CHARACTER(len=1) :: lsd_str
544 : CHARACTER(len=20) :: mult_str
545 : INTEGER :: handle, i, ideriv, ispin, istate, j, &
546 : nactive, nao, nocc, nspins, nstates
547 : REAL(kind=dp) :: osc_strength
548 1426 : REAL(kind=dp), ALLOCATABLE, DIMENSION(:, :, :) :: trans_dipoles
549 : TYPE(cp_fm_struct_type), POINTER :: matrix_struct
550 : TYPE(cp_fm_type) :: dipact
551 :
552 1426 : CALL timeset(routineN, handle)
553 :
554 1426 : nspins = SIZE(evects, 1)
555 1426 : nstates = SIZE(evects, 2)
556 :
557 1426 : IF (nspins > 1) THEN
558 172 : lsd_str = 'U'
559 : ELSE
560 1254 : lsd_str = 'R'
561 : END IF
562 :
563 : ! *** summary header ***
564 1426 : IF (log_unit > 0) THEN
565 713 : CALL integer_to_string(mult, mult_str)
566 713 : WRITE (log_unit, '(/,1X,A1,A,1X,A)') lsd_str, "-TDDFPT states of multiplicity", TRIM(mult_str)
567 715 : SELECT CASE (dipole_form)
568 : CASE (tddfpt_dipole_berry)
569 2 : WRITE (log_unit, '(1X,A,/)') "Transition dipoles calculated using Berry operator formulation"
570 : CASE (tddfpt_dipole_length)
571 14 : WRITE (log_unit, '(1X,A,/)') "Transition dipoles calculated using length formulation"
572 : CASE (tddfpt_dipole_velocity)
573 695 : WRITE (log_unit, '(1X,A,/)') "Transition dipoles calculated using velocity formulation"
574 : CASE (tddfpt_dipole_velocity_old)
575 2 : WRITE (log_unit, '(1X,A,/)') "Transition dipoles calculated using old velocity formulation"
576 : CASE (tddfpt_dipole_scf_moment)
577 0 : WRITE (log_unit, '(1X,A,/)') "Transition dipoles calculated using SCF-MO moment formulation"
578 : CASE DEFAULT
579 713 : CPABORT("Unimplemented form of the dipole operator")
580 : END SELECT
581 :
582 713 : WRITE (log_unit, '(T10,A,T19,A,T37,A,T69,A)') "State", "Excitation", &
583 1426 : "Transition dipole (a.u.)", "Oscillator"
584 713 : WRITE (log_unit, '(T10,A,T19,A,T37,A,T49,A,T61,A,T67,A)') "number", "energy (eV)", &
585 1426 : "x", "y", "z", "strength (a.u.)"
586 713 : WRITE (log_unit, '(T10,72("-"))')
587 : END IF
588 :
589 : ! transition dipole moment
590 5704 : ALLOCATE (trans_dipoles(nstates, nderivs, nspins))
591 1426 : trans_dipoles(:, :, :) = 0.0_dp
592 :
593 : ! nspins == 1 .AND. mult == 3 : spin-flip transitions are forbidden due to symmetry reasons
594 1426 : IF (nspins > 1 .OR. mult == 1) THEN
595 2568 : DO ispin = 1, nspins
596 1370 : CALL cp_fm_get_info(dipole_op_mos_occ(1, ispin), nrow_global=nao, ncol_global=nocc)
597 1370 : CALL cp_fm_get_info(evects(ispin, 1), ncol_global=nactive)
598 2568 : IF (nocc == nactive) THEN
599 5416 : DO ideriv = 1, nderivs
600 : CALL cp_fm_trace(evects(ispin, :), dipole_op_mos_occ(ideriv, ispin), &
601 5416 : trans_dipoles(:, ideriv, ispin))
602 : END DO
603 : ELSE ! res
604 16 : CALL cp_fm_get_info(evects(ispin, 1), matrix_struct=matrix_struct)
605 16 : CALL cp_fm_create(dipact, matrix_struct)
606 64 : DO ideriv = 1, nderivs
607 144 : DO i = 1, nactive
608 96 : j = gs_mos(ispin)%index_active(i)
609 : CALL cp_fm_to_fm(dipole_op_mos_occ(ideriv, ispin), dipact, &
610 144 : ncol=1, source_start=j, target_start=i)
611 : END DO
612 64 : CALL cp_fm_trace(evects(ispin, :), dipact, trans_dipoles(:, ideriv, ispin))
613 : END DO
614 16 : CALL cp_fm_release(dipact)
615 : END IF
616 : END DO
617 :
618 1198 : IF (nspins == 1) THEN
619 11040 : trans_dipoles(:, :, 1) = SQRT(2.0_dp)*trans_dipoles(:, :, 1)
620 : ELSE
621 2500 : trans_dipoles(:, :, 1) = trans_dipoles(:, :, 1) + trans_dipoles(:, :, 2)
622 : END IF
623 : END IF
624 :
625 : ! *** summary information ***
626 5080 : DO istate = 1, nstates
627 :
628 3666 : SELECT CASE (dipole_form)
629 : CASE (tddfpt_dipole_berry)
630 48 : osc_strength = 2.0_dp/3.0_dp*evals(istate)*SUM(trans_dipoles(istate, :, 1)**2)
631 : CASE (tddfpt_dipole_length)
632 416 : osc_strength = 2.0_dp/3.0_dp*evals(istate)*SUM(trans_dipoles(istate, :, 1)**2)
633 : CASE (tddfpt_dipole_velocity)
634 14104 : osc_strength = 2.0_dp/3.0_dp/evals(istate)*SUM(trans_dipoles(istate, :, 1)**2)
635 : CASE (tddfpt_dipole_velocity_old)
636 48 : osc_strength = 2.0_dp/3.0_dp*evals(istate)*SUM(trans_dipoles(istate, :, 1)**2)
637 : CASE DEFAULT
638 3654 : CPABORT("Unimplemented form of the dipole operator")
639 : END SELECT
640 :
641 3654 : ostrength(istate) = osc_strength
642 5080 : IF (log_unit > 0) THEN
643 : WRITE (log_unit, '(1X,A,T9,I7,T19,F11.5,T31,3(1X,ES11.4E2),T69,ES12.5E2)') &
644 1827 : "TDDFPT|", istate, evals(istate)*evolt, trans_dipoles(istate, 1:nderivs, 1), osc_strength
645 : END IF
646 : END DO
647 :
648 : ! punch a checksum for the regs
649 1426 : IF (log_unit > 0) THEN
650 2540 : WRITE (log_unit, '(/,T2,A,E16.8)') 'TDDFPT : CheckSum E = ', SQRT(SUM(evals**2))
651 2540 : WRITE (log_unit, '(/,T2,A,E16.8)') 'TDDFPT : CheckSum F = ', SQRT(SUM(ostrength**2))
652 : END IF
653 :
654 1426 : DEALLOCATE (trans_dipoles)
655 :
656 1426 : CALL timestop(handle)
657 1426 : END SUBROUTINE tddfpt_print_summary
658 :
659 : ! **************************************************************************************************
660 : !> \brief Print excitation analysis.
661 : !> \param log_unit output unit
662 : !> \param evects TDDFPT trial vectors (SIZE(evects,1) -- number of spins;
663 : !> SIZE(evects,2) -- number of excited states to print)
664 : !> \param evals TDDFPT eigenvalues
665 : !> \param gs_mos molecular orbitals optimised for the ground state
666 : !> \param matrix_s overlap matrix
667 : !> \param spinflip ...
668 : !> \param min_amplitude the smallest excitation amplitude to print
669 : !> \par History
670 : !> * 05.2016 created as 'tddfpt_print_summary' [Sergey Chulkov]
671 : !> * 08.2018 splited of from 'tddfpt_print_summary' [Sergey Chulkov]
672 : ! **************************************************************************************************
673 1426 : SUBROUTINE tddfpt_print_excitation_analysis(log_unit, evects, evals, gs_mos, matrix_s, spinflip, &
674 : min_amplitude)
675 : INTEGER, INTENT(in) :: log_unit
676 : TYPE(cp_fm_type), DIMENSION(:, :), INTENT(in) :: evects
677 : REAL(kind=dp), DIMENSION(:), INTENT(in) :: evals
678 : TYPE(tddfpt_ground_state_mos), DIMENSION(:), &
679 : INTENT(in) :: gs_mos
680 : TYPE(dbcsr_type), POINTER :: matrix_s
681 : INTEGER :: spinflip
682 : REAL(kind=dp), INTENT(in) :: min_amplitude
683 :
684 : CHARACTER(LEN=*), PARAMETER :: routineN = 'tddfpt_print_excitation_analysis'
685 :
686 : CHARACTER(len=5) :: spin_label, spin_label2
687 : INTEGER :: handle, icol, iproc, irow, ispin, &
688 : istate, nao, ncols_local, nrows_local, &
689 : nspins, nstates, spin2, state_spin, &
690 : state_spin2
691 : INTEGER(kind=int_8) :: iexc, imo_act, imo_occ, imo_virt, ind, &
692 : nexcs, nexcs_local, nexcs_max_local, &
693 : nmo_virt_occ, nmo_virt_occ_alpha
694 1426 : INTEGER(kind=int_8), ALLOCATABLE, DIMENSION(:) :: inds_local, inds_recv, nexcs_recv
695 : INTEGER(kind=int_8), DIMENSION(1) :: nexcs_send
696 : INTEGER(kind=int_8), DIMENSION(maxspins) :: nactive8, nmo_occ8, nmo_virt8
697 1426 : INTEGER, ALLOCATABLE, DIMENSION(:) :: inds
698 1426 : INTEGER, DIMENSION(:), POINTER :: col_indices, row_indices
699 : INTEGER, DIMENSION(maxspins) :: nactive, nmo_occ, nmo_virt
700 : LOGICAL :: do_exc_analysis
701 1426 : REAL(kind=dp), ALLOCATABLE, DIMENSION(:) :: weights_local, weights_neg_abs_recv, &
702 1426 : weights_recv
703 : REAL(kind=dp), CONTIGUOUS, DIMENSION(:, :), &
704 1426 : POINTER :: local_data
705 : TYPE(cp_blacs_env_type), POINTER :: blacs_env
706 : TYPE(cp_fm_struct_type), POINTER :: fm_struct
707 1426 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:) :: S_mos_virt, weights_fm
708 : TYPE(mp_para_env_type), POINTER :: para_env
709 : TYPE(mp_request_type) :: send_handler, send_handler2
710 1426 : TYPE(mp_request_type), ALLOCATABLE, DIMENSION(:) :: recv_handlers, recv_handlers2
711 :
712 1426 : CALL timeset(routineN, handle)
713 :
714 1426 : nspins = SIZE(gs_mos, 1)
715 1426 : nstates = SIZE(evects, 2)
716 1426 : do_exc_analysis = min_amplitude < 1.0_dp
717 :
718 1426 : CALL cp_fm_get_info(gs_mos(1)%mos_occ, context=blacs_env, para_env=para_env)
719 1426 : CALL dbcsr_get_info(matrix_s, nfullrows_total=nao)
720 :
721 3046 : DO ispin = 1, nspins
722 1620 : nactive(ispin) = gs_mos(ispin)%nmo_active
723 : nactive8(ispin) = INT(nactive(ispin), kind=int_8)
724 1620 : nmo_occ(ispin) = SIZE(gs_mos(ispin)%evals_occ)
725 1620 : nmo_virt(ispin) = SIZE(gs_mos(ispin)%evals_virt)
726 1620 : nmo_occ8(ispin) = SIZE(gs_mos(ispin)%evals_occ, kind=int_8)
727 3046 : nmo_virt8(ispin) = SIZE(gs_mos(ispin)%evals_virt, kind=int_8)
728 : END DO
729 :
730 : ! *** excitation analysis ***
731 1426 : IF (do_exc_analysis) THEN
732 1426 : CPASSERT(log_unit <= 0 .OR. para_env%is_source())
733 1426 : nmo_virt_occ_alpha = INT(nmo_virt(1), int_8)*INT(nmo_occ(1), int_8)
734 :
735 1426 : IF (log_unit > 0) THEN
736 713 : WRITE (log_unit, "(1X,A)") "", &
737 713 : "-------------------------------------------------------------------------------", &
738 713 : "- Excitation analysis -", &
739 1426 : "-------------------------------------------------------------------------------"
740 713 : WRITE (log_unit, '(8X,A,T27,A,T49,A,T69,A)') "State", "Occupied", "Virtual", "Excitation"
741 713 : WRITE (log_unit, '(8X,A,T28,A,T49,A,T69,A)') "number", "orbital", "orbital", "amplitude"
742 713 : WRITE (log_unit, '(1X,79("-"))')
743 :
744 713 : IF (nspins == 1) THEN
745 616 : state_spin = 1
746 616 : state_spin2 = 2
747 616 : spin_label = ' '
748 616 : spin_label2 = ' '
749 97 : ELSE IF (spinflip /= no_sf_tddfpt) THEN
750 11 : state_spin = 1
751 11 : state_spin2 = 2
752 11 : spin_label = '(alp)'
753 11 : spin_label2 = '(bet)'
754 : END IF
755 : END IF
756 :
757 8900 : ALLOCATE (S_mos_virt(SIZE(evects, 1)), weights_fm(SIZE(evects, 1)))
758 3024 : DO ispin = 1, SIZE(evects, 1)
759 1598 : IF (spinflip == no_sf_tddfpt) THEN
760 : spin2 = ispin
761 : ELSE
762 22 : spin2 = 2
763 : END IF
764 1598 : CALL cp_fm_get_info(gs_mos(spin2)%mos_virt, matrix_struct=fm_struct)
765 1598 : CALL cp_fm_create(S_mos_virt(ispin), fm_struct)
766 : CALL cp_dbcsr_sm_fm_multiply(matrix_s, &
767 : gs_mos(spin2)%mos_virt, &
768 : S_mos_virt(ispin), &
769 1598 : ncol=nmo_virt(spin2), alpha=1.0_dp, beta=0.0_dp)
770 :
771 1598 : NULLIFY (fm_struct)
772 : CALL cp_fm_struct_create(fm_struct, nrow_global=nmo_virt(spin2), ncol_global=nactive(ispin), &
773 1598 : context=blacs_env)
774 1598 : CALL cp_fm_create(weights_fm(ispin), fm_struct)
775 1598 : CALL cp_fm_set_all(weights_fm(ispin), 0.0_dp)
776 3024 : CALL cp_fm_struct_release(fm_struct)
777 : END DO
778 :
779 3024 : nexcs_max_local = 0
780 3024 : DO ispin = 1, SIZE(evects, 1)
781 1598 : CALL cp_fm_get_info(weights_fm(ispin), nrow_local=nrows_local, ncol_local=ncols_local)
782 3024 : nexcs_max_local = nexcs_max_local + INT(nrows_local, int_8)*INT(ncols_local, int_8)
783 : END DO
784 :
785 5704 : ALLOCATE (weights_local(nexcs_max_local), inds_local(nexcs_max_local))
786 :
787 5080 : DO istate = 1, nstates
788 7912 : nexcs_local = 0
789 7912 : nmo_virt_occ = 0
790 :
791 : ! analyse matrix elements locally and transfer only significant
792 : ! excitations to the master node for subsequent ordering
793 7912 : DO ispin = 1, SIZE(evects, 1)
794 4258 : IF (spinflip == no_sf_tddfpt) THEN
795 : spin2 = ispin
796 : ELSE
797 90 : spin2 = 2
798 : END IF
799 : ! compute excitation amplitudes
800 : CALL parallel_gemm('T', 'N', nmo_virt(spin2), nactive(ispin), nao, 1.0_dp, S_mos_virt(ispin), &
801 4258 : evects(ispin, istate), 0.0_dp, weights_fm(ispin))
802 :
803 : CALL cp_fm_get_info(weights_fm(ispin), nrow_local=nrows_local, ncol_local=ncols_local, &
804 4258 : row_indices=row_indices, col_indices=col_indices, local_data=local_data)
805 :
806 : ! locate single excitations with significant amplitudes (>= min_amplitude)
807 23404 : DO icol = 1, ncols_local
808 237641 : DO irow = 1, nrows_local
809 233383 : IF (ABS(local_data(irow, icol)) >= min_amplitude) THEN
810 : ! number of non-negligible excitations
811 2925 : nexcs_local = nexcs_local + 1
812 : ! excitation amplitude
813 2925 : weights_local(nexcs_local) = local_data(irow, icol)
814 : ! index of single excitation (ivirt, iocc, ispin) in compressed form
815 : inds_local(nexcs_local) = nmo_virt_occ + INT(row_indices(irow), int_8) + &
816 2925 : INT(col_indices(icol) - 1, int_8)*nmo_virt8(spin2)
817 : END IF
818 : END DO
819 : END DO
820 :
821 12170 : nmo_virt_occ = nmo_virt_occ + nmo_virt8(spin2)*nmo_occ8(ispin)
822 : END DO
823 :
824 3654 : IF (para_env%is_source()) THEN
825 : ! master node
826 18270 : ALLOCATE (nexcs_recv(para_env%num_pe), recv_handlers(para_env%num_pe), recv_handlers2(para_env%num_pe))
827 :
828 : ! collect number of non-negligible excitations from other nodes
829 5481 : DO iproc = 1, para_env%num_pe
830 5481 : IF (iproc - 1 /= para_env%mepos) THEN
831 1827 : CALL para_env%irecv(nexcs_recv(iproc:iproc), iproc - 1, recv_handlers(iproc), 0)
832 : ELSE
833 1827 : nexcs_recv(iproc) = nexcs_local
834 : END IF
835 : END DO
836 :
837 5481 : DO iproc = 1, para_env%num_pe
838 5481 : IF (iproc - 1 /= para_env%mepos) THEN
839 1827 : CALL recv_handlers(iproc)%wait()
840 : END IF
841 : END DO
842 :
843 : ! compute total number of non-negligible excitations
844 1827 : nexcs = 0
845 5481 : DO iproc = 1, para_env%num_pe
846 5481 : nexcs = nexcs + nexcs_recv(iproc)
847 : END DO
848 :
849 : ! receive indices and amplitudes of selected excitations
850 7308 : ALLOCATE (weights_recv(nexcs), weights_neg_abs_recv(nexcs))
851 7308 : ALLOCATE (inds_recv(nexcs), inds(nexcs))
852 :
853 5481 : nmo_virt_occ = 0
854 5481 : DO iproc = 1, para_env%num_pe
855 5481 : IF (nexcs_recv(iproc) > 0) THEN
856 1915 : IF (iproc - 1 /= para_env%mepos) THEN
857 : ! excitation amplitudes
858 : CALL para_env%irecv(weights_recv(nmo_virt_occ + 1:nmo_virt_occ + nexcs_recv(iproc)), &
859 253 : iproc - 1, recv_handlers(iproc), 1)
860 : ! compressed indices
861 : CALL para_env%irecv(inds_recv(nmo_virt_occ + 1:nmo_virt_occ + nexcs_recv(iproc)), &
862 253 : iproc - 1, recv_handlers2(iproc), 2)
863 : ELSE
864 : ! data on master node
865 4294 : weights_recv(nmo_virt_occ + 1:nmo_virt_occ + nexcs_recv(iproc)) = weights_local(1:nexcs_recv(iproc))
866 4294 : inds_recv(nmo_virt_occ + 1:nmo_virt_occ + nexcs_recv(iproc)) = inds_local(1:nexcs_recv(iproc))
867 : END IF
868 :
869 1915 : nmo_virt_occ = nmo_virt_occ + nexcs_recv(iproc)
870 : END IF
871 : END DO
872 :
873 5481 : DO iproc = 1, para_env%num_pe
874 5481 : IF (iproc - 1 /= para_env%mepos .AND. nexcs_recv(iproc) > 0) THEN
875 253 : CALL recv_handlers(iproc)%wait()
876 253 : CALL recv_handlers2(iproc)%wait()
877 : END IF
878 : END DO
879 :
880 1827 : DEALLOCATE (nexcs_recv, recv_handlers, recv_handlers2)
881 : ELSE
882 : ! working node: send the number of selected excited states to the master node
883 1827 : nexcs_send(1) = nexcs_local
884 1827 : CALL para_env%isend(nexcs_send, para_env%source, send_handler, 0)
885 1827 : CALL send_handler%wait()
886 :
887 1827 : IF (nexcs_local > 0) THEN
888 : ! send excitation amplitudes
889 253 : CALL para_env%isend(weights_local(1:nexcs_local), para_env%source, send_handler, 1)
890 : ! send compressed indices
891 253 : CALL para_env%isend(inds_local(1:nexcs_local), para_env%source, send_handler2, 2)
892 :
893 253 : CALL send_handler%wait()
894 253 : CALL send_handler2%wait()
895 : END IF
896 : END IF
897 :
898 : ! sort non-negligible excitations on the master node according to their amplitudes,
899 : ! uncompress indices and print summary information
900 3654 : IF (para_env%is_source() .AND. log_unit > 0) THEN
901 4752 : weights_neg_abs_recv(:) = -ABS(weights_recv)
902 1827 : CALL sort(weights_neg_abs_recv, INT(nexcs), inds)
903 :
904 1827 : WRITE (log_unit, '(T7,I8,F10.5,A)') istate, evals(istate)*evolt, " eV"
905 :
906 : ! This reinitialization is needed to prevent the intel fortran compiler from introduce
907 : ! a bug when using optimization level 3 flag
908 1827 : state_spin = 1
909 1827 : state_spin2 = 1
910 1827 : IF (spinflip /= no_sf_tddfpt) THEN
911 45 : state_spin = 1
912 45 : state_spin2 = 2
913 : END IF
914 4752 : DO iexc = 1, nexcs
915 2925 : ind = inds_recv(inds(iexc)) - 1
916 2925 : IF ((nspins > 1) .AND. (spinflip == no_sf_tddfpt)) THEN
917 627 : IF (ind < nmo_virt_occ_alpha) THEN
918 285 : state_spin = 1
919 285 : state_spin2 = 1
920 285 : spin_label = '(alp)'
921 285 : spin_label2 = '(alp)'
922 : ELSE
923 342 : state_spin = 2
924 342 : state_spin2 = 2
925 342 : ind = ind - nmo_virt_occ_alpha
926 342 : spin_label = '(bet)'
927 342 : spin_label2 = '(bet)'
928 : END IF
929 : END IF
930 2925 : imo_act = ind/nmo_virt8(state_spin2) + 1
931 2925 : imo_occ = gs_mos(state_spin)%index_active(imo_act)
932 2925 : imo_virt = MOD(ind, nmo_virt8(state_spin2)) + 1
933 :
934 2925 : WRITE (log_unit, '(T27,I8,1X,A5,T48,I8,1X,A5,T70,F9.6)') imo_occ, spin_label, &
935 7677 : nmo_occ8(state_spin2) + imo_virt, spin_label2, weights_recv(inds(iexc))
936 : END DO
937 : END IF
938 :
939 : ! deallocate temporary arrays
940 5080 : IF (para_env%is_source()) THEN
941 1827 : DEALLOCATE (weights_recv, weights_neg_abs_recv, inds_recv, inds)
942 : END IF
943 : END DO
944 :
945 1426 : DEALLOCATE (weights_local, inds_local)
946 1426 : IF (log_unit > 0) THEN
947 : WRITE (log_unit, "(1X,A)") &
948 713 : "-------------------------------------------------------------------------------"
949 : END IF
950 : END IF
951 :
952 1426 : CALL cp_fm_release(weights_fm)
953 1426 : CALL cp_fm_release(S_mos_virt)
954 :
955 1426 : CALL timestop(handle)
956 :
957 2852 : END SUBROUTINE tddfpt_print_excitation_analysis
958 :
959 : ! **************************************************************************************************
960 : !> \brief Print natural transition orbital analysis.
961 : !> \param qs_env Information on Kinds and Particles
962 : !> \param evects TDDFPT trial vectors (SIZE(evects,1) -- number of spins;
963 : !> SIZE(evects,2) -- number of excited states to print)
964 : !> \param evals TDDFPT eigenvalues
965 : !> \param ostrength ...
966 : !> \param gs_mos molecular orbitals optimised for the ground state
967 : !> \param matrix_s overlap matrix
968 : !> \param print_section ...
969 : !> \par History
970 : !> * 06.2019 created [JGH]
971 : ! **************************************************************************************************
972 1426 : SUBROUTINE tddfpt_print_nto_analysis(qs_env, evects, evals, ostrength, gs_mos, matrix_s, print_section)
973 : TYPE(qs_environment_type), POINTER :: qs_env
974 : TYPE(cp_fm_type), DIMENSION(:, :), INTENT(in) :: evects
975 : REAL(kind=dp), DIMENSION(:), INTENT(in) :: evals, ostrength
976 : TYPE(tddfpt_ground_state_mos), DIMENSION(:), &
977 : INTENT(in) :: gs_mos
978 : TYPE(dbcsr_type), POINTER :: matrix_s
979 : TYPE(section_vals_type), POINTER :: print_section
980 :
981 : CHARACTER(LEN=*), PARAMETER :: routineN = 'tddfpt_print_nto_analysis'
982 : INTEGER, PARAMETER :: ntomax = 10
983 :
984 : CHARACTER(LEN=20), DIMENSION(2) :: nto_name
985 : INTEGER :: handle, i, ia, icg, iounit, ispin, &
986 : istate, j, nao, nlist, nmax, nmo, &
987 : nnto, nspins, nstates
988 : INTEGER, DIMENSION(2) :: iv
989 : INTEGER, DIMENSION(2, ntomax) :: ia_index
990 1426 : INTEGER, DIMENSION(:), POINTER :: slist, stride
991 : LOGICAL :: append_cube, cube_file, explicit
992 : REAL(KIND=dp) :: os_threshold, sume, threshold
993 1426 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:) :: eigvals
994 1426 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :) :: eigenvalues
995 : REAL(KIND=dp), DIMENSION(ntomax) :: ia_eval
996 : TYPE(cell_type), POINTER :: cell
997 : TYPE(cp_fm_struct_type), POINTER :: fm_mo_struct, fm_struct
998 : TYPE(cp_fm_type) :: Sev, smat, tmat, wmat, work, wvec
999 1426 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:) :: teig
1000 : TYPE(cp_logger_type), POINTER :: logger
1001 1426 : TYPE(mo_set_type), ALLOCATABLE, DIMENSION(:) :: nto_set
1002 1426 : TYPE(particle_type), DIMENSION(:), POINTER :: particle_set
1003 1426 : TYPE(qs_kind_type), DIMENSION(:), POINTER :: qs_kind_set
1004 : TYPE(section_vals_type), POINTER :: molden_section, nto_section
1005 :
1006 1426 : CALL timeset(routineN, handle)
1007 :
1008 1426 : logger => cp_get_default_logger()
1009 1426 : iounit = cp_logger_get_default_io_unit(logger)
1010 :
1011 1426 : IF (BTEST(cp_print_key_should_output(logger%iter_info, print_section, &
1012 : "NTO_ANALYSIS"), cp_p_file)) THEN
1013 :
1014 224 : CALL cite_reference(Martin2003)
1015 :
1016 224 : CALL section_vals_val_get(print_section, "NTO_ANALYSIS%THRESHOLD", r_val=threshold)
1017 224 : CALL section_vals_val_get(print_section, "NTO_ANALYSIS%INTENSITY_THRESHOLD", r_val=os_threshold)
1018 224 : CALL section_vals_val_get(print_section, "NTO_ANALYSIS%STATE_LIST", EXPLICIT=explicit)
1019 :
1020 224 : IF (explicit) THEN
1021 4 : CALL section_vals_val_get(print_section, "NTO_ANALYSIS%STATE_LIST", i_vals=slist)
1022 4 : nlist = SIZE(slist)
1023 : ELSE
1024 : nlist = 0
1025 : END IF
1026 :
1027 224 : IF (iounit > 0) THEN
1028 112 : WRITE (iounit, "(1X,A)") "", &
1029 112 : "-------------------------------------------------------------------------------", &
1030 112 : "- Natural Orbital analysis -", &
1031 224 : "-------------------------------------------------------------------------------"
1032 : END IF
1033 :
1034 224 : nspins = SIZE(evects, 1)
1035 224 : nstates = SIZE(evects, 2)
1036 224 : CALL dbcsr_get_info(matrix_s, nfullrows_total=nao)
1037 :
1038 548 : DO istate = 1, nstates
1039 324 : IF (os_threshold > ostrength(istate)) THEN
1040 54 : IF (iounit > 0) THEN
1041 27 : WRITE (iounit, "(1X,A,I6)") " Skipping state ", istate
1042 : END IF
1043 : CYCLE
1044 : END IF
1045 270 : IF (nlist > 0) THEN
1046 0 : IF (.NOT. ANY(slist == istate)) THEN
1047 0 : IF (iounit > 0) THEN
1048 0 : WRITE (iounit, "(1X,A,I6)") " Skipping state ", istate
1049 : END IF
1050 : CYCLE
1051 : END IF
1052 : END IF
1053 270 : IF (iounit > 0) THEN
1054 135 : WRITE (iounit, "(1X,A,I6,T30,F10.5,A)") " STATE NR. ", istate, evals(istate)*evolt, " eV"
1055 : END IF
1056 : nmax = 0
1057 546 : DO ispin = 1, nspins
1058 276 : CALL cp_fm_get_info(evects(ispin, istate), matrix_struct=fm_struct, ncol_global=nmo)
1059 546 : nmax = MAX(nmax, nmo)
1060 : END DO
1061 1080 : ALLOCATE (eigenvalues(nmax, nspins))
1062 270 : eigenvalues = 0.0_dp
1063 : ! SET 1: Hole states
1064 : ! SET 2: Particle states
1065 270 : nto_name(1) = 'Hole_states'
1066 270 : nto_name(2) = 'Particle_states'
1067 810 : ALLOCATE (nto_set(2))
1068 810 : DO i = 1, 2
1069 540 : CALL allocate_mo_set(nto_set(i), nao, ntomax, 0, 0.0_dp, 1.0_dp, 0.0_dp)
1070 540 : CALL cp_fm_get_info(evects(1, istate), matrix_struct=fm_struct)
1071 : CALL cp_fm_struct_create(fmstruct=fm_mo_struct, template_fmstruct=fm_struct, &
1072 540 : ncol_global=ntomax)
1073 540 : CALL cp_fm_create(tmat, fm_mo_struct)
1074 540 : CALL init_mo_set(nto_set(i), fm_ref=tmat, name=nto_name(i))
1075 540 : CALL cp_fm_release(tmat)
1076 1350 : CALL cp_fm_struct_release(fm_mo_struct)
1077 : END DO
1078 : !
1079 1086 : ALLOCATE (teig(nspins))
1080 : ! hole states
1081 : ! Diagonalize X(T)*S*X
1082 546 : DO ispin = 1, nspins
1083 : ASSOCIATE (ev => evects(ispin, istate))
1084 276 : CALL cp_fm_get_info(ev, matrix_struct=fm_struct, ncol_global=nmo)
1085 276 : CALL cp_fm_create(Sev, fm_struct)
1086 : CALL cp_fm_struct_create(fmstruct=fm_mo_struct, template_fmstruct=fm_struct, &
1087 276 : nrow_global=nmo, ncol_global=nmo)
1088 276 : CALL cp_fm_create(tmat, fm_mo_struct)
1089 276 : CALL cp_fm_create(teig(ispin), fm_mo_struct)
1090 276 : CALL cp_dbcsr_sm_fm_multiply(matrix_s, ev, Sev, ncol=nmo, alpha=1.0_dp, beta=0.0_dp)
1091 276 : CALL parallel_gemm('T', 'N', nmo, nmo, nao, 1.0_dp, ev, Sev, 0.0_dp, tmat)
1092 : END ASSOCIATE
1093 :
1094 276 : CALL choose_eigv_solver(tmat, teig(ispin), eigenvalues(1:nmo, ispin))
1095 :
1096 276 : CALL cp_fm_struct_release(fm_mo_struct)
1097 276 : CALL cp_fm_release(tmat)
1098 1098 : CALL cp_fm_release(Sev)
1099 : END DO
1100 : ! find major determinants i->a
1101 270 : ia_index = 0
1102 270 : sume = 0.0_dp
1103 270 : nnto = 0
1104 326 : DO i = 1, ntomax
1105 3452 : iv = MAXLOC(eigenvalues)
1106 326 : ia_eval(i) = eigenvalues(iv(1), iv(2))
1107 978 : ia_index(1:2, i) = iv(1:2)
1108 326 : sume = sume + ia_eval(i)
1109 326 : eigenvalues(iv(1), iv(2)) = 0.0_dp
1110 326 : nnto = nnto + 1
1111 326 : IF (sume > threshold) EXIT
1112 : END DO
1113 : ! store hole states
1114 270 : CALL set_mo_set(nto_set(1), nmo=nnto)
1115 596 : DO i = 1, nnto
1116 326 : ia = ia_index(1, i)
1117 326 : ispin = ia_index(2, i)
1118 326 : CALL cp_fm_get_info(gs_mos(ispin)%mos_occ, ncol_global=nmo)
1119 326 : CALL cp_fm_get_info(teig(ispin), matrix_struct=fm_struct)
1120 : CALL cp_fm_struct_create(fmstruct=fm_mo_struct, template_fmstruct=fm_struct, &
1121 326 : nrow_global=nmo, ncol_global=1)
1122 326 : CALL cp_fm_create(tmat, fm_mo_struct)
1123 326 : CALL cp_fm_struct_release(fm_mo_struct)
1124 326 : CALL cp_fm_get_info(gs_mos(1)%mos_occ, matrix_struct=fm_struct)
1125 : CALL cp_fm_struct_create(fmstruct=fm_mo_struct, template_fmstruct=fm_struct, &
1126 326 : ncol_global=1)
1127 326 : CALL cp_fm_create(wvec, fm_mo_struct)
1128 326 : CALL cp_fm_struct_release(fm_mo_struct)
1129 326 : CALL cp_fm_to_fm(teig(ispin), tmat, 1, ia, 1)
1130 : CALL parallel_gemm('N', 'N', nao, 1, nmo, 1.0_dp, gs_mos(ispin)%mos_occ, &
1131 326 : tmat, 0.0_dp, wvec)
1132 326 : CALL cp_fm_to_fm(wvec, nto_set(1)%mo_coeff, 1, 1, i)
1133 326 : CALL cp_fm_release(wvec)
1134 1574 : CALL cp_fm_release(tmat)
1135 : END DO
1136 : ! particle states
1137 : ! Solve generalized eigenvalue equation: (S*X)*(S*X)(T)*v = lambda*S*v
1138 270 : CALL set_mo_set(nto_set(2), nmo=nnto)
1139 546 : DO ispin = 1, nspins
1140 : ASSOCIATE (ev => evects(ispin, istate))
1141 276 : CALL cp_fm_get_info(ev, matrix_struct=fm_struct, nrow_global=nao, ncol_global=nmo)
1142 828 : ALLOCATE (eigvals(nao))
1143 276 : eigvals = 0.0_dp
1144 276 : CALL cp_fm_create(Sev, fm_struct)
1145 552 : CALL cp_dbcsr_sm_fm_multiply(matrix_s, ev, Sev, ncol=nmo, alpha=1.0_dp, beta=0.0_dp)
1146 : END ASSOCIATE
1147 : CALL cp_fm_struct_create(fmstruct=fm_mo_struct, template_fmstruct=fm_struct, &
1148 276 : nrow_global=nao, ncol_global=nao)
1149 276 : CALL cp_fm_create(tmat, fm_mo_struct)
1150 276 : CALL cp_fm_create(smat, fm_mo_struct)
1151 276 : CALL cp_fm_create(wmat, fm_mo_struct)
1152 276 : CALL cp_fm_create(work, fm_mo_struct)
1153 276 : CALL cp_fm_struct_release(fm_mo_struct)
1154 276 : CALL copy_dbcsr_to_fm(matrix_s, smat)
1155 276 : CALL parallel_gemm('N', 'T', nao, nao, nmo, 1.0_dp, Sev, Sev, 0.0_dp, tmat)
1156 276 : CALL cp_fm_geeig(tmat, smat, wmat, eigvals, work)
1157 610 : DO i = 1, nnto
1158 610 : IF (ispin == ia_index(2, i)) THEN
1159 326 : icg = 0
1160 7984 : DO j = 1, nao
1161 7984 : IF (ABS(eigvals(j) - ia_eval(i)) < 1.E-6_dp) THEN
1162 326 : icg = j
1163 326 : EXIT
1164 : END IF
1165 : END DO
1166 326 : IF (icg == 0) THEN
1167 : CALL cp_warn(__LOCATION__, &
1168 0 : "Could not locate particle state associated with hole state.")
1169 : ELSE
1170 326 : CALL cp_fm_to_fm(wmat, nto_set(2)%mo_coeff, 1, icg, i)
1171 : END IF
1172 : END IF
1173 : END DO
1174 276 : DEALLOCATE (eigvals)
1175 276 : CALL cp_fm_release(Sev)
1176 276 : CALL cp_fm_release(tmat)
1177 276 : CALL cp_fm_release(smat)
1178 276 : CALL cp_fm_release(wmat)
1179 822 : CALL cp_fm_release(work)
1180 : END DO
1181 : ! print
1182 270 : IF (iounit > 0) THEN
1183 135 : sume = 0.0_dp
1184 298 : DO i = 1, nnto
1185 163 : sume = sume + ia_eval(i)
1186 : WRITE (iounit, "(T6,A,i2,T30,A,i1,T42,A,F8.5,T63,A,F8.5)") &
1187 163 : "Particle-Hole state:", i, " Spin:", ia_index(2, i), &
1188 461 : "Eigenvalue:", ia_eval(i), " Sum Eigv:", sume
1189 : END DO
1190 : END IF
1191 : ! Cube and Molden files
1192 270 : nto_section => section_vals_get_subs_vals(print_section, "NTO_ANALYSIS")
1193 270 : CALL section_vals_val_get(nto_section, "CUBE_FILES", l_val=cube_file)
1194 270 : CALL section_vals_val_get(nto_section, "STRIDE", i_vals=stride)
1195 270 : CALL section_vals_val_get(nto_section, "APPEND", l_val=append_cube)
1196 270 : IF (cube_file) THEN
1197 8 : CALL print_nto_cubes(qs_env, nto_set, istate, stride, append_cube, nto_section)
1198 : END IF
1199 270 : CALL get_qs_env(qs_env, qs_kind_set=qs_kind_set, particle_set=particle_set, cell=cell)
1200 270 : molden_section => section_vals_get_subs_vals(print_section, "MOS_MOLDEN")
1201 270 : CALL write_mos_molden(nto_set, qs_kind_set, particle_set, molden_section, cell=cell, qs_env=qs_env)
1202 : !
1203 270 : DEALLOCATE (eigenvalues)
1204 270 : CALL cp_fm_release(teig)
1205 : !
1206 810 : DO i = 1, 2
1207 810 : CALL deallocate_mo_set(nto_set(i))
1208 : END DO
1209 1034 : DEALLOCATE (nto_set)
1210 : END DO
1211 :
1212 224 : IF (iounit > 0) THEN
1213 : WRITE (iounit, "(1X,A)") &
1214 112 : "-------------------------------------------------------------------------------"
1215 : END IF
1216 :
1217 : END IF
1218 :
1219 1426 : CALL timestop(handle)
1220 :
1221 2852 : END SUBROUTINE tddfpt_print_nto_analysis
1222 :
1223 : ! **************************************************************************************************
1224 : !> \brief Print exciton descriptors, cf. Mewes et al., JCTC 14, 710-725 (2018)
1225 : !> \param log_unit output unit
1226 : !> \param evects TDDFPT trial vectors (SIZE(evects,1) -- number of spins;
1227 : !> SIZE(evects,2) -- number of excited states to print)
1228 : !> \param gs_mos molecular orbitals optimised for the ground state
1229 : !> \param matrix_s overlap matrix
1230 : !> \param do_directional_exciton_descriptors flag for computing descriptors for each (cartesian) direction
1231 : !> \param qs_env Information on particles/geometry
1232 : !> \par History
1233 : !> * 12.2024 created as 'tddfpt_print_exciton_descriptors' [Maximilian Graml]
1234 : ! **************************************************************************************************
1235 2 : SUBROUTINE tddfpt_print_exciton_descriptors(log_unit, evects, gs_mos, matrix_s, &
1236 : do_directional_exciton_descriptors, qs_env)
1237 : INTEGER, INTENT(in) :: log_unit
1238 : TYPE(cp_fm_type), DIMENSION(:, :), INTENT(in) :: evects
1239 : TYPE(tddfpt_ground_state_mos), DIMENSION(:), &
1240 : INTENT(in) :: gs_mos
1241 : TYPE(dbcsr_type), POINTER :: matrix_s
1242 : LOGICAL, INTENT(IN) :: do_directional_exciton_descriptors
1243 : TYPE(qs_environment_type), INTENT(IN), POINTER :: qs_env
1244 :
1245 : CHARACTER(LEN=*), PARAMETER :: routineN = 'tddfpt_print_exciton_descriptors'
1246 :
1247 : CHARACTER(LEN=4) :: prefix_output
1248 : INTEGER :: handle, ispin, istate, n_moments_quad, &
1249 : nactive, nao, nspins, nstates
1250 : INTEGER, DIMENSION(maxspins) :: nmo_occ, nmo_virt
1251 : LOGICAL :: print_checkvalue
1252 2 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:) :: ref_point_multipole
1253 : TYPE(cp_blacs_env_type), POINTER :: blacs_env
1254 : TYPE(cp_fm_struct_type), POINTER :: fm_struct_mo_coeff, &
1255 : fm_struct_S_mos_virt, fm_struct_X_ia_n
1256 2 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:) :: eigvec_X_ia_n, fm_multipole_ab, &
1257 2 : fm_multipole_ai, fm_multipole_ij, &
1258 2 : S_mos_virt
1259 2 : TYPE(cp_fm_type), DIMENSION(:), POINTER :: mo_coeff
1260 : TYPE(exciton_descr_type), ALLOCATABLE, &
1261 2 : DIMENSION(:) :: exc_descr
1262 :
1263 2 : CALL timeset(routineN, handle)
1264 :
1265 2 : nspins = SIZE(evects, 1)
1266 2 : nstates = SIZE(evects, 2)
1267 :
1268 2 : CPASSERT(nspins == 1) ! Other spins are not yet implemented for exciton descriptors
1269 :
1270 2 : CALL cp_fm_get_info(gs_mos(1)%mos_occ, context=blacs_env)
1271 2 : CALL dbcsr_get_info(matrix_s, nfullrows_total=nao)
1272 :
1273 4 : DO ispin = 1, nspins
1274 2 : nmo_occ(ispin) = SIZE(gs_mos(ispin)%evals_occ)
1275 4 : nmo_virt(ispin) = SIZE(gs_mos(ispin)%evals_virt)
1276 : END DO
1277 :
1278 : ! Prepare fm with all MO coefficents, i.e. nao x nao
1279 8 : ALLOCATE (mo_coeff(nspins))
1280 : CALL cp_fm_struct_create(fm_struct_mo_coeff, nrow_global=nao, ncol_global=nao, &
1281 2 : context=blacs_env)
1282 4 : DO ispin = 1, nspins
1283 2 : CALL cp_fm_create(mo_coeff(ispin), fm_struct_mo_coeff)
1284 : CALL cp_fm_to_fm_submat_general(gs_mos(ispin)%mos_occ, &
1285 : mo_coeff(ispin), &
1286 : nao, &
1287 : nmo_occ(ispin), &
1288 : 1, &
1289 : 1, &
1290 : 1, &
1291 : 1, &
1292 2 : blacs_env)
1293 : CALL cp_fm_to_fm_submat_general(gs_mos(ispin)%mos_virt, &
1294 : mo_coeff(ispin), &
1295 : nao, &
1296 : nmo_virt(ispin), &
1297 : 1, &
1298 : 1, &
1299 : 1, &
1300 : nmo_occ(ispin) + 1, &
1301 4 : blacs_env)
1302 : END DO
1303 2 : CALL cp_fm_struct_release(fm_struct_mo_coeff)
1304 :
1305 : ! Compute multipole moments
1306 : ! fm_multipole_XY have structure inherited by libint, i.e. x, y, z, xx, xy, xz, yy, yz, zz
1307 2 : n_moments_quad = 9
1308 2 : ALLOCATE (ref_point_multipole(3))
1309 20 : ALLOCATE (fm_multipole_ij(n_moments_quad))
1310 20 : ALLOCATE (fm_multipole_ab(n_moments_quad))
1311 20 : ALLOCATE (fm_multipole_ai(n_moments_quad))
1312 :
1313 : CALL get_multipoles_mo(fm_multipole_ai, fm_multipole_ij, fm_multipole_ab, &
1314 : qs_env, mo_coeff, ref_point_multipole, 2, &
1315 2 : nmo_occ(1), nmo_virt(1), blacs_env)
1316 :
1317 2 : CALL cp_fm_release(mo_coeff)
1318 :
1319 : ! Compute eigenvector X of the Casida equation from trial vectors
1320 10 : ALLOCATE (S_mos_virt(nspins), eigvec_X_ia_n(nspins))
1321 4 : DO ispin = 1, nspins
1322 2 : CALL cp_fm_get_info(gs_mos(ispin)%mos_virt, matrix_struct=fm_struct_S_mos_virt)
1323 2 : CALL cp_fm_create(S_mos_virt(ispin), fm_struct_S_mos_virt)
1324 2 : NULLIFY (fm_struct_S_mos_virt)
1325 : CALL cp_dbcsr_sm_fm_multiply(matrix_s, &
1326 : gs_mos(ispin)%mos_virt, &
1327 : S_mos_virt(ispin), &
1328 2 : ncol=nmo_virt(ispin), alpha=1.0_dp, beta=0.0_dp)
1329 :
1330 : CALL cp_fm_struct_create(fm_struct_X_ia_n, nrow_global=nmo_occ(ispin), ncol_global=nmo_virt(ispin), &
1331 2 : context=blacs_env)
1332 2 : CALL cp_fm_create(eigvec_X_ia_n(ispin), fm_struct_X_ia_n)
1333 4 : CALL cp_fm_struct_release(fm_struct_X_ia_n)
1334 : END DO
1335 172 : ALLOCATE (exc_descr(nstates))
1336 12 : DO istate = 1, nstates
1337 22 : DO ispin = 1, nspins
1338 10 : CALL cp_fm_set_all(eigvec_X_ia_n(ispin), 0.0_dp)
1339 : ! compute eigenvectors X of the TDA equation
1340 : ! Reshuffle multiplication from
1341 : ! X_ai = S_ma ^T * C_mi
1342 : ! to
1343 : ! X_ia = C_mi ^T * S_ma
1344 : ! for compatibility with the structure needed for get_exciton_descriptors of bse_properties.F
1345 10 : CALL cp_fm_get_info(evects(ispin, istate), ncol_global=nactive)
1346 10 : IF (nactive /= nmo_occ(ispin)) THEN
1347 : CALL cp_abort(__LOCATION__, &
1348 0 : "Reduced active space excitations not implemented")
1349 : END IF
1350 : CALL parallel_gemm('T', 'N', nmo_occ(ispin), nmo_virt(ispin), nao, 1.0_dp, &
1351 10 : evects(ispin, istate), S_mos_virt(ispin), 0.0_dp, eigvec_X_ia_n(ispin))
1352 :
1353 : CALL get_exciton_descriptors(exc_descr, eigvec_X_ia_n(ispin), &
1354 : fm_multipole_ij, fm_multipole_ab, &
1355 : fm_multipole_ai, &
1356 30 : istate, nmo_occ(ispin), nmo_virt(ispin))
1357 : END DO
1358 : END DO
1359 2 : CALL cp_fm_release(eigvec_X_ia_n)
1360 2 : CALL cp_fm_release(S_mos_virt)
1361 2 : CALL cp_fm_release(fm_multipole_ai)
1362 2 : CALL cp_fm_release(fm_multipole_ij)
1363 2 : CALL cp_fm_release(fm_multipole_ab)
1364 :
1365 : ! Actual printing
1366 2 : print_checkvalue = .TRUE.
1367 2 : prefix_output = ' '
1368 : CALL print_exciton_descriptors(exc_descr, ref_point_multipole, log_unit, &
1369 : nstates, print_checkvalue, do_directional_exciton_descriptors, &
1370 2 : prefix_output, qs_env)
1371 :
1372 2 : DEALLOCATE (ref_point_multipole)
1373 2 : DEALLOCATE (exc_descr)
1374 :
1375 2 : CALL timestop(handle)
1376 :
1377 6 : END SUBROUTINE tddfpt_print_exciton_descriptors
1378 :
1379 : ! **************************************************************************************************
1380 : !> \brief ...
1381 : !> \param vin ...
1382 : !> \param vout ...
1383 : !> \param mos_occ ...
1384 : !> \param matrix_s ...
1385 : ! **************************************************************************************************
1386 0 : SUBROUTINE project_vector(vin, vout, mos_occ, matrix_s)
1387 : TYPE(dbcsr_type) :: vin, vout
1388 : TYPE(cp_fm_type), INTENT(IN) :: mos_occ
1389 : TYPE(dbcsr_type), POINTER :: matrix_s
1390 :
1391 : CHARACTER(LEN=*), PARAMETER :: routineN = 'project_vector'
1392 :
1393 : INTEGER :: handle, nao, nmo
1394 : REAL(KIND=dp) :: norm(1)
1395 : TYPE(cp_fm_struct_type), POINTER :: fm_struct, fm_vec_struct
1396 : TYPE(cp_fm_type) :: csvec, svec, vec
1397 :
1398 0 : CALL timeset(routineN, handle)
1399 :
1400 0 : CALL cp_fm_get_info(mos_occ, matrix_struct=fm_struct, nrow_global=nao, ncol_global=nmo)
1401 : CALL cp_fm_struct_create(fmstruct=fm_vec_struct, template_fmstruct=fm_struct, &
1402 0 : nrow_global=nao, ncol_global=1)
1403 0 : CALL cp_fm_create(vec, fm_vec_struct)
1404 0 : CALL cp_fm_create(svec, fm_vec_struct)
1405 0 : CALL cp_fm_struct_release(fm_vec_struct)
1406 : CALL cp_fm_struct_create(fmstruct=fm_vec_struct, template_fmstruct=fm_struct, &
1407 0 : nrow_global=nmo, ncol_global=1)
1408 0 : CALL cp_fm_create(csvec, fm_vec_struct)
1409 0 : CALL cp_fm_struct_release(fm_vec_struct)
1410 :
1411 0 : CALL copy_dbcsr_to_fm(vin, vec)
1412 0 : CALL cp_dbcsr_sm_fm_multiply(matrix_s, vec, svec, ncol=1, alpha=1.0_dp, beta=0.0_dp)
1413 0 : CALL parallel_gemm('T', 'N', nmo, 1, nao, 1.0_dp, mos_occ, svec, 0.0_dp, csvec)
1414 0 : CALL parallel_gemm('N', 'N', nao, 1, nmo, -1.0_dp, mos_occ, csvec, 1.0_dp, vec)
1415 0 : CALL cp_fm_vectorsnorm(vec, norm)
1416 0 : CPASSERT(norm(1) > 1.e-14_dp)
1417 0 : norm(1) = SQRT(1._dp/norm(1))
1418 0 : CALL cp_fm_scale(norm(1), vec)
1419 0 : CALL copy_fm_to_dbcsr(vec, vout, keep_sparsity=.FALSE.)
1420 :
1421 0 : CALL cp_fm_release(csvec)
1422 0 : CALL cp_fm_release(svec)
1423 0 : CALL cp_fm_release(vec)
1424 :
1425 0 : CALL timestop(handle)
1426 :
1427 0 : END SUBROUTINE project_vector
1428 :
1429 : ! **************************************************************************************************
1430 : !> \brief ...
1431 : !> \param va ...
1432 : !> \param vb ...
1433 : !> \param res ...
1434 : ! **************************************************************************************************
1435 0 : SUBROUTINE vec_product(va, vb, res)
1436 : TYPE(dbcsr_type) :: va, vb
1437 : REAL(KIND=dp), INTENT(OUT) :: res
1438 :
1439 : CHARACTER(LEN=*), PARAMETER :: routineN = 'vec_product'
1440 :
1441 : INTEGER :: handle, icol, irow
1442 : LOGICAL :: found
1443 0 : REAL(KIND=dp), DIMENSION(:, :), POINTER :: vba, vbb
1444 : TYPE(dbcsr_iterator_type) :: iter
1445 : TYPE(mp_comm_type) :: group
1446 :
1447 0 : CALL timeset(routineN, handle)
1448 :
1449 0 : res = 0.0_dp
1450 :
1451 0 : CALL dbcsr_get_info(va, group=group)
1452 0 : CALL dbcsr_iterator_start(iter, va)
1453 0 : DO WHILE (dbcsr_iterator_blocks_left(iter))
1454 0 : CALL dbcsr_iterator_next_block(iter, irow, icol, vba)
1455 0 : CALL dbcsr_get_block_p(vb, row=irow, col=icol, block=vbb, found=found)
1456 0 : res = res + SUM(vba*vbb)
1457 0 : CPASSERT(found)
1458 : END DO
1459 0 : CALL dbcsr_iterator_stop(iter)
1460 0 : CALL group%sum(res)
1461 :
1462 0 : CALL timestop(handle)
1463 :
1464 0 : END SUBROUTINE vec_product
1465 :
1466 : ! **************************************************************************************************
1467 : !> \brief ...
1468 : !> \param qs_env ...
1469 : !> \param mos ...
1470 : !> \param istate ...
1471 : !> \param stride ...
1472 : !> \param append_cube ...
1473 : !> \param print_section ...
1474 : ! **************************************************************************************************
1475 8 : SUBROUTINE print_nto_cubes(qs_env, mos, istate, stride, append_cube, print_section)
1476 :
1477 : TYPE(qs_environment_type), POINTER :: qs_env
1478 : TYPE(mo_set_type), DIMENSION(:), INTENT(IN) :: mos
1479 : INTEGER, INTENT(IN) :: istate
1480 : INTEGER, DIMENSION(:), POINTER :: stride
1481 : LOGICAL, INTENT(IN) :: append_cube
1482 : TYPE(section_vals_type), POINTER :: print_section
1483 :
1484 : CHARACTER(LEN=default_path_length) :: filename, my_pos_cube, title
1485 : INTEGER :: i, iset, nmo, unit_nr
1486 : LOGICAL :: mpi_io
1487 8 : TYPE(atomic_kind_type), DIMENSION(:), POINTER :: atomic_kind_set
1488 : TYPE(cell_type), POINTER :: cell
1489 : TYPE(cp_fm_type), POINTER :: mo_coeff
1490 : TYPE(cp_logger_type), POINTER :: logger
1491 : TYPE(dft_control_type), POINTER :: dft_control
1492 : TYPE(particle_list_type), POINTER :: particles
1493 8 : TYPE(particle_type), DIMENSION(:), POINTER :: particle_set
1494 : TYPE(pw_c1d_gs_type) :: wf_g
1495 : TYPE(pw_env_type), POINTER :: pw_env
1496 8 : TYPE(pw_pool_p_type), DIMENSION(:), POINTER :: pw_pools
1497 : TYPE(pw_pool_type), POINTER :: auxbas_pw_pool
1498 : TYPE(pw_r3d_rs_type) :: wf_r
1499 8 : TYPE(qs_kind_type), DIMENSION(:), POINTER :: qs_kind_set
1500 : TYPE(qs_subsys_type), POINTER :: subsys
1501 :
1502 16 : logger => cp_get_default_logger()
1503 :
1504 8 : CALL get_qs_env(qs_env=qs_env, dft_control=dft_control, pw_env=pw_env)
1505 8 : CALL pw_env_get(pw_env, auxbas_pw_pool=auxbas_pw_pool, pw_pools=pw_pools)
1506 8 : CALL auxbas_pw_pool%create_pw(wf_r)
1507 8 : CALL auxbas_pw_pool%create_pw(wf_g)
1508 :
1509 8 : CALL get_qs_env(qs_env, subsys=subsys)
1510 8 : CALL qs_subsys_get(subsys, particles=particles)
1511 :
1512 8 : my_pos_cube = "REWIND"
1513 8 : IF (append_cube) THEN
1514 0 : my_pos_cube = "APPEND"
1515 : END IF
1516 :
1517 : CALL get_qs_env(qs_env=qs_env, &
1518 : atomic_kind_set=atomic_kind_set, &
1519 : qs_kind_set=qs_kind_set, &
1520 : cell=cell, &
1521 8 : particle_set=particle_set)
1522 :
1523 24 : DO iset = 1, 2
1524 16 : CALL get_mo_set(mo_set=mos(iset), mo_coeff=mo_coeff, nmo=nmo)
1525 44 : DO i = 1, nmo
1526 : CALL calculate_wavefunction(mo_coeff, i, wf_r, wf_g, atomic_kind_set, qs_kind_set, &
1527 20 : cell, dft_control, particle_set, pw_env)
1528 20 : IF (iset == 1) THEN
1529 10 : WRITE (filename, '(a4,I3.3,I2.2,a11)') "NTO_STATE", istate, i, "_Hole_State"
1530 10 : ELSE IF (iset == 2) THEN
1531 10 : WRITE (filename, '(a4,I3.3,I2.2,a15)') "NTO_STATE", istate, i, "_Particle_State"
1532 : END IF
1533 20 : mpi_io = .TRUE.
1534 : unit_nr = cp_print_key_unit_nr(logger, print_section, '', extension=".cube", &
1535 : middle_name=TRIM(filename), file_position=my_pos_cube, &
1536 20 : log_filename=.FALSE., ignore_should_output=.TRUE., mpi_io=mpi_io)
1537 20 : IF (iset == 1) THEN
1538 10 : WRITE (title, *) "Natural Transition Orbital Hole State", i
1539 10 : ELSE IF (iset == 2) THEN
1540 10 : WRITE (title, *) "Natural Transition Orbital Particle State", i
1541 : END IF
1542 20 : CALL cp_pw_to_cube(wf_r, unit_nr, title, particles=particles, stride=stride, mpi_io=mpi_io)
1543 : CALL cp_print_key_finished_output(unit_nr, logger, print_section, '', &
1544 36 : ignore_should_output=.TRUE., mpi_io=mpi_io)
1545 : END DO
1546 : END DO
1547 :
1548 8 : CALL auxbas_pw_pool%give_back_pw(wf_g)
1549 8 : CALL auxbas_pw_pool%give_back_pw(wf_r)
1550 :
1551 8 : END SUBROUTINE print_nto_cubes
1552 :
1553 : END MODULE qs_tddfpt2_properties
|
