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 : ! **************************************************************************************************
9 : !> \brief Routines to calculate RI-RPA energy
10 : !> \par History
11 : !> 06.2012 created [Mauro Del Ben]
12 : !> 04.2015 GW routines added [Jan Wilhelm]
13 : !> 10.2015 Cubic-scaling RPA routines added [Jan Wilhelm]
14 : !> 10.2018 Cubic-scaling SOS-MP2 added [Frederick Stein]
15 : !> 03.2019 Refactoring [Frederick Stein]
16 : ! **************************************************************************************************
17 : MODULE rpa_main
18 : USE bibliography, ONLY: &
19 : Bates2013, DelBen2013, DelBen2015, Freeman1977, Gruneis2009, Ren2011, Ren2013, &
20 : Wilhelm2016a, Wilhelm2016b, Wilhelm2017, Wilhelm2018, cite_reference
21 : USE bse_main, ONLY: start_bse_calculation
22 : USE cp_blacs_env, ONLY: cp_blacs_env_create,&
23 : cp_blacs_env_release,&
24 : cp_blacs_env_type
25 : USE cp_cfm_types, ONLY: cp_cfm_type
26 : USE cp_dbcsr_api, ONLY: dbcsr_add,&
27 : dbcsr_clear,&
28 : dbcsr_get_info,&
29 : dbcsr_p_type,&
30 : dbcsr_type
31 : USE cp_fm_basic_linalg, ONLY: cp_fm_scale_and_add
32 : USE cp_fm_struct, ONLY: cp_fm_struct_create,&
33 : cp_fm_struct_release,&
34 : cp_fm_struct_type
35 : USE cp_fm_types, ONLY: cp_fm_create,&
36 : cp_fm_release,&
37 : cp_fm_set_all,&
38 : cp_fm_to_fm,&
39 : cp_fm_type
40 : USE dbt_api, ONLY: dbt_type
41 : USE dgemm_counter_types, ONLY: dgemm_counter_init,&
42 : dgemm_counter_type,&
43 : dgemm_counter_write
44 : USE group_dist_types, ONLY: create_group_dist,&
45 : get_group_dist,&
46 : group_dist_d1_type,&
47 : maxsize,&
48 : release_group_dist
49 : USE hfx_types, ONLY: block_ind_type,&
50 : hfx_compression_type
51 : USE input_constants, ONLY: rpa_exchange_axk,&
52 : rpa_exchange_none,&
53 : rpa_exchange_sosex,&
54 : sigma_none,&
55 : wfc_mm_style_gemm
56 : USE kinds, ONLY: dp,&
57 : int_8
58 : USE kpoint_types, ONLY: kpoint_type
59 : USE machine, ONLY: m_flush,&
60 : m_memory
61 : USE mathconstants, ONLY: pi,&
62 : z_zero
63 : USE message_passing, ONLY: mp_comm_type,&
64 : mp_para_env_release,&
65 : mp_para_env_type
66 : USE minimax_exp, ONLY: check_exp_minimax_range
67 : USE mp2_grids, ONLY: get_clenshaw_grid,&
68 : get_minimax_grid
69 : USE mp2_laplace, ONLY: SOS_MP2_postprocessing
70 : USE mp2_ri_grad_util, ONLY: array2fm
71 : USE mp2_types, ONLY: mp2_type,&
72 : three_dim_real_array,&
73 : two_dim_int_array,&
74 : two_dim_real_array
75 : USE qs_environment_types, ONLY: get_qs_env,&
76 : qs_environment_type
77 : USE rpa_exchange, ONLY: rpa_exchange_needed_mem,&
78 : rpa_exchange_work_type
79 : USE rpa_grad, ONLY: rpa_grad_copy_Q,&
80 : rpa_grad_create,&
81 : rpa_grad_finalize,&
82 : rpa_grad_matrix_operations,&
83 : rpa_grad_needed_mem,&
84 : rpa_grad_type
85 : USE rpa_gw, ONLY: allocate_matrices_gw,&
86 : allocate_matrices_gw_im_time,&
87 : compute_GW_self_energy,&
88 : compute_QP_energies,&
89 : compute_W_cubic_GW,&
90 : deallocate_matrices_gw,&
91 : deallocate_matrices_gw_im_time,&
92 : get_fermi_level_offset
93 : USE rpa_gw_ic, ONLY: calculate_ic_correction
94 : USE rpa_gw_kpoints_util, ONLY: get_bandstruc_and_k_dependent_MOs,&
95 : invert_eps_compute_W_and_Erpa_kp
96 : USE rpa_im_time, ONLY: compute_mat_P_omega,&
97 : zero_mat_P_omega
98 : USE rpa_im_time_force_methods, ONLY: calc_laplace_loop_forces,&
99 : calc_post_loop_forces,&
100 : calc_rpa_loop_forces,&
101 : init_im_time_forces,&
102 : keep_initial_quad
103 : USE rpa_im_time_force_types, ONLY: im_time_force_release,&
104 : im_time_force_type
105 : USE rpa_sigma_functional, ONLY: finalize_rpa_sigma,&
106 : rpa_sigma_create,&
107 : rpa_sigma_matrix_spectral,&
108 : rpa_sigma_type
109 : USE rpa_util, ONLY: Q_trace_and_add_unit_matrix,&
110 : alloc_im_time,&
111 : calc_mat_Q,&
112 : compute_Erpa_by_freq_int,&
113 : contract_P_omega_with_mat_L,&
114 : dealloc_im_time,&
115 : remove_scaling_factor_rpa
116 : USE util, ONLY: get_limit
117 : #include "./base/base_uses.f90"
118 :
119 : IMPLICIT NONE
120 :
121 : PRIVATE
122 :
123 : CHARACTER(len=*), PARAMETER, PRIVATE :: moduleN = 'rpa_main'
124 :
125 : PUBLIC :: rpa_ri_compute_en
126 :
127 : CONTAINS
128 :
129 : ! **************************************************************************************************
130 : !> \brief ...
131 : !> \param qs_env ...
132 : !> \param Erpa ...
133 : !> \param mp2_env ...
134 : !> \param BIb_C ...
135 : !> \param BIb_C_gw ...
136 : !> \param BIb_C_bse_ij ...
137 : !> \param BIb_C_bse_ab ...
138 : !> \param para_env ...
139 : !> \param para_env_sub ...
140 : !> \param color_sub ...
141 : !> \param gd_array ...
142 : !> \param gd_B_virtual ...
143 : !> \param gd_B_all ...
144 : !> \param gd_B_occ_bse ...
145 : !> \param gd_B_virt_bse ...
146 : !> \param mo_coeff ...
147 : !> \param fm_matrix_PQ ...
148 : !> \param fm_matrix_L_kpoints ...
149 : !> \param fm_matrix_Minv_L_kpoints ...
150 : !> \param fm_matrix_Minv ...
151 : !> \param fm_matrix_Minv_Vtrunc_Minv ...
152 : !> \param kpoints ...
153 : !> \param Eigenval ...
154 : !> \param nmo ...
155 : !> \param homo ...
156 : !> \param dimen_RI ...
157 : !> \param dimen_RI_red ...
158 : !> \param gw_corr_lev_occ ...
159 : !> \param gw_corr_lev_virt ...
160 : !> \param bse_lev_virt ...
161 : !> \param unit_nr ...
162 : !> \param do_ri_sos_laplace_mp2 ...
163 : !> \param my_do_gw ...
164 : !> \param do_im_time ...
165 : !> \param do_bse ...
166 : !> \param matrix_s ...
167 : !> \param mat_munu ...
168 : !> \param mat_P_global ...
169 : !> \param t_3c_M ...
170 : !> \param t_3c_O ...
171 : !> \param t_3c_O_compressed ...
172 : !> \param t_3c_O_ind ...
173 : !> \param starts_array_mc ...
174 : !> \param ends_array_mc ...
175 : !> \param starts_array_mc_block ...
176 : !> \param ends_array_mc_block ...
177 : !> \param calc_forces ...
178 : ! **************************************************************************************************
179 1480 : SUBROUTINE rpa_ri_compute_en(qs_env, Erpa, mp2_env, BIb_C, BIb_C_gw, BIb_C_bse_ij, BIb_C_bse_ab, &
180 : para_env, para_env_sub, color_sub, &
181 296 : gd_array, gd_B_virtual, gd_B_all, gd_B_occ_bse, gd_B_virt_bse, &
182 296 : mo_coeff, fm_matrix_PQ, fm_matrix_L_kpoints, fm_matrix_Minv_L_kpoints, &
183 : fm_matrix_Minv, fm_matrix_Minv_Vtrunc_Minv, kpoints, &
184 592 : Eigenval, nmo, homo, dimen_RI, dimen_RI_red, gw_corr_lev_occ, gw_corr_lev_virt, &
185 : bse_lev_virt, &
186 : unit_nr, do_ri_sos_laplace_mp2, my_do_gw, do_im_time, do_bse, matrix_s, &
187 : mat_munu, mat_P_global, t_3c_M, t_3c_O, t_3c_O_compressed, t_3c_O_ind, &
188 : starts_array_mc, ends_array_mc, &
189 : starts_array_mc_block, ends_array_mc_block, calc_forces)
190 :
191 : TYPE(qs_environment_type), POINTER :: qs_env
192 : REAL(KIND=dp), INTENT(OUT) :: Erpa
193 : TYPE(mp2_type), INTENT(INOUT) :: mp2_env
194 : TYPE(three_dim_real_array), DIMENSION(:), &
195 : INTENT(INOUT) :: BIb_C, BIb_C_gw
196 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :, :), &
197 : INTENT(INOUT) :: BIb_C_bse_ij, BIb_C_bse_ab
198 : TYPE(mp_para_env_type), POINTER :: para_env, para_env_sub
199 : INTEGER, INTENT(INOUT) :: color_sub
200 : TYPE(group_dist_d1_type), INTENT(INOUT) :: gd_array
201 : TYPE(group_dist_d1_type), DIMENSION(:), &
202 : INTENT(INOUT) :: gd_B_virtual
203 : TYPE(group_dist_d1_type), INTENT(INOUT) :: gd_B_all, gd_B_occ_bse, gd_B_virt_bse
204 : TYPE(cp_fm_type), DIMENSION(:), INTENT(IN) :: mo_coeff
205 : TYPE(cp_fm_type), INTENT(IN) :: fm_matrix_PQ
206 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:, :) :: fm_matrix_L_kpoints, &
207 : fm_matrix_Minv_L_kpoints, &
208 : fm_matrix_Minv, &
209 : fm_matrix_Minv_Vtrunc_Minv
210 : TYPE(kpoint_type), POINTER :: kpoints
211 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :), &
212 : INTENT(INOUT) :: Eigenval
213 : INTEGER, INTENT(IN) :: nmo
214 : INTEGER, DIMENSION(:), INTENT(IN) :: homo
215 : INTEGER, INTENT(IN) :: dimen_RI, dimen_RI_red
216 : INTEGER, DIMENSION(:), INTENT(IN) :: gw_corr_lev_occ, gw_corr_lev_virt
217 : INTEGER, INTENT(IN) :: bse_lev_virt, unit_nr
218 : LOGICAL, INTENT(IN) :: do_ri_sos_laplace_mp2, my_do_gw, &
219 : do_im_time, do_bse
220 : TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: matrix_s
221 : TYPE(dbcsr_p_type), INTENT(IN) :: mat_munu
222 : TYPE(dbcsr_p_type), INTENT(INOUT) :: mat_P_global
223 : TYPE(dbt_type) :: t_3c_M
224 : TYPE(dbt_type), ALLOCATABLE, DIMENSION(:, :) :: t_3c_O
225 : TYPE(hfx_compression_type), ALLOCATABLE, &
226 : DIMENSION(:, :, :), INTENT(INOUT) :: t_3c_O_compressed
227 : TYPE(block_ind_type), ALLOCATABLE, &
228 : DIMENSION(:, :, :), INTENT(INOUT) :: t_3c_O_ind
229 : INTEGER, ALLOCATABLE, DIMENSION(:), INTENT(IN) :: starts_array_mc, ends_array_mc, &
230 : starts_array_mc_block, &
231 : ends_array_mc_block
232 : LOGICAL, INTENT(IN) :: calc_forces
233 :
234 : CHARACTER(LEN=*), PARAMETER :: routineN = 'rpa_ri_compute_en'
235 :
236 : INTEGER :: best_integ_group_size, best_num_integ_point, color_rpa_group, dimen_homo_square, &
237 : dimen_nm_gw, dimen_virt_square, handle, handle2, handle3, ierr, iiB, &
238 : input_num_integ_groups, integ_group_size, ispin, jjB, min_integ_group_size, &
239 : my_ab_comb_bse_end, my_ab_comb_bse_size, my_ab_comb_bse_start, my_group_L_end, &
240 : my_group_L_size, my_group_L_start, my_ij_comb_bse_end, my_ij_comb_bse_size, &
241 : my_ij_comb_bse_start, my_nm_gw_end, my_nm_gw_size, my_nm_gw_start, ncol_block_mat, &
242 : ngroup, nrow_block_mat, nspins, num_integ_group, num_integ_points, pos_integ_group
243 : INTEGER(KIND=int_8) :: mem
244 592 : INTEGER, ALLOCATABLE, DIMENSION(:) :: dimen_ia, my_ia_end, my_ia_size, &
245 296 : my_ia_start, virtual
246 : LOGICAL :: do_kpoints_from_Gamma, do_minimax_quad, &
247 : my_open_shell, skip_integ_group_opt
248 : REAL(KIND=dp) :: allowed_memory, avail_mem, E_Range, Emax, Emin, mem_for_iaK, mem_for_QK, &
249 : mem_min, mem_per_group, mem_per_rank, mem_per_repl, mem_real
250 296 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :, :) :: Eigenval_kp
251 592 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:) :: fm_mat_Q, fm_mat_Q_gemm, fm_mat_S, &
252 296 : fm_mat_S_gw
253 888 : TYPE(cp_fm_type), DIMENSION(1) :: fm_mat_R_gw, fm_mat_S_ab_bse, &
254 592 : fm_mat_S_ij_bse
255 : TYPE(mp_para_env_type), POINTER :: para_env_RPA
256 : TYPE(two_dim_real_array), ALLOCATABLE, &
257 592 : DIMENSION(:) :: BIb_C_2D, BIb_C_2D_gw
258 2960 : TYPE(two_dim_real_array), DIMENSION(1) :: BIb_C_2D_bse_ab, BIb_C_2D_bse_ij
259 :
260 296 : CALL timeset(routineN, handle)
261 :
262 296 : CALL cite_reference(DelBen2013)
263 296 : CALL cite_reference(DelBen2015)
264 :
265 296 : IF (mp2_env%ri_rpa%exchange_correction == rpa_exchange_axk) THEN
266 10 : CALL cite_reference(Bates2013)
267 286 : ELSE IF (mp2_env%ri_rpa%exchange_correction == rpa_exchange_sosex) THEN
268 2 : CALL cite_reference(Freeman1977)
269 2 : CALL cite_reference(Gruneis2009)
270 : END IF
271 296 : IF (mp2_env%ri_rpa%do_rse) THEN
272 4 : CALL cite_reference(Ren2011)
273 4 : CALL cite_reference(Ren2013)
274 : END IF
275 :
276 296 : IF (my_do_gw) THEN
277 106 : CALL cite_reference(Wilhelm2016a)
278 106 : CALL cite_reference(Wilhelm2017)
279 106 : CALL cite_reference(Wilhelm2018)
280 : END IF
281 :
282 296 : IF (do_im_time) THEN
283 134 : CALL cite_reference(Wilhelm2016b)
284 : END IF
285 :
286 296 : nspins = SIZE(homo)
287 296 : my_open_shell = (nspins == 2)
288 2072 : ALLOCATE (virtual(nspins), dimen_ia(nspins), my_ia_end(nspins), my_ia_start(nspins), my_ia_size(nspins))
289 650 : virtual(:) = nmo - homo(:)
290 650 : dimen_ia(:) = virtual(:)*homo(:)
291 :
292 1184 : ALLOCATE (Eigenval_kp(nmo, 1, nspins))
293 8930 : Eigenval_kp(:, 1, :) = Eigenval(:, :)
294 :
295 296 : IF (do_im_time) mp2_env%ri_rpa%minimax_quad = .TRUE.
296 296 : do_minimax_quad = mp2_env%ri_rpa%minimax_quad
297 :
298 296 : IF (do_ri_sos_laplace_mp2) THEN
299 58 : num_integ_points = mp2_env%ri_laplace%n_quadrature
300 58 : input_num_integ_groups = mp2_env%ri_laplace%num_integ_groups
301 :
302 : ! check the range for the minimax approximation
303 58 : E_Range = mp2_env%e_range
304 58 : IF (mp2_env%e_range <= 1.0_dp .OR. mp2_env%e_gap <= 0.0_dp) THEN
305 : Emin = HUGE(dp)
306 : Emax = 0.0_dp
307 88 : DO ispin = 1, nspins
308 88 : IF (homo(ispin) > 0) THEN
309 50 : Emin = MIN(Emin, 2.0_dp*(Eigenval(homo(ispin) + 1, ispin) - Eigenval(homo(ispin), ispin)))
310 2740 : Emax = MAX(Emax, 2.0_dp*(MAXVAL(Eigenval(:, ispin)) - MINVAL(Eigenval(:, ispin))))
311 : END IF
312 : END DO
313 38 : E_Range = Emax/Emin
314 : END IF
315 58 : IF (E_Range < 2.0_dp) E_Range = 2.0_dp
316 : ierr = 0
317 58 : CALL check_exp_minimax_range(num_integ_points, E_Range, ierr)
318 58 : IF (ierr /= 0) THEN
319 : jjB = num_integ_points - 1
320 0 : DO iiB = 1, jjB
321 0 : num_integ_points = num_integ_points - 1
322 : ierr = 0
323 0 : CALL check_exp_minimax_range(num_integ_points, E_Range, ierr)
324 0 : IF (ierr == 0) EXIT
325 : END DO
326 : END IF
327 58 : CPASSERT(num_integ_points >= 1)
328 : ELSE
329 238 : num_integ_points = mp2_env%ri_rpa%rpa_num_quad_points
330 238 : input_num_integ_groups = mp2_env%ri_rpa%rpa_num_integ_groups
331 238 : IF (my_do_gw .AND. do_minimax_quad) THEN
332 46 : IF (num_integ_points > 34) THEN
333 0 : IF (unit_nr > 0) &
334 : CALL cp_warn(__LOCATION__, &
335 : "The required number of quadrature point exceeds the maximum possible in the "// &
336 0 : "Minimax quadrature scheme. The number of quadrature point has been reset to 30.")
337 0 : num_integ_points = 30
338 : END IF
339 : ELSE
340 192 : IF (do_minimax_quad .AND. num_integ_points > 20) THEN
341 0 : IF (unit_nr > 0) &
342 : CALL cp_warn(__LOCATION__, &
343 : "The required number of quadrature point exceeds the maximum possible in the "// &
344 0 : "Minimax quadrature scheme. The number of quadrature point has been reset to 20.")
345 0 : num_integ_points = 20
346 : END IF
347 : END IF
348 : END IF
349 296 : allowed_memory = mp2_env%mp2_memory
350 :
351 296 : CALL get_group_dist(gd_array, color_sub, my_group_L_start, my_group_L_end, my_group_L_size)
352 :
353 296 : ngroup = para_env%num_pe/para_env_sub%num_pe
354 :
355 : ! for imaginary time or periodic GW or BSE, we use all processors for a single frequency/time point
356 296 : IF (do_im_time .OR. mp2_env%ri_g0w0%do_periodic .OR. do_bse) THEN
357 :
358 168 : integ_group_size = ngroup
359 168 : best_num_integ_point = num_integ_points
360 :
361 : ELSE
362 :
363 : ! Calculate available memory and create integral group according to that
364 : ! mem_for_iaK is the memory needed for storing the 3 centre integrals
365 284 : mem_for_iaK = REAL(SUM(dimen_ia), KIND=dp)*dimen_RI_red*8.0_dp/(1024_dp**2)
366 128 : mem_for_QK = REAL(dimen_RI_red, KIND=dp)*nspins*dimen_RI_red*8.0_dp/(1024_dp**2)
367 :
368 128 : CALL m_memory(mem)
369 128 : mem_real = (mem + 1024*1024 - 1)/(1024*1024)
370 128 : CALL para_env%min(mem_real)
371 :
372 128 : mem_per_rank = 0.0_dp
373 :
374 : ! B_ia_P
375 : mem_per_repl = mem_for_iaK
376 : ! Q (regular and for dgemm)
377 128 : mem_per_repl = mem_per_repl + 2.0_dp*mem_for_QK
378 :
379 128 : IF (calc_forces) CALL rpa_grad_needed_mem(homo, virtual, dimen_RI_red, mem_per_rank, mem_per_repl, do_ri_sos_laplace_mp2)
380 128 : CALL rpa_exchange_needed_mem(mp2_env, homo, virtual, dimen_RI_red, para_env, mem_per_rank, mem_per_repl)
381 :
382 128 : mem_min = mem_per_repl/para_env%num_pe + mem_per_rank
383 :
384 128 : IF (unit_nr > 0) THEN
385 64 : WRITE (unit_nr, '(T3,A,T68,F9.2,A4)') 'RI_INFO| Minimum required memory per MPI process:', mem_min, ' MiB'
386 64 : WRITE (unit_nr, '(T3,A,T68,F9.2,A4)') 'RI_INFO| Available memory per MPI process:', mem_real, ' MiB'
387 : END IF
388 :
389 : ! Use only the allowed amount of memory
390 128 : mem_real = MIN(mem_real, allowed_memory)
391 : ! For the memory estimate, we require the amount of required memory per replication group and the available memory
392 128 : mem_real = mem_real - mem_per_rank
393 :
394 128 : mem_per_group = mem_real*para_env_sub%num_pe
395 :
396 : ! here we try to find the best rpa/laplace group size
397 128 : skip_integ_group_opt = .FALSE.
398 :
399 : ! Check the input number of integration groups
400 128 : IF (input_num_integ_groups > 0) THEN
401 2 : IF (MOD(num_integ_points, input_num_integ_groups) == 0) THEN
402 0 : IF (MOD(ngroup, input_num_integ_groups) == 0) THEN
403 0 : best_integ_group_size = ngroup/input_num_integ_groups
404 0 : best_num_integ_point = num_integ_points/input_num_integ_groups
405 0 : skip_integ_group_opt = .TRUE.
406 : ELSE
407 0 : IF (unit_nr > 0) WRITE (unit_nr, '(T3,A)') 'Total number of groups not multiple of NUM_INTEG_GROUPS'
408 : END IF
409 : ELSE
410 2 : IF (unit_nr > 0) WRITE (unit_nr, '(T3,A)') 'NUM_QUAD_POINTS not multiple of NUM_INTEG_GROUPS'
411 : END IF
412 : END IF
413 :
414 128 : IF (.NOT. skip_integ_group_opt) THEN
415 128 : best_integ_group_size = ngroup
416 128 : best_num_integ_point = num_integ_points
417 :
418 128 : min_integ_group_size = MAX(1, ngroup/num_integ_points)
419 :
420 128 : integ_group_size = min_integ_group_size - 1
421 142 : DO iiB = min_integ_group_size + 1, ngroup
422 106 : integ_group_size = integ_group_size + 1
423 :
424 : ! check that the ngroup is a multiple of integ_group_size
425 106 : IF (MOD(ngroup, integ_group_size) /= 0) CYCLE
426 :
427 : ! check for memory
428 106 : avail_mem = integ_group_size*mem_per_group
429 106 : IF (avail_mem < mem_per_repl) CYCLE
430 :
431 : ! check that the integration groups have the same size
432 106 : num_integ_group = ngroup/integ_group_size
433 106 : IF (MOD(num_integ_points, num_integ_group) /= 0) CYCLE
434 :
435 92 : best_num_integ_point = num_integ_points/num_integ_group
436 92 : best_integ_group_size = integ_group_size
437 :
438 142 : EXIT
439 :
440 : END DO
441 : END IF
442 :
443 128 : integ_group_size = best_integ_group_size
444 :
445 : END IF
446 :
447 296 : IF (unit_nr > 0 .AND. .NOT. do_im_time) THEN
448 81 : IF (do_ri_sos_laplace_mp2) THEN
449 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
450 14 : "RI_INFO| Group size for laplace numerical integration:", integ_group_size*para_env_sub%num_pe
451 : WRITE (UNIT=unit_nr, FMT="(T3,A)") &
452 14 : "INTEG_INFO| MINIMAX approximation"
453 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
454 14 : "INTEG_INFO| Number of integration points:", num_integ_points
455 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
456 14 : "INTEG_INFO| Number of integration points per Laplace group:", best_num_integ_point
457 : ELSE
458 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
459 67 : "RI_INFO| Group size for frequency integration:", integ_group_size*para_env_sub%num_pe
460 67 : IF (do_minimax_quad) THEN
461 : WRITE (UNIT=unit_nr, FMT="(T3,A)") &
462 19 : "INTEG_INFO| MINIMAX quadrature"
463 : ELSE
464 : WRITE (UNIT=unit_nr, FMT="(T3,A)") &
465 48 : "INTEG_INFO| Clenshaw-Curtius quadrature"
466 : END IF
467 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
468 67 : "INTEG_INFO| Number of integration points:", num_integ_points
469 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
470 67 : "INTEG_INFO| Number of integration points per RPA group:", best_num_integ_point
471 : END IF
472 81 : CALL m_flush(unit_nr)
473 : END IF
474 :
475 296 : num_integ_group = ngroup/integ_group_size
476 :
477 296 : pos_integ_group = MOD(color_sub, integ_group_size)
478 296 : color_rpa_group = color_sub/integ_group_size
479 :
480 296 : CALL timeset(routineN//"_reorder", handle2)
481 :
482 : ! not necessary for imaginary time
483 :
484 1242 : ALLOCATE (BIb_C_2D(nspins))
485 :
486 296 : IF (.NOT. do_im_time) THEN
487 :
488 : ! reorder the local data in such a way to help the next stage of matrix creation
489 : ! now the data inside the group are divided into a ia x K matrix
490 352 : DO ispin = 1, nspins
491 : CALL calculate_BIb_C_2D(BIb_C_2D(ispin)%array, BIb_C(ispin)%array, para_env_sub, dimen_ia(ispin), &
492 : homo(ispin), virtual(ispin), gd_B_virtual(ispin), &
493 190 : my_ia_size(ispin), my_ia_start(ispin), my_ia_end(ispin), my_group_L_size)
494 :
495 190 : DEALLOCATE (BIb_C(ispin)%array)
496 542 : CALL release_group_dist(gd_B_virtual(ispin))
497 :
498 : END DO
499 :
500 : ! in the GW case, BIb_C_2D_gw is an nm x K matrix, with n: number of corr GW levels, m=nmo
501 162 : IF (my_do_gw) THEN
502 184 : ALLOCATE (BIb_C_2D_gw(nspins))
503 :
504 60 : CALL timeset(routineN//"_reorder_gw", handle3)
505 :
506 60 : dimen_nm_gw = nmo*(gw_corr_lev_occ(1) + gw_corr_lev_virt(1))
507 :
508 : ! The same for open shell
509 124 : DO ispin = 1, nspins
510 : CALL calculate_BIb_C_2D(BIb_C_2D_gw(ispin)%array, BIb_C_gw(ispin)%array, para_env_sub, dimen_nm_gw, &
511 : gw_corr_lev_occ(ispin) + gw_corr_lev_virt(ispin), nmo, gd_B_all, &
512 64 : my_nm_gw_size, my_nm_gw_start, my_nm_gw_end, my_group_L_size)
513 124 : DEALLOCATE (BIb_C_gw(ispin)%array)
514 : END DO
515 :
516 60 : CALL release_group_dist(gd_B_all)
517 :
518 120 : CALL timestop(handle3)
519 :
520 : END IF
521 : END IF
522 :
523 360 : IF (do_bse) THEN
524 :
525 32 : CALL timeset(routineN//"_reorder_bse1", handle3)
526 :
527 32 : dimen_homo_square = homo(1)**2
528 : ! We do not implement an explicit bse_lev_occ different to homo here, because the small number of occupied levels
529 : ! does not critically influence the memory
530 : CALL calculate_BIb_C_2D(BIb_C_2D_bse_ij(1)%array, BIb_C_bse_ij, para_env_sub, dimen_homo_square, &
531 : homo(1), homo(1), gd_B_occ_bse, &
532 32 : my_ij_comb_bse_size, my_ij_comb_bse_start, my_ij_comb_bse_end, my_group_L_size)
533 :
534 32 : DEALLOCATE (BIb_C_bse_ij)
535 32 : CALL release_group_dist(gd_B_occ_bse)
536 :
537 32 : CALL timestop(handle3)
538 :
539 32 : CALL timeset(routineN//"_reorder_bse2", handle3)
540 :
541 32 : dimen_virt_square = bse_lev_virt**2
542 :
543 : CALL calculate_BIb_C_2D(BIb_C_2D_bse_ab(1)%array, BIb_C_bse_ab, para_env_sub, dimen_virt_square, &
544 : bse_lev_virt, bse_lev_virt, gd_B_virt_bse, &
545 32 : my_ab_comb_bse_size, my_ab_comb_bse_start, my_ab_comb_bse_end, my_group_L_size)
546 :
547 32 : DEALLOCATE (BIb_C_bse_ab)
548 32 : CALL release_group_dist(gd_B_virt_bse)
549 :
550 32 : CALL timestop(handle3)
551 :
552 : END IF
553 :
554 296 : CALL timestop(handle2)
555 :
556 296 : IF (num_integ_group > 1) THEN
557 92 : ALLOCATE (para_env_RPA)
558 92 : CALL para_env_RPA%from_split(para_env, color_rpa_group)
559 : ELSE
560 204 : para_env_RPA => para_env
561 : END IF
562 :
563 : ! now create the matrices needed for the calculation, Q, S and G
564 : ! Q and G will have omega dependence
565 :
566 296 : IF (do_im_time) THEN
567 834 : ALLOCATE (fm_mat_Q(nspins), fm_mat_Q_gemm(1), fm_mat_S(1))
568 : ELSE
569 1380 : ALLOCATE (fm_mat_Q(nspins), fm_mat_Q_gemm(nspins), fm_mat_S(nspins))
570 : END IF
571 :
572 : CALL create_integ_mat(BIb_C_2D, para_env, para_env_sub, color_sub, ngroup, integ_group_size, &
573 : dimen_RI_red, dimen_ia, color_rpa_group, &
574 : mp2_env%block_size_row, mp2_env%block_size_col, unit_nr, &
575 : my_ia_size, my_ia_start, my_ia_end, &
576 : my_group_L_size, my_group_L_start, my_group_L_end, &
577 : para_env_RPA, fm_mat_S, nrow_block_mat, ncol_block_mat, &
578 : dimen_ia_for_block_size=dimen_ia(1), &
579 296 : do_im_time=do_im_time, fm_mat_Q_gemm=fm_mat_Q_gemm, fm_mat_Q=fm_mat_Q, qs_env=qs_env)
580 :
581 650 : DEALLOCATE (BIb_C_2D, my_ia_end, my_ia_size, my_ia_start)
582 :
583 : ! for GW, we need other matrix fm_mat_S, always allocate the container to prevent crying compilers
584 1242 : ALLOCATE (fm_mat_S_gw(nspins))
585 296 : IF (my_do_gw .AND. .NOT. do_im_time) THEN
586 :
587 : CALL create_integ_mat(BIb_C_2D_gw, para_env, para_env_sub, color_sub, ngroup, integ_group_size, &
588 : dimen_RI_red, [dimen_nm_gw, dimen_nm_gw], color_rpa_group, &
589 : mp2_env%block_size_row, mp2_env%block_size_col, unit_nr, &
590 : [my_nm_gw_size, my_nm_gw_size], [my_nm_gw_start, my_nm_gw_start], [my_nm_gw_end, my_nm_gw_end], &
591 : my_group_L_size, my_group_L_start, my_group_L_end, &
592 : para_env_RPA, fm_mat_S_gw, nrow_block_mat, ncol_block_mat, &
593 : fm_mat_Q(1)%matrix_struct%context, fm_mat_Q(1)%matrix_struct%context, &
594 540 : fm_mat_Q=fm_mat_R_gw)
595 124 : DEALLOCATE (BIb_C_2D_gw)
596 :
597 : END IF
598 :
599 : ! for Bethe-Salpeter, we need other matrix fm_mat_S
600 296 : IF (do_bse) THEN
601 : CALL create_integ_mat(BIb_C_2D_bse_ij, para_env, para_env_sub, color_sub, ngroup, integ_group_size, &
602 : dimen_RI_red, [dimen_homo_square], color_rpa_group, &
603 : mp2_env%block_size_row, mp2_env%block_size_col, unit_nr, &
604 : [my_ij_comb_bse_size], [my_ij_comb_bse_start], [my_ij_comb_bse_end], &
605 : my_group_L_size, my_group_L_start, my_group_L_end, &
606 : para_env_RPA, fm_mat_S_ij_bse, nrow_block_mat, ncol_block_mat, &
607 160 : fm_mat_Q(1)%matrix_struct%context, fm_mat_Q(1)%matrix_struct%context)
608 :
609 : CALL create_integ_mat(BIb_C_2D_bse_ab, para_env, para_env_sub, color_sub, ngroup, integ_group_size, &
610 : dimen_RI_red, [dimen_virt_square], color_rpa_group, &
611 : mp2_env%block_size_row, mp2_env%block_size_col, unit_nr, &
612 : [my_ab_comb_bse_size], [my_ab_comb_bse_start], [my_ab_comb_bse_end], &
613 : my_group_L_size, my_group_L_start, my_group_L_end, &
614 : para_env_RPA, fm_mat_S_ab_bse, nrow_block_mat, ncol_block_mat, &
615 160 : fm_mat_Q(1)%matrix_struct%context, fm_mat_Q(1)%matrix_struct%context)
616 :
617 : END IF
618 :
619 296 : do_kpoints_from_Gamma = qs_env%mp2_env%ri_rpa_im_time%do_kpoints_from_Gamma
620 296 : IF (do_kpoints_from_Gamma) THEN
621 18 : CALL get_bandstruc_and_k_dependent_MOs(qs_env, Eigenval_kp)
622 : END IF
623 :
624 : ! Now start the RPA calculation
625 : ! fm_mo_coeff_occ, fm_mo_coeff_virt will be deallocated here
626 : CALL rpa_num_int(qs_env, Erpa, mp2_env, para_env, para_env_RPA, para_env_sub, unit_nr, &
627 : homo, virtual, dimen_RI, dimen_RI_red, dimen_ia, dimen_nm_gw, &
628 : Eigenval_kp, num_integ_points, num_integ_group, color_rpa_group, &
629 : fm_matrix_PQ, fm_mat_S, fm_mat_Q_gemm, fm_mat_Q, fm_mat_S_gw, fm_mat_R_gw(1), &
630 : fm_mat_S_ij_bse(1), fm_mat_S_ab_bse(1), &
631 : my_do_gw, do_bse, gw_corr_lev_occ, gw_corr_lev_virt, &
632 : bse_lev_virt, &
633 : do_minimax_quad, &
634 : do_im_time, mo_coeff, &
635 : fm_matrix_L_kpoints, fm_matrix_Minv_L_kpoints, &
636 : fm_matrix_Minv, fm_matrix_Minv_Vtrunc_Minv, mat_munu, mat_P_global, &
637 : t_3c_M, t_3c_O, t_3c_O_compressed, t_3c_O_ind, &
638 : starts_array_mc, ends_array_mc, &
639 : starts_array_mc_block, ends_array_mc_block, &
640 : matrix_s, do_kpoints_from_Gamma, kpoints, gd_array, color_sub, &
641 296 : do_ri_sos_laplace_mp2=do_ri_sos_laplace_mp2, calc_forces=calc_forces)
642 :
643 296 : CALL release_group_dist(gd_array)
644 :
645 296 : IF (num_integ_group > 1) CALL mp_para_env_release(para_env_RPA)
646 :
647 296 : IF (.NOT. do_im_time) THEN
648 162 : CALL cp_fm_release(fm_mat_Q_gemm)
649 162 : CALL cp_fm_release(fm_mat_S)
650 : END IF
651 296 : CALL cp_fm_release(fm_mat_Q)
652 :
653 296 : IF (my_do_gw .AND. .NOT. do_im_time) THEN
654 60 : CALL cp_fm_release(fm_mat_S_gw)
655 60 : CALL cp_fm_release(fm_mat_R_gw(1))
656 : END IF
657 :
658 296 : IF (do_bse) THEN
659 32 : CALL cp_fm_release(fm_mat_S_ij_bse(1))
660 32 : CALL cp_fm_release(fm_mat_S_ab_bse(1))
661 : END IF
662 :
663 296 : CALL timestop(handle)
664 :
665 1184 : END SUBROUTINE rpa_ri_compute_en
666 :
667 : ! **************************************************************************************************
668 : !> \brief reorder the local data in such a way to help the next stage of matrix creation;
669 : !> now the data inside the group are divided into a ia x K matrix (BIb_C_2D);
670 : !> Subroutine created to avoid massive double coding
671 : !> \param BIb_C_2D ...
672 : !> \param BIb_C ...
673 : !> \param para_env_sub ...
674 : !> \param dimen_ia ...
675 : !> \param homo ...
676 : !> \param virtual ...
677 : !> \param gd_B_virtual ...
678 : !> \param my_ia_size ...
679 : !> \param my_ia_start ...
680 : !> \param my_ia_end ...
681 : !> \param my_group_L_size ...
682 : !> \author Jan Wilhelm, 03/2015
683 : ! **************************************************************************************************
684 318 : SUBROUTINE calculate_BIb_C_2D(BIb_C_2D, BIb_C, para_env_sub, dimen_ia, homo, virtual, &
685 : gd_B_virtual, &
686 : my_ia_size, my_ia_start, my_ia_end, my_group_L_size)
687 :
688 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :), &
689 : INTENT(OUT) :: BIb_C_2D
690 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :, :), &
691 : INTENT(IN) :: BIb_C
692 : TYPE(mp_para_env_type), INTENT(IN) :: para_env_sub
693 : INTEGER, INTENT(IN) :: dimen_ia, homo, virtual
694 : TYPE(group_dist_d1_type), INTENT(INOUT) :: gd_B_virtual
695 : INTEGER :: my_ia_size, my_ia_start, my_ia_end, &
696 : my_group_L_size
697 :
698 : INTEGER, PARAMETER :: occ_chunk = 128
699 :
700 : INTEGER :: ia_global, iiB, itmp(2), jjB, my_B_size, my_B_virtual_start, occ_high, occ_low, &
701 : proc_receive, proc_send, proc_shift, rec_B_size, rec_B_virtual_end, rec_B_virtual_start
702 318 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:), TARGET :: BIb_C_rec_1D
703 318 : REAL(KIND=dp), DIMENSION(:, :, :), POINTER :: BIb_C_rec
704 :
705 318 : itmp = get_limit(dimen_ia, para_env_sub%num_pe, para_env_sub%mepos)
706 318 : my_ia_start = itmp(1)
707 318 : my_ia_end = itmp(2)
708 318 : my_ia_size = my_ia_end - my_ia_start + 1
709 :
710 318 : CALL get_group_dist(gd_B_virtual, para_env_sub%mepos, sizes=my_B_size, starts=my_B_virtual_start)
711 :
712 : ! reorder data
713 1266 : ALLOCATE (BIb_C_2D(my_group_L_size, my_ia_size))
714 :
715 : !$OMP PARALLEL DO DEFAULT(NONE) PRIVATE(jjB,ia_global) &
716 : !$OMP SHARED(homo,my_B_size,virtual,my_B_virtual_start,my_ia_start,my_ia_end,BIb_C,BIb_C_2D,&
717 318 : !$OMP my_group_L_size)
718 : DO iiB = 1, homo
719 : DO jjB = 1, my_B_size
720 : ia_global = (iiB - 1)*virtual + my_B_virtual_start + jjB - 1
721 : IF (ia_global >= my_ia_start .AND. ia_global <= my_ia_end) THEN
722 : BIb_C_2D(1:my_group_L_size, ia_global - my_ia_start + 1) = BIb_C(1:my_group_L_size, jjB, iiB)
723 : END IF
724 : END DO
725 : END DO
726 :
727 318 : IF (para_env_sub%num_pe > 1) THEN
728 30 : ALLOCATE (BIb_C_rec_1D(INT(my_group_L_size, int_8)*maxsize(gd_B_virtual)*MIN(homo, occ_chunk)))
729 20 : DO proc_shift = 1, para_env_sub%num_pe - 1
730 10 : proc_send = MODULO(para_env_sub%mepos + proc_shift, para_env_sub%num_pe)
731 10 : proc_receive = MODULO(para_env_sub%mepos - proc_shift, para_env_sub%num_pe)
732 :
733 10 : CALL get_group_dist(gd_B_virtual, proc_receive, rec_B_virtual_start, rec_B_virtual_end, rec_B_size)
734 :
735 : ! do this in chunks to avoid high memory overhead
736 20 : DO occ_low = 1, homo, occ_chunk
737 10 : occ_high = MIN(homo, occ_low + occ_chunk - 1)
738 : BIb_C_rec(1:my_group_L_size, 1:rec_B_size, 1:occ_high - occ_low + 1) => &
739 10 : BIb_C_rec_1D(1:INT(my_group_L_size, int_8)*rec_B_size*(occ_high - occ_low + 1))
740 : CALL para_env_sub%sendrecv(BIb_C(:, :, occ_low:occ_high), proc_send, &
741 31970 : BIb_C_rec(:, :, 1:occ_high - occ_low + 1), proc_receive)
742 : !$OMP PARALLEL DO DEFAULT(NONE) PRIVATE(jjB,ia_global) &
743 : !$OMP SHARED(occ_low,occ_high,rec_B_size,virtual,rec_B_virtual_start,my_ia_start,my_ia_end,BIb_C_rec,BIb_C_2D,&
744 10 : !$OMP my_group_L_size)
745 : DO iiB = occ_low, occ_high
746 : DO jjB = 1, rec_B_size
747 : ia_global = (iiB - 1)*virtual + rec_B_virtual_start + jjB - 1
748 : IF (ia_global >= my_ia_start .AND. ia_global <= my_ia_end) THEN
749 : BIb_C_2D(1:my_group_L_size, ia_global - my_ia_start + 1) = BIb_C_rec(1:my_group_L_size, jjB, iiB - occ_low + 1)
750 : END IF
751 : END DO
752 : END DO
753 : END DO
754 :
755 : END DO
756 10 : DEALLOCATE (BIb_C_rec_1D)
757 : END IF
758 :
759 318 : END SUBROUTINE calculate_BIb_C_2D
760 :
761 : ! **************************************************************************************************
762 : !> \brief ...
763 : !> \param BIb_C_2D ...
764 : !> \param para_env ...
765 : !> \param para_env_sub ...
766 : !> \param color_sub ...
767 : !> \param ngroup ...
768 : !> \param integ_group_size ...
769 : !> \param dimen_RI ...
770 : !> \param dimen_ia ...
771 : !> \param color_rpa_group ...
772 : !> \param ext_row_block_size ...
773 : !> \param ext_col_block_size ...
774 : !> \param unit_nr ...
775 : !> \param my_ia_size ...
776 : !> \param my_ia_start ...
777 : !> \param my_ia_end ...
778 : !> \param my_group_L_size ...
779 : !> \param my_group_L_start ...
780 : !> \param my_group_L_end ...
781 : !> \param para_env_RPA ...
782 : !> \param fm_mat_S ...
783 : !> \param nrow_block_mat ...
784 : !> \param ncol_block_mat ...
785 : !> \param blacs_env_ext ...
786 : !> \param blacs_env_ext_S ...
787 : !> \param dimen_ia_for_block_size ...
788 : !> \param do_im_time ...
789 : !> \param fm_mat_Q_gemm ...
790 : !> \param fm_mat_Q ...
791 : !> \param qs_env ...
792 : ! **************************************************************************************************
793 420 : SUBROUTINE create_integ_mat(BIb_C_2D, para_env, para_env_sub, color_sub, ngroup, integ_group_size, &
794 420 : dimen_RI, dimen_ia, color_rpa_group, &
795 : ext_row_block_size, ext_col_block_size, unit_nr, &
796 420 : my_ia_size, my_ia_start, my_ia_end, &
797 : my_group_L_size, my_group_L_start, my_group_L_end, &
798 420 : para_env_RPA, fm_mat_S, nrow_block_mat, ncol_block_mat, &
799 : blacs_env_ext, blacs_env_ext_S, dimen_ia_for_block_size, &
800 420 : do_im_time, fm_mat_Q_gemm, fm_mat_Q, qs_env)
801 :
802 : TYPE(two_dim_real_array), DIMENSION(:), &
803 : INTENT(INOUT) :: BIb_C_2D
804 : TYPE(mp_para_env_type), INTENT(IN) :: para_env, para_env_sub
805 : INTEGER, INTENT(IN) :: color_sub, ngroup, integ_group_size, &
806 : dimen_RI
807 : INTEGER, DIMENSION(:), INTENT(IN) :: dimen_ia
808 : INTEGER, INTENT(IN) :: color_rpa_group, ext_row_block_size, &
809 : ext_col_block_size, unit_nr
810 : INTEGER, DIMENSION(:), INTENT(IN) :: my_ia_size, my_ia_start, my_ia_end
811 : INTEGER, INTENT(IN) :: my_group_L_size, my_group_L_start, &
812 : my_group_L_end
813 : TYPE(mp_para_env_type), INTENT(IN), POINTER :: para_env_RPA
814 : TYPE(cp_fm_type), DIMENSION(:), INTENT(INOUT) :: fm_mat_S
815 : INTEGER, INTENT(INOUT) :: nrow_block_mat, ncol_block_mat
816 : TYPE(cp_blacs_env_type), OPTIONAL, POINTER :: blacs_env_ext, blacs_env_ext_S
817 : INTEGER, INTENT(IN), OPTIONAL :: dimen_ia_for_block_size
818 : LOGICAL, INTENT(IN), OPTIONAL :: do_im_time
819 : TYPE(cp_fm_type), DIMENSION(:), OPTIONAL :: fm_mat_Q_gemm, fm_mat_Q
820 : TYPE(qs_environment_type), INTENT(IN), OPTIONAL, &
821 : POINTER :: qs_env
822 :
823 : CHARACTER(LEN=*), PARAMETER :: routineN = 'create_integ_mat'
824 :
825 : INTEGER :: col_row_proc_ratio, grid_2D(2), handle, &
826 : iproc, iproc_col, iproc_row, ispin, &
827 : mepos_in_RPA_group
828 420 : INTEGER, ALLOCATABLE, DIMENSION(:, :) :: group_grid_2_mepos
829 : LOGICAL :: my_blacs_ext, my_blacs_S_ext, &
830 : my_do_im_time
831 : TYPE(cp_blacs_env_type), POINTER :: blacs_env, blacs_env_Q
832 : TYPE(cp_fm_struct_type), POINTER :: fm_struct
833 420 : TYPE(group_dist_d1_type) :: gd_ia, gd_L
834 :
835 420 : CALL timeset(routineN, handle)
836 :
837 420 : CPASSERT(PRESENT(blacs_env_ext) .OR. PRESENT(dimen_ia_for_block_size))
838 :
839 420 : my_blacs_ext = .FALSE.
840 420 : IF (PRESENT(blacs_env_ext)) my_blacs_ext = .TRUE.
841 :
842 420 : my_blacs_S_ext = .FALSE.
843 420 : IF (PRESENT(blacs_env_ext_S)) my_blacs_S_ext = .TRUE.
844 :
845 420 : my_do_im_time = .FALSE.
846 420 : IF (PRESENT(do_im_time)) my_do_im_time = do_im_time
847 :
848 420 : NULLIFY (blacs_env)
849 : ! create the RPA blacs env
850 420 : IF (my_blacs_S_ext) THEN
851 124 : blacs_env => blacs_env_ext_S
852 : ELSE
853 296 : IF (para_env_RPA%num_pe > 1) THEN
854 204 : col_row_proc_ratio = MAX(1, dimen_ia_for_block_size/dimen_RI)
855 :
856 204 : iproc_col = MIN(MAX(INT(SQRT(REAL(para_env_RPA%num_pe*col_row_proc_ratio, KIND=dp))), 1), para_env_RPA%num_pe) + 1
857 204 : DO iproc = 1, para_env_RPA%num_pe
858 204 : iproc_col = iproc_col - 1
859 204 : IF (MOD(para_env_RPA%num_pe, iproc_col) == 0) EXIT
860 : END DO
861 :
862 204 : iproc_row = para_env_RPA%num_pe/iproc_col
863 204 : grid_2D(1) = iproc_row
864 204 : grid_2D(2) = iproc_col
865 : ELSE
866 276 : grid_2D = 1
867 : END IF
868 296 : CALL cp_blacs_env_create(blacs_env=blacs_env, para_env=para_env_RPA, grid_2d=grid_2D)
869 :
870 296 : IF (unit_nr > 0 .AND. .NOT. my_do_im_time) THEN
871 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
872 81 : "MATRIX_INFO| Number row processes:", grid_2D(1)
873 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
874 81 : "MATRIX_INFO| Number column processes:", grid_2D(2)
875 : END IF
876 :
877 : ! define the block_size for the row
878 296 : IF (ext_row_block_size > 0) THEN
879 0 : nrow_block_mat = ext_row_block_size
880 : ELSE
881 296 : nrow_block_mat = MAX(1, dimen_RI/grid_2D(1)/2)
882 : END IF
883 :
884 : ! define the block_size for the column
885 296 : IF (ext_col_block_size > 0) THEN
886 0 : ncol_block_mat = ext_col_block_size
887 : ELSE
888 296 : ncol_block_mat = MAX(1, dimen_ia_for_block_size/grid_2D(2)/2)
889 : END IF
890 :
891 296 : IF (unit_nr > 0 .AND. .NOT. my_do_im_time) THEN
892 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
893 81 : "MATRIX_INFO| Row block size:", nrow_block_mat
894 : WRITE (UNIT=unit_nr, FMT="(T3,A,T75,i6)") &
895 81 : "MATRIX_INFO| Column block size:", ncol_block_mat
896 : END IF
897 : END IF
898 :
899 353 : IF (.NOT. my_do_im_time) THEN
900 604 : DO ispin = 1, SIZE(BIb_C_2D)
901 318 : NULLIFY (fm_struct)
902 318 : IF (my_blacs_ext) THEN
903 : CALL cp_fm_struct_create(fm_struct, context=blacs_env, nrow_global=dimen_RI, &
904 128 : ncol_global=dimen_ia(ispin), para_env=para_env_RPA)
905 : ELSE
906 : CALL cp_fm_struct_create(fm_struct, context=blacs_env, nrow_global=dimen_RI, &
907 : ncol_global=dimen_ia(ispin), para_env=para_env_RPA, &
908 190 : nrow_block=nrow_block_mat, ncol_block=ncol_block_mat, force_block=.TRUE.)
909 :
910 : END IF ! external blacs_env
911 :
912 318 : CALL create_group_dist(gd_ia, my_ia_start(ispin), my_ia_end(ispin), my_ia_size(ispin), para_env_RPA)
913 318 : CALL create_group_dist(gd_L, my_group_L_start, my_group_L_end, my_group_L_size, para_env_RPA)
914 :
915 : ! create the info array
916 :
917 318 : mepos_in_RPA_group = MOD(color_sub, integ_group_size)
918 1272 : ALLOCATE (group_grid_2_mepos(0:integ_group_size - 1, 0:para_env_sub%num_pe - 1))
919 1138 : group_grid_2_mepos = 0
920 318 : group_grid_2_mepos(mepos_in_RPA_group, para_env_sub%mepos) = para_env_RPA%mepos
921 318 : CALL para_env_RPA%sum(group_grid_2_mepos)
922 :
923 : CALL array2fm(BIb_C_2D(ispin)%array, fm_struct, my_group_L_start, my_group_L_end, &
924 : my_ia_start(ispin), my_ia_end(ispin), gd_L, gd_ia, &
925 : group_grid_2_mepos, ngroup, para_env_sub%num_pe, fm_mat_S(ispin), &
926 318 : integ_group_size, color_rpa_group)
927 :
928 318 : DEALLOCATE (group_grid_2_mepos)
929 318 : CALL cp_fm_struct_release(fm_struct)
930 :
931 : ! deallocate the info array
932 318 : CALL release_group_dist(gd_L)
933 318 : CALL release_group_dist(gd_ia)
934 :
935 : ! sum the local data across processes belonging to different RPA group.
936 604 : IF (para_env_RPA%num_pe /= para_env%num_pe) THEN
937 : BLOCK
938 : TYPE(mp_comm_type) :: comm_exchange
939 144 : comm_exchange = fm_mat_S(ispin)%matrix_struct%context%interconnect(para_env)
940 144 : CALL comm_exchange%sum(fm_mat_S(ispin)%local_data)
941 288 : CALL comm_exchange%free()
942 : END BLOCK
943 : END IF
944 : END DO
945 : END IF
946 :
947 420 : IF (PRESENT(fm_mat_Q_gemm) .AND. .NOT. my_do_im_time) THEN
948 : ! create the Q matrix dimen_RIxdimen_RI where the result of the mat-mat-mult will be stored
949 162 : NULLIFY (fm_struct)
950 : CALL cp_fm_struct_create(fm_struct, context=blacs_env, nrow_global=dimen_RI, &
951 : ncol_global=dimen_RI, para_env=para_env_RPA, &
952 162 : nrow_block=nrow_block_mat, ncol_block=ncol_block_mat, force_block=.TRUE.)
953 352 : DO ispin = 1, SIZE(fm_mat_Q_gemm)
954 352 : CALL cp_fm_create(fm_mat_Q_gemm(ispin), fm_struct, name="fm_mat_Q_gemm")
955 : END DO
956 162 : CALL cp_fm_struct_release(fm_struct)
957 : END IF
958 :
959 420 : IF (PRESENT(fm_mat_Q)) THEN
960 356 : NULLIFY (blacs_env_Q)
961 356 : IF (my_blacs_ext) THEN
962 60 : blacs_env_Q => blacs_env_ext
963 296 : ELSE IF (para_env_RPA%num_pe == para_env%num_pe .AND. PRESENT(qs_env)) THEN
964 204 : CALL get_qs_env(qs_env, blacs_env=blacs_env_Q)
965 : ELSE
966 92 : CALL cp_blacs_env_create(blacs_env=blacs_env_Q, para_env=para_env_RPA)
967 : END IF
968 356 : NULLIFY (fm_struct)
969 : CALL cp_fm_struct_create(fm_struct, context=blacs_env_Q, nrow_global=dimen_RI, &
970 356 : ncol_global=dimen_RI, para_env=para_env_RPA)
971 770 : DO ispin = 1, SIZE(fm_mat_Q)
972 770 : CALL cp_fm_create(fm_mat_Q(ispin), fm_struct, name="fm_mat_Q")
973 : END DO
974 :
975 356 : CALL cp_fm_struct_release(fm_struct)
976 :
977 356 : IF (.NOT. (my_blacs_ext .OR. (para_env_RPA%num_pe == para_env%num_pe .AND. PRESENT(qs_env)))) &
978 92 : CALL cp_blacs_env_release(blacs_env_Q)
979 : END IF
980 :
981 : ! release blacs_env
982 420 : IF (.NOT. my_blacs_S_ext) THEN
983 296 : CALL cp_blacs_env_release(blacs_env)
984 : ELSE
985 124 : NULLIFY (blacs_env)
986 : END IF
987 :
988 420 : CALL timestop(handle)
989 :
990 420 : END SUBROUTINE create_integ_mat
991 :
992 : ! **************************************************************************************************
993 : !> \brief ...
994 : !> \param qs_env ...
995 : !> \param Erpa ...
996 : !> \param mp2_env ...
997 : !> \param para_env ...
998 : !> \param para_env_RPA ...
999 : !> \param para_env_sub ...
1000 : !> \param unit_nr ...
1001 : !> \param homo ...
1002 : !> \param virtual ...
1003 : !> \param dimen_RI ...
1004 : !> \param dimen_RI_red ...
1005 : !> \param dimen_ia ...
1006 : !> \param dimen_nm_gw ...
1007 : !> \param Eigenval ...
1008 : !> \param num_integ_points ...
1009 : !> \param num_integ_group ...
1010 : !> \param color_rpa_group ...
1011 : !> \param fm_matrix_PQ ...
1012 : !> \param fm_mat_S ...
1013 : !> \param fm_mat_Q_gemm ...
1014 : !> \param fm_mat_Q ...
1015 : !> \param fm_mat_S_gw ...
1016 : !> \param fm_mat_R_gw ...
1017 : !> \param fm_mat_S_ij_bse ...
1018 : !> \param fm_mat_S_ab_bse ...
1019 : !> \param my_do_gw ...
1020 : !> \param do_bse ...
1021 : !> \param gw_corr_lev_occ ...
1022 : !> \param gw_corr_lev_virt ...
1023 : !> \param bse_lev_virt ...
1024 : !> \param do_minimax_quad ...
1025 : !> \param do_im_time ...
1026 : !> \param mo_coeff ...
1027 : !> \param fm_matrix_L_kpoints ...
1028 : !> \param fm_matrix_Minv_L_kpoints ...
1029 : !> \param fm_matrix_Minv ...
1030 : !> \param fm_matrix_Minv_Vtrunc_Minv ...
1031 : !> \param mat_munu ...
1032 : !> \param mat_P_global ...
1033 : !> \param t_3c_M ...
1034 : !> \param t_3c_O ...
1035 : !> \param t_3c_O_compressed ...
1036 : !> \param t_3c_O_ind ...
1037 : !> \param starts_array_mc ...
1038 : !> \param ends_array_mc ...
1039 : !> \param starts_array_mc_block ...
1040 : !> \param ends_array_mc_block ...
1041 : !> \param matrix_s ...
1042 : !> \param do_kpoints_from_Gamma ...
1043 : !> \param kpoints ...
1044 : !> \param gd_array ...
1045 : !> \param color_sub ...
1046 : !> \param do_ri_sos_laplace_mp2 ...
1047 : !> \param calc_forces ...
1048 : ! **************************************************************************************************
1049 296 : SUBROUTINE rpa_num_int(qs_env, Erpa, mp2_env, para_env, para_env_RPA, para_env_sub, unit_nr, &
1050 296 : homo, virtual, dimen_RI, dimen_RI_red, dimen_ia, dimen_nm_gw, &
1051 : Eigenval, num_integ_points, num_integ_group, color_rpa_group, &
1052 592 : fm_matrix_PQ, fm_mat_S, fm_mat_Q_gemm, fm_mat_Q, fm_mat_S_gw, fm_mat_R_gw, &
1053 : fm_mat_S_ij_bse, fm_mat_S_ab_bse, &
1054 296 : my_do_gw, do_bse, gw_corr_lev_occ, gw_corr_lev_virt, &
1055 : bse_lev_virt, &
1056 296 : do_minimax_quad, do_im_time, mo_coeff, &
1057 : fm_matrix_L_kpoints, fm_matrix_Minv_L_kpoints, &
1058 : fm_matrix_Minv, fm_matrix_Minv_Vtrunc_Minv, mat_munu, mat_P_global, &
1059 : t_3c_M, t_3c_O, t_3c_O_compressed, t_3c_O_ind, &
1060 : starts_array_mc, ends_array_mc, &
1061 : starts_array_mc_block, ends_array_mc_block, &
1062 : matrix_s, do_kpoints_from_Gamma, kpoints, gd_array, color_sub, &
1063 : do_ri_sos_laplace_mp2, calc_forces)
1064 :
1065 : TYPE(qs_environment_type), POINTER :: qs_env
1066 : REAL(KIND=dp), INTENT(OUT) :: Erpa
1067 : TYPE(mp2_type) :: mp2_env
1068 : TYPE(mp_para_env_type), POINTER :: para_env, para_env_RPA, para_env_sub
1069 : INTEGER, INTENT(IN) :: unit_nr
1070 : INTEGER, DIMENSION(:), INTENT(IN) :: homo, virtual
1071 : INTEGER, INTENT(IN) :: dimen_RI, dimen_RI_red
1072 : INTEGER, DIMENSION(:), INTENT(IN) :: dimen_ia
1073 : INTEGER, INTENT(IN) :: dimen_nm_gw
1074 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :, :), &
1075 : INTENT(INOUT) :: Eigenval
1076 : INTEGER, INTENT(IN) :: num_integ_points, num_integ_group, &
1077 : color_rpa_group
1078 : TYPE(cp_fm_type), INTENT(IN) :: fm_matrix_PQ
1079 : TYPE(cp_fm_type), DIMENSION(:), INTENT(INOUT) :: fm_mat_S
1080 : TYPE(cp_fm_type), DIMENSION(:), INTENT(IN) :: fm_mat_Q_gemm, fm_mat_Q, fm_mat_S_gw
1081 : TYPE(cp_fm_type), INTENT(IN) :: fm_mat_R_gw, fm_mat_S_ij_bse, &
1082 : fm_mat_S_ab_bse
1083 : LOGICAL, INTENT(IN) :: my_do_gw, do_bse
1084 : INTEGER, DIMENSION(:), INTENT(IN) :: gw_corr_lev_occ, gw_corr_lev_virt
1085 : INTEGER, INTENT(IN) :: bse_lev_virt
1086 : LOGICAL, INTENT(IN) :: do_minimax_quad, do_im_time
1087 : TYPE(cp_fm_type), DIMENSION(:), INTENT(IN) :: mo_coeff
1088 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:, :) :: fm_matrix_L_kpoints, &
1089 : fm_matrix_Minv_L_kpoints, &
1090 : fm_matrix_Minv, &
1091 : fm_matrix_Minv_Vtrunc_Minv
1092 : TYPE(dbcsr_p_type), INTENT(IN) :: mat_munu
1093 : TYPE(dbcsr_p_type), INTENT(INOUT) :: mat_P_global
1094 : TYPE(dbt_type), INTENT(INOUT) :: t_3c_M
1095 : TYPE(dbt_type), ALLOCATABLE, DIMENSION(:, :), &
1096 : INTENT(INOUT) :: t_3c_O
1097 : TYPE(hfx_compression_type), ALLOCATABLE, &
1098 : DIMENSION(:, :, :), INTENT(INOUT) :: t_3c_O_compressed
1099 : TYPE(block_ind_type), ALLOCATABLE, &
1100 : DIMENSION(:, :, :), INTENT(INOUT) :: t_3c_O_ind
1101 : INTEGER, ALLOCATABLE, DIMENSION(:), INTENT(IN) :: starts_array_mc, ends_array_mc, &
1102 : starts_array_mc_block, &
1103 : ends_array_mc_block
1104 : TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: matrix_s
1105 : LOGICAL :: do_kpoints_from_Gamma
1106 : TYPE(kpoint_type), POINTER :: kpoints
1107 : TYPE(group_dist_d1_type), INTENT(IN) :: gd_array
1108 : INTEGER, INTENT(IN) :: color_sub
1109 : LOGICAL, INTENT(IN) :: do_ri_sos_laplace_mp2, calc_forces
1110 :
1111 : CHARACTER(LEN=*), PARAMETER :: routineN = 'rpa_num_int'
1112 :
1113 : COMPLEX(KIND=dp), ALLOCATABLE, &
1114 296 : DIMENSION(:, :, :, :) :: vec_Sigma_c_gw
1115 : INTEGER :: count_ev_sc_GW, cut_memory, group_size_P, gw_corr_lev_tot, handle, handle3, i, &
1116 : ikp_local, ispin, iter_evGW, iter_sc_GW0, j, jquad, min_bsize, mm_style, nkp, &
1117 : nkp_self_energy, nmo, nspins, num_3c_repl, num_cells_dm, num_fit_points, Pspin, Qspin, &
1118 : size_P
1119 : INTEGER(int_8) :: dbcsr_nflop
1120 296 : INTEGER, ALLOCATABLE, DIMENSION(:, :) :: index_to_cell_3c
1121 296 : INTEGER, ALLOCATABLE, DIMENSION(:, :, :) :: cell_to_index_3c
1122 592 : INTEGER, DIMENSION(:), POINTER :: col_blk_size, prim_blk_sizes, &
1123 296 : RI_blk_sizes
1124 : LOGICAL :: do_apply_ic_corr_to_gw, do_gw_im_time, do_ic_model, do_kpoints_cubic_RPA, &
1125 : do_periodic, do_print, do_ri_Sigma_x, exit_ev_gw, first_cycle, &
1126 : first_cycle_periodic_correction, my_open_shell, print_ic_values
1127 296 : LOGICAL, ALLOCATABLE, DIMENSION(:, :, :, :, :) :: has_mat_P_blocks
1128 : REAL(KIND=dp) :: a_scaling, alpha, dbcsr_time, e_exchange, e_exchange_corr, eps_filter, &
1129 : eps_filter_im_time, ext_scaling, fermi_level_offset, fermi_level_offset_input, &
1130 : my_flop_rate, omega, omega_max_fit, omega_old, tau, tau_old
1131 592 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:) :: delta_corr, e_fermi, tau_tj, tau_wj, tj, &
1132 296 : trace_Qomega, vec_omega_fit_gw, wj, &
1133 296 : wkp_W
1134 296 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :) :: vec_W_gw, weights_cos_tf_t_to_w, &
1135 296 : weights_cos_tf_w_to_t, &
1136 296 : weights_sin_tf_t_to_w
1137 296 : REAL(KIND=dp), ALLOCATABLE, DIMENSION(:, :, :) :: Eigenval_last, Eigenval_scf, &
1138 296 : vec_Sigma_x_gw
1139 : TYPE(cp_cfm_type) :: cfm_mat_Q
1140 : TYPE(cp_fm_type) :: fm_mat_Q_static_bse_gemm, fm_mat_RI_global_work, fm_mat_S_ia_bse, &
1141 : fm_mat_work, fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, fm_scaled_dm_occ_tau, &
1142 : fm_scaled_dm_virt_tau
1143 296 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:) :: fm_mat_S_gw_work, fm_mat_W, &
1144 296 : fm_mo_coeff_occ, fm_mo_coeff_virt
1145 296 : TYPE(cp_fm_type), ALLOCATABLE, DIMENSION(:, :) :: fm_mat_L_kpoints, fm_mat_Minv_L_kpoints
1146 : TYPE(dbcsr_p_type) :: mat_dm, mat_L, mat_M_P_munu_occ, &
1147 : mat_M_P_munu_virt, mat_MinvVMinv
1148 : TYPE(dbcsr_p_type), ALLOCATABLE, &
1149 296 : DIMENSION(:, :, :) :: mat_P_omega
1150 296 : TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: matrix_berry_im_mo_mo, &
1151 296 : matrix_berry_re_mo_mo
1152 296 : TYPE(dbcsr_p_type), DIMENSION(:, :), POINTER :: mat_P_omega_kp
1153 : TYPE(dbcsr_type), POINTER :: mat_W, mat_work
1154 2072 : TYPE(dbt_type) :: t_3c_overl_int_ao_mo
1155 296 : TYPE(dbt_type), ALLOCATABLE, DIMENSION(:) :: t_3c_overl_int_gw_AO, &
1156 296 : t_3c_overl_int_gw_RI, &
1157 296 : t_3c_overl_nnP_ic, &
1158 296 : t_3c_overl_nnP_ic_reflected
1159 : TYPE(dgemm_counter_type) :: dgemm_counter
1160 : TYPE(hfx_compression_type), ALLOCATABLE, &
1161 296 : DIMENSION(:) :: t_3c_O_mo_compressed
1162 21904 : TYPE(im_time_force_type) :: force_data
1163 296 : TYPE(rpa_exchange_work_type) :: exchange_work
1164 1480 : TYPE(rpa_grad_type) :: rpa_grad
1165 296 : TYPE(rpa_sigma_type) :: rpa_sigma
1166 296 : TYPE(two_dim_int_array), ALLOCATABLE, DIMENSION(:) :: t_3c_O_mo_ind
1167 :
1168 296 : CALL timeset(routineN, handle)
1169 :
1170 296 : nspins = SIZE(homo)
1171 296 : nmo = homo(1) + virtual(1)
1172 :
1173 296 : my_open_shell = (nspins == 2)
1174 :
1175 296 : do_gw_im_time = my_do_gw .AND. do_im_time
1176 296 : do_ri_Sigma_x = mp2_env%ri_g0w0%do_ri_Sigma_x
1177 296 : do_ic_model = mp2_env%ri_g0w0%do_ic_model
1178 296 : print_ic_values = mp2_env%ri_g0w0%print_ic_values
1179 296 : do_periodic = mp2_env%ri_g0w0%do_periodic
1180 296 : do_kpoints_cubic_RPA = mp2_env%ri_rpa_im_time%do_im_time_kpoints
1181 :
1182 : ! For SOS-MP2 only gemm is implemented
1183 296 : mm_style = wfc_mm_style_gemm
1184 296 : IF (.NOT. do_ri_sos_laplace_mp2) mm_style = mp2_env%ri_rpa%mm_style
1185 :
1186 296 : IF (my_do_gw) THEN
1187 106 : ext_scaling = 0.2_dp
1188 106 : omega_max_fit = mp2_env%ri_g0w0%omega_max_fit
1189 106 : fermi_level_offset_input = mp2_env%ri_g0w0%fermi_level_offset
1190 106 : iter_evGW = mp2_env%ri_g0w0%iter_evGW
1191 106 : iter_sc_GW0 = mp2_env%ri_g0w0%iter_sc_GW0
1192 106 : IF ((.NOT. do_im_time)) THEN
1193 60 : IF (iter_sc_GW0 .NE. 1 .AND. iter_evGW .NE. 1) CPABORT("Mixed scGW0/evGW not implemented.")
1194 : ! in case of scGW0 with the N^4 algorithm, we use the evGW code but use the DFT eigenvalues for W
1195 60 : IF (iter_sc_GW0 .NE. 1) iter_evGW = iter_sc_GW0
1196 : END IF
1197 : ELSE
1198 190 : ext_scaling = 0.0_dp
1199 190 : iter_evGW = 1
1200 190 : iter_sc_GW0 = 1
1201 : END IF
1202 :
1203 296 : IF (do_kpoints_cubic_RPA .AND. do_ri_sos_laplace_mp2) THEN
1204 0 : CPABORT("RI-SOS-Laplace-MP2 with k-point-sampling is not implemented.")
1205 : END IF
1206 :
1207 296 : do_apply_ic_corr_to_gw = .FALSE.
1208 296 : IF (mp2_env%ri_g0w0%ic_corr_list(1)%array(1) > 0.0_dp) do_apply_ic_corr_to_gw = .TRUE.
1209 :
1210 296 : IF (do_im_time) THEN
1211 134 : CPASSERT(do_minimax_quad .OR. do_ri_sos_laplace_mp2)
1212 :
1213 134 : group_size_P = mp2_env%ri_rpa_im_time%group_size_P
1214 134 : cut_memory = mp2_env%ri_rpa_im_time%cut_memory
1215 134 : eps_filter = mp2_env%ri_rpa_im_time%eps_filter
1216 : eps_filter_im_time = mp2_env%ri_rpa_im_time%eps_filter* &
1217 134 : mp2_env%ri_rpa_im_time%eps_filter_factor
1218 :
1219 134 : min_bsize = mp2_env%ri_rpa_im_time%min_bsize
1220 :
1221 : CALL alloc_im_time(qs_env, para_env, dimen_RI, dimen_RI_red, &
1222 : num_integ_points, nspins, fm_mat_Q(1), fm_mo_coeff_occ, fm_mo_coeff_virt, &
1223 : fm_matrix_Minv_L_kpoints, fm_matrix_L_kpoints, mat_P_global, &
1224 : t_3c_O, matrix_s, kpoints, eps_filter_im_time, &
1225 : cut_memory, nkp, num_cells_dm, num_3c_repl, &
1226 : size_P, ikp_local, &
1227 : index_to_cell_3c, &
1228 : cell_to_index_3c, &
1229 : col_blk_size, &
1230 : do_ic_model, do_kpoints_cubic_RPA, &
1231 : do_kpoints_from_Gamma, do_ri_Sigma_x, my_open_shell, &
1232 : has_mat_P_blocks, wkp_W, &
1233 : cfm_mat_Q, fm_mat_Minv_L_kpoints, fm_mat_L_kpoints, &
1234 : fm_mat_RI_global_work, fm_mat_work, fm_mo_coeff_occ_scaled, &
1235 : fm_mo_coeff_virt_scaled, mat_dm, mat_L, mat_M_P_munu_occ, mat_M_P_munu_virt, &
1236 : mat_MinvVMinv, mat_P_omega, mat_P_omega_kp, mat_work, mo_coeff, &
1237 134 : fm_scaled_dm_occ_tau, fm_scaled_dm_virt_tau, homo, nmo)
1238 :
1239 134 : IF (calc_forces) CALL init_im_time_forces(force_data, fm_matrix_PQ, t_3c_M, unit_nr, mp2_env, qs_env)
1240 :
1241 134 : IF (my_do_gw) THEN
1242 :
1243 : CALL dbcsr_get_info(mat_P_global%matrix, &
1244 46 : row_blk_size=RI_blk_sizes)
1245 :
1246 : CALL dbcsr_get_info(matrix_s(1)%matrix, &
1247 46 : row_blk_size=prim_blk_sizes)
1248 :
1249 46 : gw_corr_lev_tot = gw_corr_lev_occ(1) + gw_corr_lev_virt(1)
1250 :
1251 46 : IF (.NOT. do_kpoints_cubic_RPA) THEN
1252 : CALL allocate_matrices_gw_im_time(gw_corr_lev_occ, gw_corr_lev_virt, homo, nmo, &
1253 : num_integ_points, unit_nr, &
1254 : RI_blk_sizes, do_ic_model, &
1255 : para_env, fm_mat_W, fm_mat_Q(1), &
1256 : mo_coeff, &
1257 : t_3c_overl_int_ao_mo, t_3c_O_mo_compressed, t_3c_O_mo_ind, &
1258 : t_3c_overl_int_gw_RI, t_3c_overl_int_gw_AO, &
1259 : starts_array_mc, ends_array_mc, &
1260 : t_3c_overl_nnP_ic, t_3c_overl_nnP_ic_reflected, &
1261 : matrix_s, mat_W, t_3c_O, &
1262 : t_3c_O_compressed, t_3c_O_ind, &
1263 46 : qs_env)
1264 :
1265 : END IF
1266 : END IF
1267 :
1268 : END IF
1269 296 : IF (do_ic_model) THEN
1270 : ! image charge model only implemented for cubic scaling GW
1271 2 : CPASSERT(do_gw_im_time)
1272 2 : CPASSERT(.NOT. do_periodic)
1273 2 : IF (cut_memory .NE. 1) CPABORT("For IC, use MEMORY_CUT 1 in the LOW_SCALING section.")
1274 : END IF
1275 :
1276 888 : ALLOCATE (e_fermi(nspins))
1277 296 : IF (do_minimax_quad .OR. do_ri_sos_laplace_mp2) THEN
1278 200 : do_print = .NOT. do_ic_model
1279 : CALL get_minimax_grid(para_env, unit_nr, homo, Eigenval, num_integ_points, do_im_time, &
1280 : do_ri_sos_laplace_mp2, do_print, &
1281 : tau_tj, tau_wj, qs_env, do_gw_im_time, &
1282 : do_kpoints_cubic_RPA, e_fermi(1), tj, wj, &
1283 : weights_cos_tf_t_to_w, weights_cos_tf_w_to_t, weights_sin_tf_t_to_w, &
1284 200 : qs_env%mp2_env%ri_g0w0%regularization_minimax)
1285 : !For sos_laplace_mp2 and low-scaling RPA, potentially need to store/retrieve the initial weights
1286 200 : IF (qs_env%mp2_env%ri_rpa_im_time%keep_quad) THEN
1287 : CALL keep_initial_quad(tj, wj, tau_tj, tau_wj, weights_cos_tf_t_to_w, &
1288 : weights_cos_tf_w_to_t, do_ri_sos_laplace_mp2, do_im_time, &
1289 200 : num_integ_points, unit_nr, qs_env)
1290 : END IF
1291 : ELSE
1292 96 : IF (calc_forces) CPABORT("Forces with Clenshaw-Curtis grid not implemented.")
1293 : CALL get_clenshaw_grid(para_env, para_env_RPA, unit_nr, homo, virtual, Eigenval, num_integ_points, &
1294 : num_integ_group, color_rpa_group, fm_mat_S, my_do_gw, &
1295 96 : ext_scaling, a_scaling, tj, wj)
1296 : END IF
1297 :
1298 : ! This array is needed for RPA
1299 296 : IF (.NOT. do_ri_sos_laplace_mp2) THEN
1300 714 : ALLOCATE (trace_Qomega(dimen_RI_red))
1301 : END IF
1302 :
1303 296 : IF (do_ri_sos_laplace_mp2 .AND. .NOT. do_im_time) THEN
1304 28 : alpha = 1.0_dp
1305 268 : ELSE IF (my_open_shell .OR. do_ri_sos_laplace_mp2) THEN
1306 72 : alpha = 2.0_dp
1307 : ELSE
1308 196 : alpha = 4.0_dp
1309 : END IF
1310 296 : IF (my_do_gw) THEN
1311 : CALL allocate_matrices_gw(vec_Sigma_c_gw, color_rpa_group, dimen_nm_gw, &
1312 : gw_corr_lev_occ, gw_corr_lev_virt, homo, &
1313 : nmo, num_integ_group, num_integ_points, unit_nr, &
1314 : gw_corr_lev_tot, num_fit_points, omega_max_fit, &
1315 : do_minimax_quad, do_periodic, do_ri_Sigma_x,.NOT. do_im_time, &
1316 : first_cycle_periodic_correction, &
1317 : a_scaling, Eigenval, tj, vec_omega_fit_gw, vec_Sigma_x_gw, &
1318 : delta_corr, Eigenval_last, Eigenval_scf, vec_W_gw, &
1319 : fm_mat_S_gw, fm_mat_S_gw_work, &
1320 : para_env, mp2_env, kpoints, nkp, nkp_self_energy, &
1321 106 : do_kpoints_cubic_RPA, do_kpoints_from_Gamma)
1322 :
1323 106 : IF (do_bse) THEN
1324 :
1325 32 : CALL cp_fm_create(fm_mat_Q_static_bse_gemm, fm_mat_Q_gemm(1)%matrix_struct)
1326 32 : CALL cp_fm_to_fm(fm_mat_Q_gemm(1), fm_mat_Q_static_bse_gemm)
1327 32 : CALL cp_fm_set_all(fm_mat_Q_static_bse_gemm, 0.0_dp)
1328 :
1329 : END IF
1330 :
1331 : END IF
1332 :
1333 296 : IF (calc_forces .AND. .NOT. do_im_time) CALL rpa_grad_create(rpa_grad, fm_mat_Q(1), &
1334 : fm_mat_S, homo, virtual, mp2_env, Eigenval(:, 1, :), &
1335 40 : unit_nr, do_ri_sos_laplace_mp2)
1336 : !doubtful
1337 296 : IF (.NOT. do_im_time .AND. .NOT. do_ri_sos_laplace_mp2) &
1338 : CALL exchange_work%create(qs_env, para_env_sub, mat_munu, dimen_RI_red, &
1339 134 : fm_mat_S, fm_mat_Q(1), fm_mat_Q_gemm(1), homo, virtual)
1340 296 : Erpa = 0.0_dp
1341 296 : IF (mp2_env%ri_rpa%exchange_correction /= rpa_exchange_none) e_exchange = 0.0_dp
1342 296 : first_cycle = .TRUE.
1343 296 : omega_old = 0.0_dp
1344 296 : CALL dgemm_counter_init(dgemm_counter, unit_nr, mp2_env%ri_rpa%print_dgemm_info)
1345 :
1346 702 : DO count_ev_sc_GW = 1, iter_evGW
1347 426 : dbcsr_time = 0.0_dp
1348 426 : dbcsr_nflop = 0
1349 :
1350 426 : IF (do_ic_model) CYCLE
1351 :
1352 : ! reset some values, important when doing eigenvalue self-consistent GW
1353 424 : IF (my_do_gw) THEN
1354 234 : Erpa = 0.0_dp
1355 158320 : vec_Sigma_c_gw = z_zero
1356 234 : first_cycle = .TRUE.
1357 : END IF
1358 :
1359 : ! calculate Q_PQ(it)
1360 424 : IF (do_im_time) THEN ! not using Imaginary time
1361 :
1362 146 : IF (.NOT. do_kpoints_cubic_RPA) THEN
1363 314 : DO ispin = 1, nspins
1364 314 : e_fermi(ispin) = (Eigenval(homo(ispin), 1, ispin) + Eigenval(homo(ispin) + 1, 1, ispin))*0.5_dp
1365 : END DO
1366 : END IF
1367 :
1368 146 : tau = 0.0_dp
1369 146 : tau_old = 0.0_dp
1370 :
1371 146 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(/T3,A,T66,i15)") &
1372 73 : "MEMORY_INFO| Memory cut:", cut_memory
1373 146 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(T3,A,T66,ES15.2)") &
1374 73 : "SPARSITY_INFO| Eps filter for M virt/occ tensors:", eps_filter
1375 146 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(T3,A,T66,ES15.2)") &
1376 73 : "SPARSITY_INFO| Eps filter for P matrix:", eps_filter_im_time
1377 146 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(T3,A,T66,i15)") &
1378 73 : "SPARSITY_INFO| Minimum tensor block size:", min_bsize
1379 :
1380 : ! for evGW, we have to ensure that mat_P_omega is zero
1381 146 : CALL zero_mat_P_omega(mat_P_omega(:, :, 1))
1382 :
1383 : ! compute the matrix Q(it) and Fourier transform it directly to mat_P_omega(iw)
1384 : CALL compute_mat_P_omega(mat_P_omega(:, :, 1), fm_scaled_dm_occ_tau, &
1385 : fm_scaled_dm_virt_tau, fm_mo_coeff_occ(1), fm_mo_coeff_virt(1), &
1386 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, &
1387 : mat_P_global, matrix_s, 1, &
1388 : t_3c_M, t_3c_O, t_3c_O_compressed, t_3c_O_ind, &
1389 : starts_array_mc, ends_array_mc, &
1390 : starts_array_mc_block, ends_array_mc_block, &
1391 : weights_cos_tf_t_to_w, tj, tau_tj, e_fermi(1), eps_filter, alpha, &
1392 : eps_filter_im_time, Eigenval(:, 1, 1), nmo, &
1393 : num_integ_points, cut_memory, &
1394 : unit_nr, mp2_env, para_env, &
1395 : qs_env, do_kpoints_from_Gamma, &
1396 : index_to_cell_3c, cell_to_index_3c, &
1397 : has_mat_P_blocks, do_ri_sos_laplace_mp2, &
1398 146 : dbcsr_time, dbcsr_nflop)
1399 :
1400 : ! the same for open shell, use fm_mo_coeff_occ_beta and fm_mo_coeff_virt_beta
1401 146 : IF (my_open_shell) THEN
1402 30 : CALL zero_mat_P_omega(mat_P_omega(:, :, 2))
1403 : CALL compute_mat_P_omega(mat_P_omega(:, :, 2), fm_scaled_dm_occ_tau, &
1404 : fm_scaled_dm_virt_tau, fm_mo_coeff_occ(2), &
1405 : fm_mo_coeff_virt(2), &
1406 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, &
1407 : mat_P_global, matrix_s, 2, &
1408 : t_3c_M, t_3c_O, t_3c_O_compressed, t_3c_O_ind, &
1409 : starts_array_mc, ends_array_mc, &
1410 : starts_array_mc_block, ends_array_mc_block, &
1411 : weights_cos_tf_t_to_w, tj, tau_tj, e_fermi(2), eps_filter, alpha, &
1412 : eps_filter_im_time, Eigenval(:, 1, 2), nmo, &
1413 : num_integ_points, cut_memory, &
1414 : unit_nr, mp2_env, para_env, &
1415 : qs_env, do_kpoints_from_Gamma, &
1416 : index_to_cell_3c, cell_to_index_3c, &
1417 : has_mat_P_blocks, do_ri_sos_laplace_mp2, &
1418 30 : dbcsr_time, dbcsr_nflop)
1419 :
1420 : !For RPA, we sum up the P matrices. If no force needed, can clean-up the beta spin one
1421 30 : IF (.NOT. do_ri_sos_laplace_mp2) THEN
1422 144 : DO j = 1, SIZE(mat_P_omega, 2)
1423 928 : DO i = 1, SIZE(mat_P_omega, 1)
1424 784 : CALL dbcsr_add(mat_P_omega(i, j, 1)%matrix, mat_P_omega(i, j, 2)%matrix, 1.0_dp, 1.0_dp)
1425 906 : IF (.NOT. calc_forces) CALL dbcsr_clear(mat_P_omega(i, j, 2)%matrix)
1426 : END DO
1427 : END DO
1428 : END IF
1429 : END IF ! my_open_shell
1430 :
1431 : END IF ! do im time
1432 :
1433 424 : IF (mp2_env%ri_rpa%sigma_param /= sigma_none) THEN
1434 10 : CALL rpa_sigma_create(rpa_sigma, mp2_env%ri_rpa%sigma_param, fm_mat_Q(1), unit_nr, para_env)
1435 : END IF
1436 :
1437 12626 : DO jquad = 1, num_integ_points
1438 12202 : IF (MODULO(jquad, num_integ_group) /= color_rpa_group) CYCLE
1439 :
1440 11562 : CALL timeset(routineN//"_RPA_matrix_operations", handle3)
1441 :
1442 11562 : IF (do_ri_sos_laplace_mp2) THEN
1443 206 : omega = tau_tj(jquad)
1444 : ELSE
1445 11356 : IF (do_minimax_quad) THEN
1446 946 : omega = tj(jquad)
1447 : ELSE
1448 10410 : omega = a_scaling/TAN(tj(jquad))
1449 : END IF
1450 : END IF ! do_ri_sos_laplace_mp2
1451 :
1452 11562 : IF (do_im_time) THEN
1453 : ! in case we do imag time, we already calculated fm_mat_Q by a Fourier transform from im. time
1454 :
1455 928 : IF (.NOT. (do_kpoints_cubic_RPA .OR. do_kpoints_from_Gamma)) THEN
1456 :
1457 1784 : DO ispin = 1, SIZE(mat_P_omega, 3)
1458 : CALL contract_P_omega_with_mat_L(mat_P_omega(jquad, 1, ispin)%matrix, mat_L%matrix, mat_work, &
1459 : eps_filter_im_time, fm_mat_work, dimen_RI, dimen_RI_red, &
1460 1784 : fm_mat_Minv_L_kpoints(1, 1), fm_mat_Q(ispin))
1461 : END DO
1462 : END IF
1463 :
1464 : ELSE
1465 10634 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(T3, A, 1X, I3, 1X, A, 1X, I3)") &
1466 5317 : "INTEG_INFO| Started with Integration point", jquad, "of", num_integ_points
1467 :
1468 10634 : IF (first_cycle .AND. count_ev_sc_gw > 1) THEN
1469 116 : IF (iter_sc_gw0 == 1) THEN
1470 156 : DO ispin = 1, nspins
1471 : CALL remove_scaling_factor_rpa(fm_mat_S(ispin), virtual(ispin), &
1472 156 : Eigenval_last(:, 1, ispin), homo(ispin), omega_old)
1473 : END DO
1474 : ELSE
1475 84 : DO ispin = 1, nspins
1476 : CALL remove_scaling_factor_rpa(fm_mat_S(ispin), virtual(ispin), &
1477 84 : Eigenval_scf(:, 1, ispin), homo(ispin), omega_old)
1478 : END DO
1479 : END IF
1480 : END IF
1481 :
1482 10634 : IF (iter_sc_GW0 > 1) THEN
1483 8140 : DO ispin = 1, nspins
1484 : CALL calc_mat_Q(fm_mat_S(ispin), do_ri_sos_laplace_mp2, first_cycle, virtual(ispin), &
1485 : Eigenval_scf(:, 1, ispin), homo(ispin), omega, omega_old, jquad, mm_style, &
1486 : dimen_RI_red, dimen_ia(ispin), alpha, fm_mat_Q(ispin), &
1487 : fm_mat_Q_gemm(ispin), do_bse, fm_mat_Q_static_bse_gemm, dgemm_counter, &
1488 8140 : num_integ_points, count_ev_sc_GW)
1489 : END DO
1490 :
1491 : ! For SOS-MP2 we need both matrices separately
1492 4070 : IF (.NOT. do_ri_sos_laplace_mp2) THEN
1493 4070 : DO ispin = 2, nspins
1494 4070 : CALL cp_fm_scale_and_add(alpha=1.0_dp, matrix_a=fm_mat_Q(1), beta=1.0_dp, matrix_b=fm_mat_Q(ispin))
1495 : END DO
1496 : END IF
1497 : ELSE
1498 13292 : DO ispin = 1, nspins
1499 : CALL calc_mat_Q(fm_mat_S(ispin), do_ri_sos_laplace_mp2, first_cycle, virtual(ispin), &
1500 : Eigenval(:, 1, ispin), homo(ispin), omega, omega_old, jquad, mm_style, &
1501 : dimen_RI_red, dimen_ia(ispin), alpha, fm_mat_Q(ispin), &
1502 : fm_mat_Q_gemm(ispin), do_bse, fm_mat_Q_static_bse_gemm, dgemm_counter, &
1503 13292 : num_integ_points, count_ev_sc_GW)
1504 : END DO
1505 :
1506 : ! For SOS-MP2 we need both matrices separately
1507 6564 : IF (.NOT. do_ri_sos_laplace_mp2) THEN
1508 6600 : DO ispin = 2, nspins
1509 6600 : CALL cp_fm_scale_and_add(alpha=1.0_dp, matrix_a=fm_mat_Q(1), beta=1.0_dp, matrix_b=fm_mat_Q(ispin))
1510 : END DO
1511 : END IF
1512 :
1513 : END IF
1514 :
1515 : END IF ! im time
1516 :
1517 : ! Calculate RPA exchange energy correction
1518 11562 : IF (mp2_env%ri_rpa%exchange_correction /= rpa_exchange_none) THEN
1519 12 : e_exchange_corr = 0.0_dp
1520 12 : CALL exchange_work%compute(fm_mat_Q(1), Eigenval(:, 1, :), fm_mat_S, omega, e_exchange_corr, mp2_env)
1521 :
1522 : ! Evaluate the final exchange energy correction
1523 12 : e_exchange = e_exchange + e_exchange_corr*wj(jquad)
1524 : END IF
1525 :
1526 : ! for developing Sigma functional closed and open shell are taken cared for
1527 11562 : IF (mp2_env%ri_rpa%sigma_param /= sigma_none) THEN
1528 30 : CALL rpa_sigma_matrix_spectral(rpa_sigma, fm_mat_Q(1), wj(jquad), para_env_RPA)
1529 : END IF
1530 :
1531 11562 : IF (do_ri_sos_laplace_mp2) THEN
1532 :
1533 206 : CALL SOS_MP2_postprocessing(fm_mat_Q, Erpa, tau_wj(jquad))
1534 :
1535 206 : IF (calc_forces .AND. .NOT. do_im_time) CALL rpa_grad_matrix_operations(mp2_env, rpa_grad, do_ri_sos_laplace_mp2, &
1536 : fm_mat_Q, fm_mat_Q_gemm, dgemm_counter, fm_mat_S, omega, homo, virtual, &
1537 60 : Eigenval(:, 1, :), tau_wj(jquad), unit_nr)
1538 : ELSE
1539 11356 : IF (calc_forces .AND. .NOT. do_im_time) CALL rpa_grad_copy_Q(fm_mat_Q(1), rpa_grad)
1540 :
1541 11356 : CALL Q_trace_and_add_unit_matrix(dimen_RI_red, trace_Qomega, fm_mat_Q(1))
1542 :
1543 11356 : IF (do_kpoints_cubic_RPA .OR. do_kpoints_from_Gamma) THEN
1544 : CALL invert_eps_compute_W_and_Erpa_kp(dimen_RI, num_integ_points, jquad, nkp, count_ev_sc_GW, para_env, &
1545 : Erpa, tau_tj, tj, wj, weights_cos_tf_w_to_t, &
1546 : wkp_W, do_gw_im_time, do_ri_Sigma_x, do_kpoints_from_Gamma, &
1547 : cfm_mat_Q, ikp_local, &
1548 : mat_P_omega(:, :, 1), mat_P_omega_kp, qs_env, eps_filter_im_time, unit_nr, &
1549 : kpoints, fm_mat_Minv_L_kpoints, fm_mat_L_kpoints, &
1550 : fm_mat_W, fm_mat_RI_global_work, mat_MinvVMinv, &
1551 132 : fm_matrix_Minv, fm_matrix_Minv_Vtrunc_Minv)
1552 : ELSE
1553 11224 : CALL compute_Erpa_by_freq_int(dimen_RI_red, trace_Qomega, fm_mat_Q(1), para_env_RPA, Erpa, wj(jquad))
1554 : END IF
1555 :
1556 11356 : IF (calc_forces .AND. .NOT. do_im_time) CALL rpa_grad_matrix_operations(mp2_env, rpa_grad, do_ri_sos_laplace_mp2, &
1557 : fm_mat_Q, fm_mat_Q_gemm, dgemm_counter, fm_mat_S, omega, homo, virtual, &
1558 48 : Eigenval(:, 1, :), wj(jquad), unit_nr)
1559 : END IF ! do_ri_sos_laplace_mp2
1560 :
1561 : ! save omega and reset the first_cycle flag
1562 11562 : first_cycle = .FALSE.
1563 11562 : omega_old = omega
1564 :
1565 11562 : CALL timestop(handle3)
1566 :
1567 11562 : IF (my_do_gw) THEN
1568 :
1569 10788 : CALL get_fermi_level_offset(fermi_level_offset, fermi_level_offset_input, Eigenval(:, 1, :), homo)
1570 :
1571 : ! do_im_time = TRUE means low-scaling calculation
1572 10788 : IF (do_im_time) THEN
1573 : ! only for molecules
1574 578 : IF (.NOT. (do_kpoints_cubic_RPA .OR. do_kpoints_from_Gamma)) THEN
1575 : CALL compute_W_cubic_GW(fm_mat_W, fm_mat_Q(1), fm_mat_work, dimen_RI, fm_mat_Minv_L_kpoints, num_integ_points, &
1576 470 : tj, tau_tj, weights_cos_tf_w_to_t, jquad, omega)
1577 : END IF
1578 : ELSE
1579 : CALL compute_GW_self_energy(vec_Sigma_c_gw, dimen_nm_gw, dimen_RI_red, gw_corr_lev_occ, &
1580 : gw_corr_lev_virt, homo, jquad, nmo, num_fit_points, &
1581 : do_im_time, do_periodic, first_cycle_periodic_correction, &
1582 : fermi_level_offset, &
1583 : omega, Eigenval(:, 1, :), delta_corr, vec_omega_fit_gw, vec_W_gw, wj, &
1584 : fm_mat_Q(1), fm_mat_R_gw, fm_mat_S_gw, &
1585 : fm_mat_S_gw_work, mo_coeff(1), para_env, &
1586 : para_env_RPA, matrix_berry_im_mo_mo, matrix_berry_re_mo_mo, &
1587 10210 : kpoints, qs_env, mp2_env)
1588 : END IF
1589 : END IF
1590 :
1591 11562 : IF (unit_nr > 0) CALL m_flush(unit_nr)
1592 24188 : CALL para_env%sync() ! sync to see output
1593 :
1594 : END DO ! jquad
1595 :
1596 424 : IF (mp2_env%ri_rpa%sigma_param /= sigma_none) THEN
1597 10 : CALL finalize_rpa_sigma(rpa_sigma, unit_nr, mp2_env%ri_rpa%e_sigma_corr, para_env, do_minimax_quad)
1598 10 : IF (do_minimax_quad) mp2_env%ri_rpa%e_sigma_corr = mp2_env%ri_rpa%e_sigma_corr/2.0_dp
1599 10 : CALL para_env%sum(mp2_env%ri_rpa%e_sigma_corr)
1600 : END IF
1601 :
1602 424 : CALL para_env%sum(Erpa)
1603 :
1604 424 : IF (.NOT. (do_ri_sos_laplace_mp2)) THEN
1605 366 : Erpa = Erpa/(pi*2.0_dp)
1606 366 : IF (do_minimax_quad) Erpa = Erpa/2.0_dp
1607 : END IF
1608 :
1609 424 : IF (mp2_env%ri_rpa%exchange_correction /= rpa_exchange_none) THEN
1610 12 : CALL para_env%sum(E_exchange)
1611 12 : E_exchange = E_exchange/(pi*2.0_dp)
1612 12 : IF (do_minimax_quad) E_exchange = E_exchange/2.0_dp
1613 12 : mp2_env%ri_rpa%ener_exchange = E_exchange
1614 : END IF
1615 :
1616 424 : IF (calc_forces .AND. do_ri_sos_laplace_mp2 .AND. do_im_time) THEN
1617 22 : IF (my_open_shell) THEN
1618 4 : Pspin = 1
1619 4 : Qspin = 2
1620 : CALL calc_laplace_loop_forces(force_data, mat_P_omega(:, 1, :), t_3c_M, t_3c_O(1, 1), &
1621 : t_3c_O_compressed(1, 1, :), t_3c_O_ind(1, 1, :), fm_scaled_dm_occ_tau, &
1622 : fm_scaled_dm_virt_tau, fm_mo_coeff_occ, fm_mo_coeff_virt, &
1623 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, &
1624 : starts_array_mc, ends_array_mc, starts_array_mc_block, &
1625 : ends_array_mc_block, num_integ_points, nmo, Eigenval(:, 1, :), &
1626 : tau_tj, tau_wj, cut_memory, Pspin, Qspin, my_open_shell, &
1627 4 : unit_nr, dbcsr_time, dbcsr_nflop, mp2_env, qs_env)
1628 4 : Pspin = 2
1629 4 : Qspin = 1
1630 : CALL calc_laplace_loop_forces(force_data, mat_P_omega(:, 1, :), t_3c_M, t_3c_O(1, 1), &
1631 : t_3c_O_compressed(1, 1, :), t_3c_O_ind(1, 1, :), fm_scaled_dm_occ_tau, &
1632 : fm_scaled_dm_virt_tau, fm_mo_coeff_occ, fm_mo_coeff_virt, &
1633 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, &
1634 : starts_array_mc, ends_array_mc, starts_array_mc_block, &
1635 : ends_array_mc_block, num_integ_points, nmo, Eigenval(:, 1, :), &
1636 : tau_tj, tau_wj, cut_memory, Pspin, Qspin, my_open_shell, &
1637 4 : unit_nr, dbcsr_time, dbcsr_nflop, mp2_env, qs_env)
1638 :
1639 : ELSE
1640 18 : Pspin = 1
1641 18 : Qspin = 1
1642 : CALL calc_laplace_loop_forces(force_data, mat_P_omega(:, 1, :), t_3c_M, t_3c_O(1, 1), &
1643 : t_3c_O_compressed(1, 1, :), t_3c_O_ind(1, 1, :), fm_scaled_dm_occ_tau, &
1644 : fm_scaled_dm_virt_tau, fm_mo_coeff_occ, fm_mo_coeff_virt, &
1645 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, &
1646 : starts_array_mc, ends_array_mc, starts_array_mc_block, &
1647 : ends_array_mc_block, num_integ_points, nmo, Eigenval(:, 1, :), &
1648 : tau_tj, tau_wj, cut_memory, Pspin, Qspin, my_open_shell, &
1649 18 : unit_nr, dbcsr_time, dbcsr_nflop, mp2_env, qs_env)
1650 : END IF
1651 22 : CALL calc_post_loop_forces(force_data, unit_nr, qs_env)
1652 : END IF !laplace SOS-MP2
1653 :
1654 424 : IF (calc_forces .AND. do_im_time .AND. .NOT. do_ri_sos_laplace_mp2) THEN
1655 64 : DO ispin = 1, nspins
1656 : CALL calc_rpa_loop_forces(force_data, mat_P_omega(:, 1, :), t_3c_M, t_3c_O(1, 1), &
1657 : t_3c_O_compressed(1, 1, :), t_3c_O_ind(1, 1, :), fm_scaled_dm_occ_tau, &
1658 : fm_scaled_dm_virt_tau, fm_mo_coeff_occ, fm_mo_coeff_virt, &
1659 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, &
1660 : starts_array_mc, ends_array_mc, starts_array_mc_block, &
1661 : ends_array_mc_block, num_integ_points, nmo, Eigenval(:, 1, :), &
1662 : e_fermi(ispin), weights_cos_tf_t_to_w, weights_cos_tf_w_to_t, tj, &
1663 : wj, tau_tj, cut_memory, ispin, my_open_shell, unit_nr, dbcsr_time, &
1664 64 : dbcsr_nflop, mp2_env, qs_env)
1665 : END DO
1666 28 : CALL calc_post_loop_forces(force_data, unit_nr, qs_env)
1667 : END IF
1668 :
1669 424 : IF (do_im_time) THEN
1670 :
1671 146 : my_flop_rate = REAL(dbcsr_nflop, dp)/(1.0E09_dp*dbcsr_time)
1672 146 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(/T3,A,T73,ES8.2)") &
1673 73 : "PERFORMANCE| DBCSR total number of flops:", REAL(dbcsr_nflop*para_env%num_pe, dp)
1674 146 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(T3,A,T66,F15.2)") &
1675 73 : "PERFORMANCE| DBCSR total execution time:", dbcsr_time
1676 146 : IF (unit_nr > 0) WRITE (UNIT=unit_nr, FMT="(T3,A,T66,F15.2)") &
1677 73 : "PERFORMANCE| DBCSR flop rate (Gflops / MPI rank):", my_flop_rate
1678 :
1679 : ELSE
1680 :
1681 278 : CALL dgemm_counter_write(dgemm_counter, para_env)
1682 :
1683 : END IF
1684 :
1685 : ! GW: for low-scaling calculation: Compute self-energy Sigma(i*tau), Sigma(i*omega)
1686 : ! for low-scaling and ordinary-scaling: analytic continuation from Sigma(iw) -> Sigma(w)
1687 : ! and correction of quasiparticle energies e_n^GW
1688 720 : IF (my_do_gw) THEN
1689 :
1690 : CALL compute_QP_energies(vec_Sigma_c_gw, count_ev_sc_GW, gw_corr_lev_occ, &
1691 : gw_corr_lev_tot, gw_corr_lev_virt, homo, &
1692 : nmo, num_fit_points, num_integ_points, &
1693 : unit_nr, do_apply_ic_corr_to_gw, do_im_time, &
1694 : do_periodic, do_ri_Sigma_x, first_cycle_periodic_correction, &
1695 : e_fermi, eps_filter, fermi_level_offset, &
1696 : delta_corr, Eigenval, &
1697 : Eigenval_last, Eigenval_scf, iter_sc_GW0, exit_ev_gw, tau_tj, tj, &
1698 : vec_omega_fit_gw, vec_Sigma_x_gw, mp2_env%ri_g0w0%ic_corr_list, &
1699 : weights_cos_tf_t_to_w, weights_sin_tf_t_to_w, &
1700 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, fm_mo_coeff_occ, &
1701 : fm_mo_coeff_virt, fm_scaled_dm_occ_tau, fm_scaled_dm_virt_tau, &
1702 : mo_coeff(1), fm_mat_W, para_env, para_env_RPA, mat_dm, mat_MinvVMinv, &
1703 : t_3c_O, t_3c_M, t_3c_overl_int_ao_mo, t_3c_O_compressed, t_3c_O_mo_compressed, &
1704 : t_3c_O_ind, t_3c_O_mo_ind, &
1705 : t_3c_overl_int_gw_RI, t_3c_overl_int_gw_AO, &
1706 : matrix_berry_im_mo_mo, matrix_berry_re_mo_mo, mat_W, matrix_s, &
1707 : kpoints, mp2_env, qs_env, nkp_self_energy, do_kpoints_cubic_RPA, &
1708 234 : starts_array_mc, ends_array_mc)
1709 :
1710 : ! if HOMO-LUMO gap differs by less than mp2_env%ri_g0w0%eps_ev_sc_iter, exit ev sc GW loop
1711 234 : IF (exit_ev_gw) EXIT
1712 :
1713 : END IF ! my_do_gw if
1714 :
1715 : END DO ! evGW loop
1716 :
1717 296 : IF (do_ic_model) THEN
1718 :
1719 2 : IF (my_open_shell) THEN
1720 :
1721 : CALL calculate_ic_correction(Eigenval(:, 1, 1), mat_MinvVMinv%matrix, &
1722 : t_3c_overl_nnP_ic(1), t_3c_overl_nnP_ic_reflected(1), &
1723 : gw_corr_lev_tot, &
1724 : gw_corr_lev_occ(1), gw_corr_lev_virt(1), homo(1), unit_nr, &
1725 0 : print_ic_values, para_env, do_alpha=.TRUE.)
1726 :
1727 : CALL calculate_ic_correction(Eigenval(:, 1, 2), mat_MinvVMinv%matrix, &
1728 : t_3c_overl_nnP_ic(2), t_3c_overl_nnP_ic_reflected(2), &
1729 : gw_corr_lev_tot, &
1730 : gw_corr_lev_occ(2), gw_corr_lev_virt(2), homo(2), unit_nr, &
1731 0 : print_ic_values, para_env, do_beta=.TRUE.)
1732 :
1733 : ELSE
1734 :
1735 : CALL calculate_ic_correction(Eigenval(:, 1, 1), mat_MinvVMinv%matrix, &
1736 : t_3c_overl_nnP_ic(1), t_3c_overl_nnP_ic_reflected(1), &
1737 : gw_corr_lev_tot, &
1738 : gw_corr_lev_occ(1), gw_corr_lev_virt(1), homo(1), unit_nr, &
1739 2 : print_ic_values, para_env)
1740 :
1741 : END IF
1742 :
1743 : END IF
1744 :
1745 : ! postprocessing after GW for Bethe-Salpeter
1746 296 : IF (do_bse) THEN
1747 : ! Check used GW flavor; in Case of evGW we use W0 for BSE
1748 : ! Use environment variable, since local iter_evGW is overwritten if evGW0 is invoked
1749 32 : IF (mp2_env%ri_g0w0%iter_evGW > 1) THEN
1750 8 : IF (unit_nr > 0) THEN
1751 : CALL cp_warn(__LOCATION__, &
1752 : "BSE@evGW applies W0, i.e. screening with DFT energies! "// &
1753 4 : "Use BSE@evGW0 to avoid this warning!")
1754 : END IF
1755 : END IF
1756 : ! Create a copy of fm_mat_S for usage in BSE
1757 32 : CALL cp_fm_create(fm_mat_S_ia_bse, fm_mat_S(1)%matrix_struct)
1758 32 : CALL cp_fm_to_fm(fm_mat_S(1), fm_mat_S_ia_bse)
1759 : ! Remove energy/frequency factor from 3c-Integral for BSE
1760 32 : IF (iter_sc_gw0 == 1) THEN
1761 : CALL remove_scaling_factor_rpa(fm_mat_S_ia_bse, virtual(1), &
1762 24 : Eigenval_last(:, 1, 1), homo(1), omega)
1763 : ELSE
1764 : CALL remove_scaling_factor_rpa(fm_mat_S_ia_bse, virtual(1), &
1765 8 : Eigenval_scf(:, 1, 1), homo(1), omega)
1766 : END IF
1767 : ! Main routine for all BSE postprocessing
1768 : CALL start_bse_calculation(fm_mat_S_ia_bse, fm_mat_S_ij_bse, fm_mat_S_ab_bse, &
1769 : fm_mat_Q_static_bse_gemm, &
1770 : Eigenval, Eigenval_scf, &
1771 : homo, virtual, dimen_RI, dimen_RI_red, bse_lev_virt, &
1772 32 : gd_array, color_sub, mp2_env, qs_env, mo_coeff, unit_nr)
1773 : ! Release BSE-copy of fm_mat_S
1774 32 : CALL cp_fm_release(fm_mat_S_ia_bse)
1775 : END IF
1776 :
1777 296 : IF (my_do_gw) THEN
1778 : CALL deallocate_matrices_gw(fm_mat_S_gw_work, vec_W_gw, vec_Sigma_c_gw, vec_omega_fit_gw, &
1779 : mp2_env%ri_g0w0%vec_Sigma_x_minus_vxc_gw, &
1780 : Eigenval_last, Eigenval_scf, do_periodic, matrix_berry_re_mo_mo, matrix_berry_im_mo_mo, &
1781 106 : kpoints, vec_Sigma_x_gw,.NOT. do_im_time)
1782 : END IF
1783 :
1784 296 : IF (do_im_time) THEN
1785 :
1786 : CALL dealloc_im_time(fm_mo_coeff_occ, fm_mo_coeff_virt, &
1787 : fm_scaled_dm_occ_tau, fm_scaled_dm_virt_tau, index_to_cell_3c, &
1788 : cell_to_index_3c, do_ic_model, &
1789 : do_kpoints_cubic_RPA, do_kpoints_from_Gamma, do_ri_Sigma_x, &
1790 : has_mat_P_blocks, &
1791 : wkp_W, cfm_mat_Q, fm_mat_Minv_L_kpoints, fm_mat_L_kpoints, &
1792 : fm_matrix_Minv, fm_matrix_Minv_Vtrunc_Minv, fm_mat_RI_global_work, fm_mat_work, &
1793 : fm_mo_coeff_occ_scaled, fm_mo_coeff_virt_scaled, mat_dm, mat_L, &
1794 : mat_MinvVMinv, mat_P_omega, mat_P_omega_kp, &
1795 134 : t_3c_M, t_3c_O, t_3c_O_compressed, t_3c_O_ind, mat_work, qs_env)
1796 :
1797 134 : IF (my_do_gw) THEN
1798 : CALL deallocate_matrices_gw_im_time(weights_cos_tf_w_to_t, weights_sin_tf_t_to_w, do_ic_model, &
1799 : do_kpoints_cubic_RPA, fm_mat_W, &
1800 : t_3c_overl_int_ao_mo, t_3c_O_mo_compressed, t_3c_O_mo_ind, &
1801 : t_3c_overl_int_gw_RI, t_3c_overl_int_gw_AO, &
1802 : t_3c_overl_nnP_ic, t_3c_overl_nnP_ic_reflected, &
1803 46 : mat_W, qs_env)
1804 : END IF
1805 :
1806 : END IF
1807 :
1808 296 : IF (.NOT. do_im_time .AND. .NOT. do_ri_sos_laplace_mp2) CALL exchange_work%release()
1809 :
1810 296 : IF (.NOT. do_ri_sos_laplace_mp2) THEN
1811 238 : DEALLOCATE (tj)
1812 238 : DEALLOCATE (wj)
1813 238 : DEALLOCATE (trace_Qomega)
1814 : END IF
1815 :
1816 296 : IF (do_im_time .OR. do_ri_sos_laplace_mp2) THEN
1817 162 : DEALLOCATE (tau_tj)
1818 162 : DEALLOCATE (tau_wj)
1819 : END IF
1820 :
1821 296 : IF (do_im_time .AND. calc_forces) THEN
1822 50 : CALL im_time_force_release(force_data)
1823 : END IF
1824 :
1825 296 : IF (calc_forces .AND. .NOT. do_im_time) CALL rpa_grad_finalize(rpa_grad, mp2_env, para_env_sub, para_env, &
1826 : qs_env, gd_array, color_sub, do_ri_sos_laplace_mp2, &
1827 40 : homo, virtual)
1828 :
1829 296 : CALL timestop(handle)
1830 :
1831 6050 : END SUBROUTINE rpa_num_int
1832 :
1833 : END MODULE rpa_main
|
