(git:039bc03)
Loading...
Searching...
No Matches
ec_orth_solver.F
Go to the documentation of this file.
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! **************************************************************************************************
9!> \brief AO-based conjugate-gradient response solver routines
10!>
11!>
12!> \date 09.2019
13!> \author Fabian Belleflamme
14! **************************************************************************************************
15MODULE ec_orth_solver
16 USE admm_types, ONLY: admm_type,&
17 get_admm_env
18 USE cp_control_types, ONLY: dft_control_type
19 USE cp_dbcsr_api, ONLY: &
20 dbcsr_add, dbcsr_copy, dbcsr_create, dbcsr_desymmetrize, dbcsr_filter, dbcsr_finalize, &
21 dbcsr_get_info, dbcsr_multiply, dbcsr_p_type, dbcsr_release, dbcsr_scale, dbcsr_set, &
22 dbcsr_transposed, dbcsr_type, dbcsr_type_no_symmetry
23 USE cp_dbcsr_contrib, ONLY: dbcsr_add_on_diag,&
24 dbcsr_checksum,&
25 dbcsr_dot
26 USE cp_dbcsr_operations, ONLY: dbcsr_allocate_matrix_set,&
27 dbcsr_deallocate_matrix_set
28 USE cp_external_control, ONLY: external_control
29 USE ec_methods, ONLY: create_kernel
30 USE input_constants, ONLY: do_admm_aux_exch_func_none,&
31 kg_tnadd_embed,&
32 kg_tnadd_embed_ri,&
33 ls_s_sqrt_ns,&
34 ls_s_sqrt_proot,&
35 precond_mlp
36 USE input_section_types, ONLY: section_vals_get_subs_vals,&
37 section_vals_type,&
38 section_vals_val_get
39 USE iterate_matrix, ONLY: matrix_sqrt_newton_schulz,&
40 matrix_sqrt_proot
41 USE kg_correction, ONLY: kg_ekin_subset
42 USE kinds, ONLY: dp
43 USE machine, ONLY: m_flush,&
44 m_walltime
45 USE mathlib, ONLY: abnormal_value
46 USE message_passing, ONLY: mp_para_env_type
47 USE pw_env_types, ONLY: pw_env_get,&
48 pw_env_type
49 USE pw_methods, ONLY: pw_axpy,&
50 pw_scale,&
51 pw_transfer,&
52 pw_zero
53 USE pw_poisson_methods, ONLY: pw_poisson_solve
54 USE pw_poisson_types, ONLY: pw_poisson_type
55 USE pw_pool_types, ONLY: pw_pool_p_type,&
56 pw_pool_type
57 USE pw_types, ONLY: pw_c1d_gs_type,&
58 pw_r3d_rs_type
59 USE qs_environment_types, ONLY: get_qs_env,&
60 qs_environment_type
61 USE qs_integrate_potential, ONLY: integrate_v_rspace
62 USE qs_kpp1_env_types, ONLY: qs_kpp1_env_type
63 USE qs_linres_kernel, ONLY: apply_hfx,&
64 apply_xc_admm
65 USE qs_linres_types, ONLY: linres_control_type
66 USE qs_p_env_methods, ONLY: p_env_check_i_alloc,&
67 p_env_finish_kpp1,&
68 p_env_update_rho
69 USE qs_p_env_types, ONLY: qs_p_env_type
70 USE qs_rho_types, ONLY: qs_rho_get,&
71 qs_rho_type
72 USE xc, ONLY: xc_prep_2nd_deriv
73#include "./base/base_uses.f90"
74
75 IMPLICIT NONE
76
77 PRIVATE
78
79 ! Global parameters
80
81 CHARACTER(len=*), PARAMETER, PRIVATE :: moduleN = 'ec_orth_solver'
82
83 ! Public subroutines
84
85 PUBLIC :: ec_response_ao
86
87CONTAINS
88
89! **************************************************************************************************
90!> \brief Preconditioning of the AO-based CG linear response solver
91!> M * z_0 = r_0
92!> M(X) = [F,B], with B = [X,P]
93!> for M we need F and P in ortho basis
94!> Returns z_0, the preconditioned residual in orthonormal basis
95!>
96!> All matrices are in orthonormal Lowdin basis
97!>
98!> \param qs_env ...
99!> \param matrix_ks Ground-state Kohn-Sham matrix
100!> \param matrix_p Ground-state Density matrix
101!> \param matrix_rhs Unpreconditioned residual of linear response CG
102!> \param matrix_cg_z Preconditioned residual
103!> \param eps_filter ...
104!> \param iounit ...
105!>
106!> \param silent ...
107!> \date 01.2020
108!> \author Fabian Belleflamme
109! **************************************************************************************************
110 SUBROUTINE ec_preconditioner(qs_env, matrix_ks, matrix_p, matrix_rhs, &
111 matrix_cg_z, eps_filter, iounit, silent)
112
113 TYPE(qs_environment_type), POINTER :: qs_env
114 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
115 POINTER :: matrix_ks, matrix_p, matrix_rhs
116 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(INOUT), &
117 POINTER :: matrix_cg_z
118 REAL(KIND=dp), INTENT(IN) :: eps_filter
119 INTEGER, INTENT(IN) :: iounit
120 LOGICAL, INTENT(IN), OPTIONAL :: silent
121
122 CHARACTER(len=*), PARAMETER :: routineN = 'ec_preconditioner'
123
124 INTEGER :: handle, i, ispin, max_iter, nao, nspins
125 LOGICAL :: converged, my_silent
126 REAL(KIND=dp) :: norm_res, t1, t2
127 REAL(KIND=dp), DIMENSION(:), POINTER :: alpha, beta, new_norm, norm_ca, norm_rr
128 TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: matrix_Ax, matrix_b, matrix_cg, &
129 matrix_res
130 TYPE(dft_control_type), POINTER :: dft_control
131 TYPE(linres_control_type), POINTER :: linres_control
132
133 CALL timeset(routinen, handle)
134
135 my_silent = .false.
136 IF (PRESENT(silent)) my_silent = silent
137
138 cpassert(ASSOCIATED(qs_env))
139 cpassert(ASSOCIATED(matrix_ks))
140 cpassert(ASSOCIATED(matrix_p))
141 cpassert(ASSOCIATED(matrix_rhs))
142 cpassert(ASSOCIATED(matrix_cg_z))
143
144 NULLIFY (dft_control, linres_control)
145
146 t1 = m_walltime()
147
148 CALL get_qs_env(qs_env=qs_env, &
149 dft_control=dft_control, &
150 linres_control=linres_control)
151 nspins = dft_control%nspins
152 CALL dbcsr_get_info(matrix_ks(1)%matrix, nfullrows_total=nao)
153
154 ALLOCATE (alpha(nspins), beta(nspins), new_norm(nspins), norm_ca(nspins), norm_rr(nspins))
155
156 !----------------------------------------
157 ! Create non-symmetric matrices: Ax, B, cg, res
158 !----------------------------------------
159
160 NULLIFY (matrix_ax, matrix_b, matrix_cg, matrix_res)
161 CALL dbcsr_allocate_matrix_set(matrix_ax, nspins)
162 CALL dbcsr_allocate_matrix_set(matrix_b, nspins)
163 CALL dbcsr_allocate_matrix_set(matrix_cg, nspins)
164 CALL dbcsr_allocate_matrix_set(matrix_res, nspins)
165
166 DO ispin = 1, nspins
167 ALLOCATE (matrix_ax(ispin)%matrix)
168 ALLOCATE (matrix_b(ispin)%matrix)
169 ALLOCATE (matrix_cg(ispin)%matrix)
170 ALLOCATE (matrix_res(ispin)%matrix)
171 CALL dbcsr_create(matrix_ax(ispin)%matrix, name="linop MATRIX", &
172 template=matrix_ks(1)%matrix, &
173 matrix_type=dbcsr_type_no_symmetry)
174 CALL dbcsr_create(matrix_b(ispin)%matrix, name="MATRIX B", &
175 template=matrix_ks(1)%matrix, &
176 matrix_type=dbcsr_type_no_symmetry)
177 CALL dbcsr_create(matrix_cg(ispin)%matrix, name="TRIAL MATRIX", &
178 template=matrix_ks(1)%matrix, &
179 matrix_type=dbcsr_type_no_symmetry)
180 CALL dbcsr_create(matrix_res(ispin)%matrix, name="RESIDUE", &
181 template=matrix_ks(1)%matrix, &
182 matrix_type=dbcsr_type_no_symmetry)
183 END DO
184
185 !----------------------------------------
186 ! Get righ-hand-side operators
187 !----------------------------------------
188
189 ! Initial guess z_0
190 DO ispin = 1, nspins
191 CALL dbcsr_copy(matrix_cg_z(ispin)%matrix, matrix_rhs(ispin)%matrix)
192
193 ! r_0 = b
194 CALL dbcsr_copy(matrix_res(ispin)%matrix, matrix_rhs(ispin)%matrix)
195 END DO
196
197 ! Projector on trial matrix
198 ! Projector does not need to be applied here,
199 ! as matrix_rhs already had this done before entering preconditioner
200 !CALL projector(qs_env, matrix_p, matrix_cg_z, eps_filter)
201
202 ! Mz_0
203 CALL hessian_op1(matrix_ks, matrix_p, matrix_cg_z, matrix_b, matrix_ax, eps_filter)
204
205 ! r_0 = b - Ax_0
206 DO ispin = 1, nspins
207 CALL dbcsr_add(matrix_res(ispin)%matrix, matrix_ax(ispin)%matrix, 1.0_dp, -1.0_dp)
208 END DO
209
210 ! Matrix projector T
211 CALL projector(qs_env, matrix_p, matrix_res, eps_filter)
212
213 DO ispin = 1, nspins
214 ! cg = p_0 = z_0
215 CALL dbcsr_copy(matrix_cg(ispin)%matrix, matrix_res(ispin)%matrix)
216 END DO
217
218 ! header
219 IF (iounit > 0 .AND. .NOT. my_silent) THEN
220 WRITE (iounit, "(/,T10,A)") "Preconditioning of search direction"
221 WRITE (iounit, "(/,T10,A,T25,A,T42,A,T62,A,/,T10,A)") &
222 "Iteration", "Stepsize", "Convergence", "Time", &
223 repeat("-", 58)
224 END IF
225
226 alpha(:) = 0.0_dp
227 max_iter = 200
228 converged = .false.
229 norm_res = 0.0_dp
230
231 ! start iteration
232 iteration: DO i = 1, max_iter
233
234 ! Hessian Ax = [F,B] is updated preconditioner
235 CALL hessian_op1(matrix_ks, matrix_p, matrix_cg, matrix_b, matrix_ax, eps_filter)
236
237 ! Matrix projector
238 CALL projector(qs_env, matrix_p, matrix_ax, eps_filter)
239
240 DO ispin = 1, nspins
241
242 ! Tr(r_0 * r_0)
243 CALL dbcsr_dot(matrix_res(ispin)%matrix, matrix_res(ispin)%matrix, norm_rr(ispin))
244 IF (abnormal_value(norm_rr(ispin))) &
245 cpabort("Preconditioner: Tr[r_j*r_j] is an abnormal value (NaN/Inf)")
246
247 IF (norm_rr(ispin) < 0.0_dp) cpabort("norm_rr < 0")
248 norm_res = max(norm_res, abs(norm_rr(ispin)/real(nao, dp)))
249
250 ! norm_cA = tr(Ap_j * p_j)
251 CALL dbcsr_dot(matrix_cg(ispin)%matrix, matrix_ax(ispin)%matrix, norm_ca(ispin))
252
253 ! Determine step-size
254 IF (norm_ca(ispin) < linres_control%eps) THEN
255 alpha(ispin) = 1.0_dp
256 ELSE
257 alpha(ispin) = norm_rr(ispin)/norm_ca(ispin)
258 END IF
259
260 ! x_j+1 = x_j + alpha*p_j
261 ! save contribution of this iteration
262 CALL dbcsr_add(matrix_cg_z(ispin)%matrix, matrix_cg(ispin)%matrix, 1.0_dp, alpha(ispin))
263
264 ! r_j+1 = r_j - alpha * Ap_j
265 CALL dbcsr_add(matrix_res(ispin)%matrix, matrix_ax(ispin)%matrix, 1.0_dp, -alpha(ispin))
266
267 END DO
268
269 norm_res = 0.0_dp
270
271 DO ispin = 1, nspins
272 ! Tr[r_j+1*z_j+1]
273 CALL dbcsr_dot(matrix_res(ispin)%matrix, matrix_res(ispin)%matrix, new_norm(ispin))
274 IF (new_norm(ispin) < 0.0_dp) cpabort("tr(r_j+1*z_j+1) < 0")
275 IF (abnormal_value(new_norm(ispin))) &
276 cpabort("Preconditioner: Tr[r_j+1*z_j+1] is an abnormal value (NaN/Inf)")
277 norm_res = max(norm_res, new_norm(ispin)/real(nao, dp))
278
279 IF (norm_rr(ispin) < linres_control%eps*0.001_dp &
280 .OR. new_norm(ispin) < linres_control%eps*0.001_dp) THEN
281 beta(ispin) = 0.0_dp
282 converged = .true.
283 ELSE
284 beta(ispin) = new_norm(ispin)/norm_rr(ispin)
285 END IF
286
287 ! update new search vector (matrix cg)
288 ! cg_j+1 = z_j+1 + beta*cg_j
289 CALL dbcsr_add(matrix_cg(ispin)%matrix, matrix_res(ispin)%matrix, beta(ispin), 1.0_dp)
290 CALL dbcsr_filter(matrix_cg(ispin)%matrix, eps_filter)
291
292 norm_rr(ispin) = new_norm(ispin)
293 END DO
294
295 ! Convergence criteria
296 IF (norm_res < linres_control%eps) THEN
297 converged = .true.
298 END IF
299
300 t2 = m_walltime()
301 IF (i == 1 .OR. mod(i, 1) == 0 .OR. converged) THEN
302 IF (iounit > 0 .AND. .NOT. my_silent) THEN
303 WRITE (iounit, "(T10,I5,T25,1E8.2,T33,F25.14,T58,F8.2)") &
304 i, maxval(alpha), norm_res, t2 - t1
305 ! Convergence in scientific notation
306 !WRITE (iounit, "(T10,I5,T25,1E8.2,T42,1E14.8,T58,F8.2)") &
307 ! i, MAXVAL(alpha), norm_res, t2 - t1
308 CALL m_flush(iounit)
309 END IF
310 END IF
311 IF (converged) THEN
312 IF (iounit > 0 .AND. .NOT. my_silent) THEN
313 WRITE (iounit, "(/,T10,A,I4,A,/)") "The precon solver converged in ", i, " iterations."
314 CALL m_flush(iounit)
315 END IF
316 EXIT iteration
317 END IF
318
319 ! Max number of iteration reached
320 IF (i == max_iter) THEN
321 IF (iounit > 0) THEN
322 WRITE (iounit, "(/,T10,A/)") &
323 "The precon solver didnt converge! Maximum number of iterations reached."
324 CALL m_flush(iounit)
325 END IF
326 converged = .false.
327 END IF
328
329 END DO iteration
330
331 ! Matrix projector
332 CALL projector(qs_env, matrix_p, matrix_cg_z, eps_filter)
333
334 ! Release matrices
335 CALL dbcsr_deallocate_matrix_set(matrix_ax)
336 CALL dbcsr_deallocate_matrix_set(matrix_b)
337 CALL dbcsr_deallocate_matrix_set(matrix_res)
338 CALL dbcsr_deallocate_matrix_set(matrix_cg)
339
340 DEALLOCATE (alpha, beta, new_norm, norm_ca, norm_rr)
341
342 CALL timestop(handle)
343
344 END SUBROUTINE ec_preconditioner
345
346! **************************************************************************************************
347!> \brief AO-based conjugate gradient linear response solver.
348!> In goes the right hand side B of the equation AZ=B, and the linear transformation of the
349!> Hessian matrix A on trial matrices is iteratively solved. Result are
350!> the response density matrix_pz, and the energy-weighted response density matrix_wz.
351!>
352!> \param qs_env ...
353!> \param p_env ...
354!> \param matrix_hz Right hand-side of linear response equation
355!> \param matrix_pz Response density
356!> \param matrix_wz Energy-weighted response density matrix
357!> \param iounit ...
358!> \param should_stop ...
359!>
360!> \param silent ...
361!> \date 01.2020
362!> \author Fabian Belleflamme
363! **************************************************************************************************
364 SUBROUTINE ec_response_ao(qs_env, p_env, matrix_hz, matrix_pz, matrix_wz, iounit, &
365 should_stop, silent)
366
367 TYPE(qs_environment_type), POINTER :: qs_env
368 TYPE(qs_p_env_type), POINTER :: p_env
369 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
370 POINTER :: matrix_hz
371 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(INOUT), &
372 POINTER :: matrix_pz, matrix_wz
373 INTEGER, INTENT(IN) :: iounit
374 LOGICAL, INTENT(OUT) :: should_stop
375 LOGICAL, INTENT(IN), OPTIONAL :: silent
376
377 CHARACTER(len=*), PARAMETER :: routinen = 'ec_response_ao'
378
379 INTEGER :: handle, i, ispin, max_iter_lanczos, nao, &
380 nspins, s_sqrt_method, s_sqrt_order
381 LOGICAL :: my_silent, restart
382 REAL(kind=dp) :: eps_filter, eps_lanczos, focc, &
383 min_shift, norm_res, old_conv, shift, &
384 t1, t2
385 REAL(kind=dp), DIMENSION(:), POINTER :: alpha, beta, new_norm, norm_ca, norm_rr, &
386 tr_rz00
387 TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: ksmat, matrix_ax, matrix_cg, matrix_cg_z, &
388 matrix_ks, matrix_nsc, matrix_p, matrix_res, matrix_s, matrix_z, matrix_z0, rho_ao
389 TYPE(dbcsr_type) :: matrix_s_sqrt, matrix_s_sqrt_inv, &
390 matrix_tmp
391 TYPE(dft_control_type), POINTER :: dft_control
392 TYPE(linres_control_type), POINTER :: linres_control
393 TYPE(qs_rho_type), POINTER :: rho
394 TYPE(section_vals_type), POINTER :: solver_section
395
396 CALL timeset(routinen, handle)
397
398 my_silent = .false.
399 IF (PRESENT(silent)) my_silent = silent
400
401 cpassert(ASSOCIATED(qs_env))
402 cpassert(ASSOCIATED(matrix_hz))
403 cpassert(ASSOCIATED(matrix_pz))
404 cpassert(ASSOCIATED(matrix_wz))
405
406 NULLIFY (dft_control, ksmat, matrix_s, linres_control, rho)
407
408 t1 = m_walltime()
409
410 CALL get_qs_env(qs_env=qs_env, &
411 dft_control=dft_control, &
412 linres_control=linres_control, &
413 matrix_ks=ksmat, &
414 matrix_s=matrix_s, &
415 rho=rho)
416 nspins = dft_control%nspins
417
418 CALL dbcsr_get_info(matrix_s(1)%matrix, nfullrows_total=nao)
419
420 solver_section => section_vals_get_subs_vals(qs_env%input, "DFT%ENERGY_CORRECTION%RESPONSE_SOLVER")
421 CALL section_vals_val_get(solver_section, "S_SQRT_METHOD", i_val=s_sqrt_method)
422 CALL section_vals_val_get(solver_section, "S_SQRT_ORDER", i_val=s_sqrt_order)
423 CALL section_vals_val_get(solver_section, "EPS_LANCZOS", r_val=eps_lanczos)
424 CALL section_vals_val_get(solver_section, "MAX_ITER_LANCZOS", i_val=max_iter_lanczos)
425
426 eps_filter = linres_control%eps_filter
427
428 CALL qs_rho_get(rho, rho_ao=rho_ao)
429
430 ALLOCATE (alpha(nspins), beta(nspins), new_norm(nspins), norm_ca(nspins), norm_rr(nspins))
431 ALLOCATE (tr_rz00(nspins))
432
433 ! local matrix P, KS, and NSC
434 ! to bring into orthogonal basis
435 NULLIFY (matrix_p, matrix_ks, matrix_nsc)
436 CALL dbcsr_allocate_matrix_set(matrix_p, nspins)
437 CALL dbcsr_allocate_matrix_set(matrix_ks, nspins)
438 CALL dbcsr_allocate_matrix_set(matrix_nsc, nspins)
439 DO ispin = 1, nspins
440 ALLOCATE (matrix_p(ispin)%matrix)
441 ALLOCATE (matrix_ks(ispin)%matrix)
442 ALLOCATE (matrix_nsc(ispin)%matrix)
443 CALL dbcsr_create(matrix_p(ispin)%matrix, name="P_IN ORTHO", &
444 template=ksmat(1)%matrix, &
445 matrix_type=dbcsr_type_no_symmetry)
446 CALL dbcsr_create(matrix_ks(ispin)%matrix, name="KS_IN ORTHO", &
447 template=ksmat(1)%matrix, &
448 matrix_type=dbcsr_type_no_symmetry)
449 CALL dbcsr_create(matrix_nsc(ispin)%matrix, name="NSC IN ORTHO", &
450 template=ksmat(1)%matrix, &
451 matrix_type=dbcsr_type_no_symmetry)
452
453 CALL dbcsr_desymmetrize(rho_ao(ispin)%matrix, matrix_p(ispin)%matrix)
454 CALL dbcsr_desymmetrize(ksmat(ispin)%matrix, matrix_ks(ispin)%matrix)
455 CALL dbcsr_desymmetrize(matrix_hz(ispin)%matrix, matrix_nsc(ispin)%matrix)
456 END DO
457
458 ! Scale matrix_p by factor 1/2 in closed-shell
459 IF (nspins == 1) CALL dbcsr_scale(matrix_p(1)%matrix, 0.5_dp)
460
461 ! Transform P, KS, and Harris kernel matrix into Orthonormal basis
462 CALL dbcsr_create(matrix_s_sqrt, template=matrix_s(1)%matrix, &
463 matrix_type=dbcsr_type_no_symmetry)
464 CALL dbcsr_create(matrix_s_sqrt_inv, template=matrix_s(1)%matrix, &
465 matrix_type=dbcsr_type_no_symmetry)
466
467 SELECT CASE (s_sqrt_method)
468 CASE (ls_s_sqrt_proot)
469 CALL matrix_sqrt_proot(matrix_s_sqrt, matrix_s_sqrt_inv, &
470 matrix_s(1)%matrix, eps_filter, &
471 s_sqrt_order, eps_lanczos, max_iter_lanczos, symmetrize=.true.)
472 CASE (ls_s_sqrt_ns)
473 CALL matrix_sqrt_newton_schulz(matrix_s_sqrt, matrix_s_sqrt_inv, &
474 matrix_s(1)%matrix, eps_filter, &
475 s_sqrt_order, eps_lanczos, max_iter_lanczos)
476 CASE DEFAULT
477 cpabort("Unknown sqrt method.")
478 END SELECT
479
480 ! Transform into orthonormal Lowdin basis
481 DO ispin = 1, nspins
482 CALL transform_m_orth(matrix_p(ispin)%matrix, matrix_s_sqrt, eps_filter)
483 CALL transform_m_orth(matrix_ks(ispin)%matrix, matrix_s_sqrt_inv, eps_filter)
484 CALL transform_m_orth(matrix_nsc(ispin)%matrix, matrix_s_sqrt_inv, eps_filter)
485 END DO
486
487 !----------------------------------------
488 ! Create non-symmetric work matrices: Ax, cg, res
489 ! Content of Ax, cg, cg_z, res, z0 anti-symmetric
490 ! Content of z symmetric
491 !----------------------------------------
492
493 CALL dbcsr_create(matrix_tmp, template=matrix_s(1)%matrix, matrix_type=dbcsr_type_no_symmetry)
494
495 NULLIFY (matrix_ax, matrix_cg, matrix_cg_z, matrix_res, matrix_z, matrix_z0)
496 CALL dbcsr_allocate_matrix_set(matrix_ax, nspins)
497 CALL dbcsr_allocate_matrix_set(matrix_cg, nspins)
498 CALL dbcsr_allocate_matrix_set(matrix_cg_z, nspins)
499 CALL dbcsr_allocate_matrix_set(matrix_res, nspins)
500 CALL dbcsr_allocate_matrix_set(matrix_z, nspins)
501 CALL dbcsr_allocate_matrix_set(matrix_z0, nspins)
502
503 DO ispin = 1, nspins
504 ALLOCATE (matrix_ax(ispin)%matrix)
505 ALLOCATE (matrix_cg(ispin)%matrix)
506 ALLOCATE (matrix_cg_z(ispin)%matrix)
507 ALLOCATE (matrix_res(ispin)%matrix)
508 ALLOCATE (matrix_z(ispin)%matrix)
509 ALLOCATE (matrix_z0(ispin)%matrix)
510 CALL dbcsr_create(matrix_ax(ispin)%matrix, name="linop MATRIX", &
511 template=matrix_s(1)%matrix, &
512 matrix_type=dbcsr_type_no_symmetry)
513 CALL dbcsr_create(matrix_cg(ispin)%matrix, name="TRIAL MATRIX", &
514 template=matrix_s(1)%matrix, &
515 matrix_type=dbcsr_type_no_symmetry)
516 CALL dbcsr_create(matrix_cg_z(ispin)%matrix, name="MATRIX CG-Z", &
517 template=matrix_s(1)%matrix, &
518 matrix_type=dbcsr_type_no_symmetry)
519 CALL dbcsr_create(matrix_res(ispin)%matrix, name="RESIDUE", &
520 template=matrix_s(1)%matrix, &
521 matrix_type=dbcsr_type_no_symmetry)
522 CALL dbcsr_create(matrix_z(ispin)%matrix, name="Z-Matrix", &
523 template=matrix_s(1)%matrix, &
524 matrix_type=dbcsr_type_no_symmetry)
525 CALL dbcsr_create(matrix_z0(ispin)%matrix, name="p after precondi-Matrix", &
526 template=matrix_s(1)%matrix, &
527 matrix_type=dbcsr_type_no_symmetry)
528 END DO
529
530 !----------------------------------------
531 ! Get righ-hand-side operators
532 !----------------------------------------
533
534 ! Spin factor
535 focc = -2.0_dp
536 IF (nspins == 1) focc = -4.0_dp
537
538 ! E^[1]_Harris = -4*G[\delta P]*Pin - Pin*G[\delta P] = -4*[G[\delta P], Pin]
539 CALL commutator(matrix_nsc, matrix_p, matrix_res, eps_filter, .false., alpha=focc)
540
541 ! Initial guess cg_Z
542 DO ispin = 1, nspins
543 CALL dbcsr_copy(matrix_cg_z(ispin)%matrix, matrix_res(ispin)%matrix)
544 END DO
545
546 ! Projector on trial matrix
547 CALL projector(qs_env, matrix_p, matrix_cg_z, eps_filter)
548
549 ! Ax0
550 CALL build_hessian_op(qs_env=qs_env, &
551 p_env=p_env, &
552 matrix_ks=matrix_ks, &
553 matrix_p=matrix_p, & ! p
554 matrix_s_sqrt_inv=matrix_s_sqrt_inv, &
555 matrix_cg=matrix_cg_z, & ! cg
556 matrix_ax=matrix_ax, &
557 eps_filter=eps_filter)
558
559 ! r_0 = b - Ax0
560 DO ispin = 1, nspins
561 CALL dbcsr_add(matrix_res(ispin)%matrix, matrix_ax(ispin)%matrix, 1.0_dp, -1.0_dp)
562 END DO
563
564 ! Matrix projector T
565 CALL projector(qs_env, matrix_p, matrix_res, eps_filter)
566
567 ! Preconditioner
568 linres_control%flag = ""
569 IF (linres_control%preconditioner_type == precond_mlp) THEN
570 ! M * z_0 = r_0
571 ! Conjugate gradient returns z_0
572 CALL ec_preconditioner(qs_env=qs_env, &
573 matrix_ks=matrix_ks, &
574 matrix_p=matrix_p, &
575 matrix_rhs=matrix_res, &
576 matrix_cg_z=matrix_z0, &
577 eps_filter=eps_filter, &
578 iounit=iounit, silent=silent)
579 linres_control%flag = "PCG-AO"
580 ELSE
581 ! z_0 = r_0
582 DO ispin = 1, nspins
583 CALL dbcsr_copy(matrix_z0(ispin)%matrix, matrix_res(ispin)%matrix)
584 linres_control%flag = "CG-AO"
585 END DO
586 END IF
587
588 norm_res = 0.0_dp
589
590 DO ispin = 1, nspins
591 ! cg = p_0 = z_0
592 CALL dbcsr_copy(matrix_cg(ispin)%matrix, matrix_z0(ispin)%matrix)
593
594 ! Tr(r_0 * z_0)
595 CALL dbcsr_dot(matrix_res(ispin)%matrix, matrix_cg(ispin)%matrix, norm_rr(ispin))
596
597 IF (norm_rr(ispin) < 0.0_dp) cpabort("norm_rr < 0")
598 norm_res = max(norm_res, abs(norm_rr(ispin)/real(nao, dp)))
599 END DO
600
601 ! eigenvalue shifting
602 min_shift = 0.0_dp
603 old_conv = norm_rr(1)
604 shift = min(10.0_dp, max(min_shift, 0.05_dp*old_conv))
605 old_conv = 100.0_dp
606
607 ! header
608 IF (iounit > 0 .AND. .NOT. my_silent) THEN
609 WRITE (iounit, "(/,T3,A,T16,A,T25,A,T38,A,T52,A,/,T3,A)") &
610 "Iteration", "Method", "Stepsize", "Convergence", "Time", &
611 repeat("-", 80)
612 END IF
613
614 alpha(:) = 0.0_dp
615 restart = .false.
616 should_stop = .false.
617 linres_control%converged = .false.
618
619 ! start iteration
620 iteration: DO i = 1, linres_control%max_iter
621
622 ! Convergence criteria
623 ! default for eps 10E-6 in MO_linres
624 IF (norm_res < linres_control%eps) THEN
625 linres_control%converged = .true.
626 END IF
627
628 t2 = m_walltime()
629 IF (i == 1 .OR. mod(i, 1) == 0 .OR. linres_control%converged &
630 .OR. restart .OR. should_stop) THEN
631 IF (iounit > 0 .AND. .NOT. my_silent) THEN
632 WRITE (iounit, "(T5,I5,T18,A3,T28,L1,T38,1E8.2,T48,F16.10,T68,F8.2)") &
633 i, linres_control%flag, restart, maxval(alpha), norm_res, t2 - t1
634 CALL m_flush(iounit)
635 END IF
636 END IF
637 IF (linres_control%converged) THEN
638 IF (iounit > 0) THEN
639 WRITE (iounit, "(/,T2,A,I4,A,T73,F8.2,/)") "The linear solver converged in ", &
640 i, " iterations.", t2 - t1
641 CALL m_flush(iounit)
642 END IF
643 EXIT iteration
644 ELSE IF (should_stop) THEN
645 IF (iounit > 0) THEN
646 WRITE (iounit, "(/,T2,A,I4,A,/)") "The linear solver did NOT converge! External stop"
647 CALL m_flush(iounit)
648 END IF
649 EXIT iteration
650 END IF
651
652 ! Max number of iteration reached
653 IF (i == linres_control%max_iter) THEN
654 IF (iounit > 0) THEN
655 WRITE (iounit, "(/,T2,A/)") &
656 "The linear solver didnt converge! Maximum number of iterations reached."
657 CALL m_flush(iounit)
658 END IF
659 linres_control%converged = .false.
660 END IF
661
662 ! Hessian Ax = [F,B] + [G(B),P]
663 CALL build_hessian_op(qs_env=qs_env, &
664 p_env=p_env, &
665 matrix_ks=matrix_ks, &
666 matrix_p=matrix_p, & ! p
667 matrix_s_sqrt_inv=matrix_s_sqrt_inv, &
668 matrix_cg=matrix_cg, & ! cg
669 matrix_ax=matrix_ax, &
670 eps_filter=eps_filter)
671
672 ! Matrix projector T
673 CALL projector(qs_env, matrix_p, matrix_ax, eps_filter)
674
675 DO ispin = 1, nspins
676
677 CALL dbcsr_filter(matrix_ax(ispin)%matrix, eps_filter)
678 ! norm_cA = tr(Ap_j * p_j)
679 CALL dbcsr_dot(matrix_cg(ispin)%matrix, matrix_ax(ispin)%matrix, norm_ca(ispin))
680
681 IF (norm_ca(ispin) < 0.0_dp) THEN
682
683 ! Recalculate w/o preconditioner
684 IF (i > 1) THEN
685 ! p = -z + beta*p
686 CALL dbcsr_add(matrix_cg(ispin)%matrix, matrix_z0(ispin)%matrix, &
687 beta(ispin), -1.0_dp)
688 CALL dbcsr_dot(matrix_res(ispin)%matrix, matrix_res(ispin)%matrix, new_norm(ispin))
689 beta(ispin) = new_norm(ispin)/tr_rz00(ispin)
690 CALL dbcsr_add(matrix_cg(ispin)%matrix, matrix_res(ispin)%matrix, &
691 beta(ispin), 1.0_dp)
692 norm_rr(ispin) = new_norm(ispin)
693 ELSE
694 CALL dbcsr_copy(matrix_res(ispin)%matrix, matrix_cg(ispin)%matrix)
695 CALL dbcsr_dot(matrix_res(ispin)%matrix, matrix_res(ispin)%matrix, norm_rr(ispin))
696 END IF
697
698 CALL build_hessian_op(qs_env=qs_env, &
699 p_env=p_env, &
700 matrix_ks=matrix_ks, &
701 matrix_p=matrix_p, & ! p
702 matrix_s_sqrt_inv=matrix_s_sqrt_inv, &
703 matrix_cg=matrix_cg, & ! cg
704 matrix_ax=matrix_ax, &
705 eps_filter=eps_filter)
706
707 ! Matrix projector T
708 CALL projector(qs_env, matrix_p, matrix_ax, eps_filter)
709
710 CALL dbcsr_dot(matrix_cg(ispin)%matrix, matrix_ax(ispin)%matrix, norm_ca(ispin))
711
712 cpabort("tr(Ap_j*p_j) < 0")
713 IF (abnormal_value(norm_ca(ispin))) &
714 cpabort("Preconditioner: Tr[Ap_j*p_j] is an abnormal value (NaN/Inf)")
715
716 END IF
717
718 END DO
719
720 DO ispin = 1, nspins
721 ! Determine step-size
722 IF (norm_ca(ispin) < linres_control%eps) THEN
723 alpha(ispin) = 1.0_dp
724 ELSE
725 alpha(ispin) = norm_rr(ispin)/norm_ca(ispin)
726 END IF
727
728 ! x_j+1 = x_j + alpha*p_j
729 ! save response-denisty of this iteration
730 CALL dbcsr_add(matrix_cg_z(ispin)%matrix, matrix_cg(ispin)%matrix, 1.0_dp, alpha(ispin))
731 END DO
732
733 ! need to recompute the residue
734 restart = .false.
735 IF (mod(i, linres_control%restart_every) == 0) THEN
736 !
737 ! r_j+1 = b - A * x_j+1
738 CALL build_hessian_op(qs_env=qs_env, &
739 p_env=p_env, &
740 matrix_ks=matrix_ks, &
741 matrix_p=matrix_p, &
742 matrix_s_sqrt_inv=matrix_s_sqrt_inv, &
743 matrix_cg=matrix_cg_z, & ! cg
744 matrix_ax=matrix_ax, &
745 eps_filter=eps_filter)
746 ! b
747 CALL commutator(matrix_nsc, matrix_p, matrix_res, eps_filter, .false., alpha=focc)
748
749 DO ispin = 1, nspins
750 CALL dbcsr_add(matrix_res(ispin)%matrix, matrix_ax(ispin)%matrix, 1.0_dp, -1.0_dp)
751 END DO
752
753 CALL projector(qs_env, matrix_p, matrix_res, eps_filter)
754 !
755 restart = .true.
756 ELSE
757 ! proj Ap onto the virtual subspace
758 CALL projector(qs_env, matrix_p, matrix_ax, eps_filter)
759 !
760 ! r_j+1 = r_j - alpha * Ap_j
761 DO ispin = 1, nspins
762 CALL dbcsr_add(matrix_res(ispin)%matrix, matrix_ax(ispin)%matrix, 1.0_dp, -alpha(ispin))
763 END DO
764 restart = .false.
765 END IF
766
767 ! Preconditioner
768 linres_control%flag = ""
769 IF (linres_control%preconditioner_type == precond_mlp) THEN
770 ! M * z_j+1 = r_j+1
771 ! Conjugate gradient returns z_j+1
772 CALL ec_preconditioner(qs_env=qs_env, &
773 matrix_ks=matrix_ks, &
774 matrix_p=matrix_p, &
775 matrix_rhs=matrix_res, &
776 matrix_cg_z=matrix_z0, &
777 eps_filter=eps_filter, &
778 iounit=iounit, silent=silent)
779 linres_control%flag = "PCG-AO"
780 ELSE
781 DO ispin = 1, nspins
782 CALL dbcsr_copy(matrix_z0(ispin)%matrix, matrix_res(ispin)%matrix)
783 END DO
784 linres_control%flag = "CG-AO"
785 END IF
786
787 norm_res = 0.0_dp
788
789 DO ispin = 1, nspins
790 ! Tr[r_j+1*z_j+1]
791 CALL dbcsr_dot(matrix_res(ispin)%matrix, matrix_z0(ispin)%matrix, new_norm(ispin))
792 IF (new_norm(ispin) < 0.0_dp) cpabort("tr(r_j+1*z_j+1) < 0")
793 IF (abnormal_value(new_norm(ispin))) &
794 cpabort("Preconditioner: Tr[r_j+1*z_j+1] is an abnormal value (NaN/Inf)")
795 norm_res = max(norm_res, new_norm(ispin)/real(nao, dp))
796
797 IF (norm_rr(ispin) < linres_control%eps .OR. new_norm(ispin) < linres_control%eps) THEN
798 beta(ispin) = 0.0_dp
799 linres_control%converged = .true.
800 ELSE
801 beta(ispin) = new_norm(ispin)/norm_rr(ispin)
802 END IF
803
804 ! update new search vector (matrix cg)
805 ! Here: cg_j+1 = z_j+1 + beta*cg_j
806 CALL dbcsr_add(matrix_cg(ispin)%matrix, matrix_z0(ispin)%matrix, beta(ispin), 1.0_dp)
807 CALL dbcsr_filter(matrix_cg(ispin)%matrix, eps_filter)
808
809 tr_rz00(ispin) = norm_rr(ispin)
810 norm_rr(ispin) = new_norm(ispin)
811 END DO
812
813 ! Can we exit the loop?
814 CALL external_control(should_stop, "LS_SOLVER", target_time=qs_env%target_time, &
815 start_time=qs_env%start_time)
816
817 END DO iteration
818
819 ! Matrix projector
820 CALL projector(qs_env, matrix_p, matrix_cg_z, eps_filter)
821
822 ! Z = [cg_z,P]
823 CALL commutator(matrix_cg_z, matrix_p, matrix_z, eps_filter, .true., alpha=0.5_dp)
824
825 DO ispin = 1, nspins
826 ! Transform Z-matrix back into non-orthogonal basis
827 CALL transform_m_orth(matrix_z(ispin)%matrix, matrix_s_sqrt_inv, eps_filter)
828
829 ! Export Z-Matrix
830 CALL dbcsr_copy(matrix_pz(ispin)%matrix, matrix_z(ispin)%matrix, keep_sparsity=.true.)
831 END DO
832
833 ! Calculate energy-weighted response density matrix
834 ! AO: Wz = 0.5*(Z*KS*P + P*KS*Z)
835 CALL ec_wz_matrix(qs_env, matrix_pz, matrix_wz, eps_filter)
836
837 ! Release matrices
838 CALL dbcsr_release(matrix_tmp)
839
840 CALL dbcsr_release(matrix_s_sqrt)
841 CALL dbcsr_release(matrix_s_sqrt_inv)
842
843 CALL dbcsr_deallocate_matrix_set(matrix_p)
844 CALL dbcsr_deallocate_matrix_set(matrix_ks)
845 CALL dbcsr_deallocate_matrix_set(matrix_nsc)
846 CALL dbcsr_deallocate_matrix_set(matrix_z)
847 CALL dbcsr_deallocate_matrix_set(matrix_ax)
848 CALL dbcsr_deallocate_matrix_set(matrix_res)
849 CALL dbcsr_deallocate_matrix_set(matrix_cg)
850 CALL dbcsr_deallocate_matrix_set(matrix_cg_z)
851 CALL dbcsr_deallocate_matrix_set(matrix_z0)
852
853 DEALLOCATE (alpha, beta, new_norm, norm_ca, norm_rr)
854 DEALLOCATE (tr_rz00)
855
856 CALL timestop(handle)
857
858 END SUBROUTINE ec_response_ao
859
860! **************************************************************************************************
861!> \brief Compute matrix_wz as needed for the forces
862!> Wz = 0.5*(Z*KS*P + P*KS*Z) (closed-shell)
863!> \param qs_env ...
864!> \param matrix_z The response density we just calculated
865!> \param matrix_wz The energy weighted response-density matrix
866!> \param eps_filter ...
867!> \par History
868!> 2020.2 created [Fabian Belleflamme]
869!> \author Fabian Belleflamme
870! **************************************************************************************************
871 SUBROUTINE ec_wz_matrix(qs_env, matrix_z, matrix_wz, eps_filter)
872
873 TYPE(qs_environment_type), POINTER :: qs_env
874 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
875 POINTER :: matrix_z
876 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(INOUT), &
877 POINTER :: matrix_wz
878 REAL(kind=dp), INTENT(IN) :: eps_filter
879
880 CHARACTER(len=*), PARAMETER :: routinen = 'ec_wz_matrix'
881
882 INTEGER :: handle, ispin, nspins
883 REAL(kind=dp) :: scaling
884 TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: matrix_ks, matrix_p, matrix_s
885 TYPE(dbcsr_type) :: matrix_tmp, matrix_tmp2
886 TYPE(dft_control_type), POINTER :: dft_control
887 TYPE(qs_rho_type), POINTER :: rho
888
889 CALL timeset(routinen, handle)
890
891 cpassert(ASSOCIATED(qs_env))
892 cpassert(ASSOCIATED(matrix_z))
893 cpassert(ASSOCIATED(matrix_wz))
894
895 CALL get_qs_env(qs_env=qs_env, &
896 dft_control=dft_control, &
897 matrix_ks=matrix_ks, &
898 matrix_s=matrix_s, &
899 rho=rho)
900 nspins = dft_control%nspins
901
902 CALL qs_rho_get(rho, rho_ao=matrix_p)
903
904 ! Init temp matrices
905 CALL dbcsr_create(matrix_tmp, template=matrix_z(1)%matrix, &
906 matrix_type=dbcsr_type_no_symmetry)
907 CALL dbcsr_create(matrix_tmp2, template=matrix_z(1)%matrix, &
908 matrix_type=dbcsr_type_no_symmetry)
909
910 ! Scale matrix_p by factor 1/2 in closed-shell
911 scaling = 1.0_dp
912 IF (nspins == 1) scaling = 0.5_dp
913
914 ! Whz = ZFP + PFZ = Z(FP) + (Z(FP))^T
915 DO ispin = 1, nspins
916
917 ! tmp = FP
918 CALL dbcsr_multiply("N", "N", scaling, matrix_ks(ispin)%matrix, matrix_p(ispin)%matrix, &
919 0.0_dp, matrix_tmp, filter_eps=eps_filter, retain_sparsity=.false.)
920
921 ! tmp2 = ZFP
922 CALL dbcsr_multiply("N", "N", 1.0_dp, matrix_z(ispin)%matrix, matrix_tmp, &
923 0.0_dp, matrix_tmp2, filter_eps=eps_filter, retain_sparsity=.false.)
924
925 ! tmp = (ZFP)^T
926 CALL dbcsr_transposed(matrix_tmp, matrix_tmp2)
927
928 ! tmp = ZFP + (ZFP)^T
929 CALL dbcsr_add(matrix_tmp, matrix_tmp2, 1.0_dp, 1.0_dp)
930
931 CALL dbcsr_filter(matrix_tmp, eps_filter)
932
933 ! Whz = ZFP + PFZ
934 CALL dbcsr_copy(matrix_wz(ispin)%matrix, matrix_tmp, keep_sparsity=.true.)
935
936 END DO
937
938 ! Release matrices
939 CALL dbcsr_release(matrix_tmp)
940 CALL dbcsr_release(matrix_tmp2)
941
942 CALL timestop(handle)
943
944 END SUBROUTINE ec_wz_matrix
945
946! **************************************************************************************************
947!> \brief Calculate first term of electronic Hessian M = [F, B]
948!> acting as liner transformation on trial matrix (matrix_cg)
949!> with intermediate response density B = [cg,P] = cg*P - P*cg = cg*P + (cg*P)^T
950!>
951!> All matrices are in orthonormal basis
952!>
953!> \param matrix_ks Ground-state Kohn-Sham matrix
954!> \param matrix_p Ground-state Density matrix
955!> \param matrix_cg Trial matrix
956!> \param matrix_b Intermediate response density
957!> \param matrix_Ax First term of electronic Hessian applied on trial matrix (matrix_cg)
958!>
959!> \param eps_filter ...
960!> \date 12.2019
961!> \author Fabian Belleflamme
962! **************************************************************************************************
963 SUBROUTINE hessian_op1(matrix_ks, matrix_p, matrix_cg, matrix_b, matrix_Ax, eps_filter)
964
965 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
966 POINTER :: matrix_ks, matrix_p, matrix_cg
967 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(INOUT), &
968 POINTER :: matrix_b, matrix_ax
969 REAL(kind=dp), INTENT(IN) :: eps_filter
970
971 CHARACTER(len=*), PARAMETER :: routinen = 'hessian_op1'
972
973 INTEGER :: handle
974
975 CALL timeset(routinen, handle)
976
977 cpassert(ASSOCIATED(matrix_ks))
978 cpassert(ASSOCIATED(matrix_p))
979 cpassert(ASSOCIATED(matrix_cg))
980 cpassert(ASSOCIATED(matrix_b))
981 cpassert(ASSOCIATED(matrix_ax))
982
983 ! Build intermediate density matrix
984 ! B = [cg, P] = cg*P - P*cg = cg*P + (cg*P)^T
985 CALL commutator(matrix_cg, matrix_p, matrix_b, eps_filter, .true.)
986
987 ! Build first part of operator
988 ! Ax = [F,[cg,P]] = [F,B]
989 CALL commutator(matrix_ks, matrix_b, matrix_ax, eps_filter, .false.)
990
991 CALL timestop(handle)
992
993 END SUBROUTINE hessian_op1
994
995! **************************************************************************************************
996!> \brief calculate linear transformation of Hessian matrix on a trial matrix matrix_cg
997!> which is stored in response density B = [cg,P] = cg*P - P*cg = cg*P + (cg*P)^T
998!> Ax = [F, B] + [G(B), Pin] in orthonormal basis
999!>
1000!> \param qs_env ...
1001!> \param p_env ...
1002!> \param matrix_ks Ground-state Kohn-Sham matrix
1003!> \param matrix_p Ground-state Density matrix
1004!> \param matrix_s_sqrt_inv S^(-1/2) needed for transformation to/from orthonormal basis
1005!> \param matrix_cg Trial matrix
1006!> \param matrix_Ax Electronic Hessian applied on trial matrix (matrix_cg)
1007!> \param eps_filter ...
1008!>
1009!> \date 12.2019
1010!> \author Fabian Belleflamme
1011! **************************************************************************************************
1012 SUBROUTINE build_hessian_op(qs_env, p_env, matrix_ks, matrix_p, matrix_s_sqrt_inv, &
1013 matrix_cg, matrix_Ax, eps_filter)
1014
1015 TYPE(qs_environment_type), POINTER :: qs_env
1016 TYPE(qs_p_env_type), POINTER :: p_env
1017 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
1018 POINTER :: matrix_ks, matrix_p
1019 TYPE(dbcsr_type), INTENT(IN) :: matrix_s_sqrt_inv
1020 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
1021 POINTER :: matrix_cg
1022 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(INOUT), &
1023 POINTER :: matrix_ax
1024 REAL(kind=dp), INTENT(IN) :: eps_filter
1025
1026 CHARACTER(len=*), PARAMETER :: routinen = 'build_hessian_op'
1027
1028 INTEGER :: handle, ispin, nspins
1029 REAL(kind=dp) :: chksum
1030 TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: matrix_b, rho1_ao
1031 TYPE(dft_control_type), POINTER :: dft_control
1032 TYPE(mp_para_env_type), POINTER :: para_env
1033 TYPE(qs_rho_type), POINTER :: rho
1034
1035 CALL timeset(routinen, handle)
1036
1037 cpassert(ASSOCIATED(qs_env))
1038 cpassert(ASSOCIATED(matrix_ks))
1039 cpassert(ASSOCIATED(matrix_p))
1040 cpassert(ASSOCIATED(matrix_cg))
1041 cpassert(ASSOCIATED(matrix_ax))
1042
1043 CALL get_qs_env(qs_env=qs_env, &
1044 dft_control=dft_control, &
1045 para_env=para_env, &
1046 rho=rho)
1047 nspins = dft_control%nspins
1048
1049 NULLIFY (matrix_b)
1050 CALL dbcsr_allocate_matrix_set(matrix_b, nspins)
1051 DO ispin = 1, nspins
1052 ALLOCATE (matrix_b(ispin)%matrix)
1053 CALL dbcsr_create(matrix_b(ispin)%matrix, name="[X,P] RSP DNSTY", &
1054 template=matrix_p(1)%matrix, &
1055 matrix_type=dbcsr_type_no_symmetry)
1056 END DO
1057
1058 ! Build uncoupled term of Hessian linear transformation
1059 CALL hessian_op1(matrix_ks, matrix_p, matrix_cg, matrix_b, matrix_ax, eps_filter)
1060
1061 ! Avoid the buildup of noisy blocks
1062 DO ispin = 1, nspins
1063 CALL dbcsr_filter(matrix_b(ispin)%matrix, eps_filter)
1064 END DO
1065
1066 chksum = 0.0_dp
1067 DO ispin = 1, nspins
1068 chksum = chksum + dbcsr_checksum(matrix_b(ispin)%matrix)
1069 END DO
1070
1071 ! skip the kernel if the DM is very small
1072 IF (chksum > 1.0e-14_dp) THEN
1073
1074 ! Bring matrix B as density on grid
1075
1076 ! prepare perturbation environment
1077 CALL p_env_check_i_alloc(p_env, qs_env)
1078
1079 ! Get response density matrix
1080 CALL qs_rho_get(p_env%rho1, rho_ao=rho1_ao)
1081
1082 DO ispin = 1, nspins
1083 ! Transform B into NON-ortho basis for collocation
1084 CALL transform_m_orth(matrix_b(ispin)%matrix, matrix_s_sqrt_inv, eps_filter)
1085 ! Filter
1086 CALL dbcsr_filter(matrix_b(ispin)%matrix, eps_filter)
1087 ! Keep symmetry of density matrix
1088 CALL dbcsr_copy(rho1_ao(ispin)%matrix, matrix_b(ispin)%matrix, keep_sparsity=.true.)
1089 CALL dbcsr_copy(p_env%p1(ispin)%matrix, matrix_b(ispin)%matrix, keep_sparsity=.true.)
1090 END DO
1091
1092 ! Updates densities on grid wrt density matrix
1093 CALL p_env_update_rho(p_env, qs_env)
1094
1095 DO ispin = 1, nspins
1096 CALL dbcsr_set(p_env%kpp1(ispin)%matrix, 0.0_dp)
1097 IF (ASSOCIATED(p_env%kpp1_admm)) CALL dbcsr_set(p_env%kpp1_admm(ispin)%matrix, 0.0_dp)
1098 END DO
1099
1100 ! Calculate kernel
1101 ! Ax = F*B - B*F + G(B)*P - P*G(B)
1102 ! IN/OUT IN IN IN
1103 CALL hessian_op2(qs_env, p_env, matrix_ax, matrix_p, matrix_s_sqrt_inv, eps_filter)
1104
1105 END IF
1106
1107 CALL dbcsr_deallocate_matrix_set(matrix_b)
1108
1109 CALL timestop(handle)
1110
1111 END SUBROUTINE build_hessian_op
1112
1113! **************************************************************************************************
1114!> \brief Calculate lin transformation of Hessian matrix on a trial matrix matrix_cg
1115!> which is stored in response density B = [cg,P] = cg*P - P*cg = cg*P + (cg*P)^T
1116!> Ax = [F, B] + [G(B), Pin] in orthonormal basis
1117!>
1118!> \param qs_env ...
1119!> \param p_env p-environment with trial density environment
1120!> \param matrix_Ax contains first part of Hessian linear transformation, kernel contribution
1121!> is calculated and added in this routine
1122!> \param matrix_p Density matrix in orthogonal basis
1123!> \param matrix_s_sqrt_inv contains matrix S^(-1/2) for switching to orthonormal Lowdin basis
1124!> \param eps_filter ...
1125!>
1126!> \date 12.2019
1127!> \author Fabian Belleflamme
1128! **************************************************************************************************
1129 SUBROUTINE hessian_op2(qs_env, p_env, matrix_Ax, matrix_p, matrix_s_sqrt_inv, eps_filter)
1130
1131 TYPE(qs_environment_type), POINTER :: qs_env
1132 TYPE(qs_p_env_type), POINTER :: p_env
1133 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(INOUT), &
1134 POINTER :: matrix_ax
1135 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
1136 POINTER :: matrix_p
1137 TYPE(dbcsr_type), INTENT(IN) :: matrix_s_sqrt_inv
1138 REAL(kind=dp), INTENT(IN) :: eps_filter
1139
1140 CHARACTER(len=*), PARAMETER :: routinen = 'hessian_op2'
1141
1142 INTEGER :: handle, ispin, nspins
1143 REAL(kind=dp) :: ekin_mol
1144 TYPE(admm_type), POINTER :: admm_env
1145 TYPE(dbcsr_p_type), DIMENSION(:), POINTER :: matrix_g, matrix_s, rho1_ao, rho_ao
1146 TYPE(dft_control_type), POINTER :: dft_control
1147 TYPE(mp_para_env_type), POINTER :: para_env
1148 TYPE(pw_c1d_gs_type) :: rho_tot_gspace, v_hartree_gspace
1149 TYPE(pw_c1d_gs_type), DIMENSION(:), POINTER :: rho1_g
1150 TYPE(pw_env_type), POINTER :: pw_env
1151 TYPE(pw_poisson_type), POINTER :: poisson_env
1152 TYPE(pw_pool_p_type), DIMENSION(:), POINTER :: pw_pools
1153 TYPE(pw_pool_type), POINTER :: auxbas_pw_pool
1154 TYPE(pw_r3d_rs_type) :: v_hartree_rspace
1155 TYPE(pw_r3d_rs_type), DIMENSION(:), POINTER :: rho1_r, rho_r, tau1_r, v_xc, v_xc_tau
1156 TYPE(pw_r3d_rs_type), POINTER :: weights
1157 TYPE(qs_kpp1_env_type), POINTER :: kpp1_env
1158 TYPE(qs_rho_type), POINTER :: rho, rho_aux
1159 TYPE(section_vals_type), POINTER :: input, xc_section, xc_section_aux
1160
1161 CALL timeset(routinen, handle)
1162
1163 NULLIFY (admm_env, dft_control, input, matrix_s, para_env, rho, rho_r, rho1_g, rho1_r)
1164
1165 CALL get_qs_env(qs_env=qs_env, &
1166 admm_env=admm_env, &
1167 dft_control=dft_control, &
1168 input=input, &
1169 matrix_s=matrix_s, &
1170 para_env=para_env, &
1171 rho=rho)
1172 nspins = dft_control%nspins
1173
1174 cpassert(ASSOCIATED(p_env%kpp1))
1175 cpassert(ASSOCIATED(p_env%kpp1_env))
1176 kpp1_env => p_env%kpp1_env
1177
1178 ! Get non-ortho input density matrix on grid
1179 CALL qs_rho_get(rho, rho_ao=rho_ao)
1180 ! Get non-ortho trial density stored in p_env
1181 CALL qs_rho_get(p_env%rho1, rho_g=rho1_g, rho_r=rho1_r, tau_r=tau1_r)
1182
1183 NULLIFY (pw_env)
1184 CALL get_qs_env(qs_env, pw_env=pw_env)
1185 cpassert(ASSOCIATED(pw_env))
1186
1187 NULLIFY (weights)
1188 CALL get_qs_env(qs_env, xcint_weights=weights)
1189
1190 NULLIFY (auxbas_pw_pool, poisson_env, pw_pools)
1191 ! gets the tmp grids
1192 CALL pw_env_get(pw_env=pw_env, &
1193 auxbas_pw_pool=auxbas_pw_pool, &
1194 pw_pools=pw_pools, &
1195 poisson_env=poisson_env)
1196
1197 ! Calculate the NSC Hartree potential
1198 CALL auxbas_pw_pool%create_pw(pw=v_hartree_gspace)
1199 CALL auxbas_pw_pool%create_pw(pw=rho_tot_gspace)
1200 CALL auxbas_pw_pool%create_pw(pw=v_hartree_rspace)
1201
1202 ! XC-Kernel
1203 NULLIFY (v_xc, v_xc_tau, xc_section)
1204
1205 IF (dft_control%do_admm) THEN
1206 xc_section => admm_env%xc_section_primary
1207 ELSE
1208 xc_section => section_vals_get_subs_vals(input, "DFT%XC")
1209 END IF
1210
1211 ! add xc-kernel
1212 CALL create_kernel(qs_env, &
1213 vxc=v_xc, &
1214 vxc_tau=v_xc_tau, &
1215 rho=rho, &
1216 rho1_r=rho1_r, &
1217 rho1_g=rho1_g, &
1218 tau1_r=tau1_r, &
1219 xc_section=xc_section)
1220
1221 DO ispin = 1, nspins
1222 CALL pw_scale(v_xc(ispin), v_xc(ispin)%pw_grid%dvol)
1223 IF (ASSOCIATED(v_xc_tau)) THEN
1224 CALL pw_scale(v_xc_tau(ispin), v_xc_tau(ispin)%pw_grid%dvol)
1225 END IF
1226 END DO
1227
1228 ! ADMM Correction
1229 IF (dft_control%do_admm) THEN
1230 IF (admm_env%aux_exch_func /= do_admm_aux_exch_func_none) THEN
1231 IF (.NOT. ASSOCIATED(kpp1_env%deriv_set_admm)) THEN
1232 xc_section_aux => admm_env%xc_section_aux
1233 CALL get_admm_env(qs_env%admm_env, rho_aux_fit=rho_aux)
1234 CALL qs_rho_get(rho_aux, rho_r=rho_r)
1235 ALLOCATE (kpp1_env%deriv_set_admm, kpp1_env%rho_set_admm)
1236 CALL xc_prep_2nd_deriv(kpp1_env%deriv_set_admm, kpp1_env%rho_set_admm, &
1237 rho_r, auxbas_pw_pool, weights, &
1238 xc_section=xc_section_aux)
1239 END IF
1240 END IF
1241 END IF
1242
1243 ! take trial density to build G^{H}[B]
1244 CALL pw_zero(rho_tot_gspace)
1245 DO ispin = 1, nspins
1246 CALL pw_axpy(rho1_g(ispin), rho_tot_gspace)
1247 END DO
1248
1249 ! get Hartree potential from rho_tot_gspace
1250 CALL pw_poisson_solve(poisson_env, rho_tot_gspace, &
1251 vhartree=v_hartree_gspace)
1252 CALL pw_transfer(v_hartree_gspace, v_hartree_rspace)
1253 CALL pw_scale(v_hartree_rspace, v_hartree_rspace%pw_grid%dvol)
1254
1255 ! Add v_xc + v_H
1256 DO ispin = 1, nspins
1257 CALL pw_axpy(v_hartree_rspace, v_xc(ispin))
1258 END DO
1259 IF (nspins == 1) THEN
1260 CALL pw_scale(v_xc(1), 2.0_dp)
1261 IF (ASSOCIATED(v_xc_tau)) CALL pw_scale(v_xc_tau(1), 2.0_dp)
1262 END IF
1263
1264 DO ispin = 1, nspins
1265 ! Integrate with ground-state density matrix, in non-orthogonal basis
1266 CALL integrate_v_rspace(v_rspace=v_xc(ispin), &
1267 pmat=rho_ao(ispin), &
1268 hmat=p_env%kpp1(ispin), &
1269 qs_env=qs_env, &
1270 calculate_forces=.false., &
1271 basis_type="ORB")
1272 IF (ASSOCIATED(v_xc_tau)) THEN
1273 CALL integrate_v_rspace(v_rspace=v_xc_tau(ispin), &
1274 pmat=rho_ao(ispin), &
1275 hmat=p_env%kpp1(ispin), &
1276 qs_env=qs_env, &
1277 compute_tau=.true., &
1278 calculate_forces=.false., &
1279 basis_type="ORB")
1280 END IF
1281 END DO
1282
1283 ! Hartree-Fock contribution
1284 CALL apply_hfx(qs_env, p_env)
1285 ! Calculate ADMM exchange correction to kernel
1286 CALL apply_xc_admm(qs_env, p_env)
1287 ! Add contribution from ADMM exchange correction to kernel
1288 CALL p_env_finish_kpp1(qs_env, p_env)
1289
1290 ! Calculate KG correction to kernel
1291 IF (dft_control%qs_control%do_kg) THEN
1292 IF (qs_env%kg_env%tnadd_method == kg_tnadd_embed .OR. &
1293 qs_env%kg_env%tnadd_method == kg_tnadd_embed_ri) THEN
1294
1295 cpassert(dft_control%nimages == 1)
1296 ekin_mol = 0.0_dp
1297 CALL qs_rho_get(p_env%rho1, rho_ao=rho1_ao)
1298 CALL kg_ekin_subset(qs_env=qs_env, &
1299 ks_matrix=p_env%kpp1, &
1300 ekin_mol=ekin_mol, &
1301 calc_force=.false., &
1302 do_kernel=.true., &
1303 pmat_ext=rho1_ao)
1304 END IF
1305 END IF
1306
1307 ! Init response kernel matrix
1308 ! matrix G(B)
1309 NULLIFY (matrix_g)
1310 CALL dbcsr_allocate_matrix_set(matrix_g, nspins)
1311 DO ispin = 1, nspins
1312 ALLOCATE (matrix_g(ispin)%matrix)
1313 CALL dbcsr_copy(matrix_g(ispin)%matrix, p_env%kpp1(ispin)%matrix, &
1314 name="MATRIX Kernel")
1315 END DO
1316
1317 ! Transforming G(B) into orthonormal basis
1318 ! Careful, this de-symmetrizes matrix_G
1319 DO ispin = 1, nspins
1320 CALL transform_m_orth(matrix_g(ispin)%matrix, matrix_s_sqrt_inv, eps_filter)
1321 CALL dbcsr_filter(matrix_g(ispin)%matrix, eps_filter)
1322 END DO
1323
1324 ! Hessian already contains Ax = [F,B] (ORTHO), now adding
1325 ! Ax = Ax + G(B)P - (G(B)P)^T
1326 CALL commutator(matrix_g, matrix_p, matrix_ax, eps_filter, .false., 1.0_dp, 1.0_dp)
1327
1328 ! release pw grids
1329 CALL auxbas_pw_pool%give_back_pw(v_hartree_gspace)
1330 CALL auxbas_pw_pool%give_back_pw(v_hartree_rspace)
1331 CALL auxbas_pw_pool%give_back_pw(rho_tot_gspace)
1332 DO ispin = 1, nspins
1333 CALL auxbas_pw_pool%give_back_pw(v_xc(ispin))
1334 END DO
1335 DEALLOCATE (v_xc)
1336 IF (ASSOCIATED(v_xc_tau)) THEN
1337 DO ispin = 1, nspins
1338 CALL auxbas_pw_pool%give_back_pw(v_xc_tau(ispin))
1339 END DO
1340 DEALLOCATE (v_xc_tau)
1341 END IF
1342
1343 CALL dbcsr_deallocate_matrix_set(matrix_g)
1344
1345 CALL timestop(handle)
1346
1347 END SUBROUTINE hessian_op2
1348
1349! **************************************************************************************************
1350!> \brief computes (anti-)commutator exploiting (anti-)symmetry:
1351!> A symmetric : RES = beta*RES + k*[A,B] = k*(AB-(AB)^T)
1352!> A anti-sym : RES = beta*RES + k*{A,B} = k*(AB+(AB)^T)
1353!>
1354!> \param a Matrix A
1355!> \param b Matrix B
1356!> \param res Commutator result
1357!> \param eps_filter filtering threshold for sparse matrices
1358!> \param anticomm Calculate anticommutator
1359!> \param alpha Scaling of anti-/commutator
1360!> \param beta Scaling of inital content of result matrix
1361!>
1362!> \par History
1363!> 2020.07 Fabian Belleflamme (based on commutator_symm)
1364! **************************************************************************************************
1365 SUBROUTINE commutator(a, b, res, eps_filter, anticomm, alpha, beta)
1366
1367 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
1368 POINTER :: a, b, res
1369 REAL(kind=dp) :: eps_filter
1370 LOGICAL :: anticomm
1371 REAL(kind=dp), OPTIONAL :: alpha, beta
1372
1373 CHARACTER(LEN=*), PARAMETER :: routinen = 'commutator'
1374
1375 INTEGER :: handle, ispin
1376 REAL(kind=dp) :: facc, myalpha, mybeta
1377 TYPE(dbcsr_type) :: work, work2
1378
1379 CALL timeset(routinen, handle)
1380
1381 cpassert(ASSOCIATED(a))
1382 cpassert(ASSOCIATED(b))
1383 cpassert(ASSOCIATED(res))
1384
1385 CALL dbcsr_create(work, template=a(1)%matrix, matrix_type=dbcsr_type_no_symmetry)
1386 CALL dbcsr_create(work2, template=a(1)%matrix, matrix_type=dbcsr_type_no_symmetry)
1387
1388 ! Scaling of anti-/commutator
1389 myalpha = 1.0_dp
1390 IF (PRESENT(alpha)) myalpha = alpha
1391 ! Scaling of result matrix
1392 mybeta = 0.0_dp
1393 IF (PRESENT(beta)) mybeta = beta
1394 ! Add/subtract second term when calculating anti-/commutator
1395 facc = -1.0_dp
1396 IF (anticomm) facc = 1.0_dp
1397
1398 DO ispin = 1, SIZE(a)
1399
1400 CALL dbcsr_multiply("N", "N", myalpha, a(ispin)%matrix, b(ispin)%matrix, &
1401 0.0_dp, work, filter_eps=eps_filter)
1402 CALL dbcsr_transposed(work2, work)
1403
1404 ! RES= beta*RES + alpha*{A,B} = beta*RES + alpha*[AB+(AB)T]
1405 ! RES= beta*RES + alpha*[A,B] = beta*RES + alpha*[AB-(AB)T]
1406 CALL dbcsr_add(work, work2, 1.0_dp, facc)
1407
1408 CALL dbcsr_add(res(ispin)%matrix, work, mybeta, 1.0_dp)
1409
1410 END DO
1411
1412 CALL dbcsr_release(work)
1413 CALL dbcsr_release(work2)
1414
1415 CALL timestop(handle)
1416
1417 END SUBROUTINE commutator
1418
1419! **************************************************************************************************
1420!> \brief Projector P(M) = P*M*Q^T + Q*M*P^T
1421!> with P = D
1422!> with Q = (1-D)
1423!>
1424!> \param qs_env ...
1425!> \param matrix_p Ground-state density in orthonormal basis
1426!> \param matrix_io Matrix to which projector is applied.
1427!>
1428!> \param eps_filter ...
1429!> \date 06.2020
1430!> \author Fabian Belleflamme
1431! **************************************************************************************************
1432 SUBROUTINE projector(qs_env, matrix_p, matrix_io, eps_filter)
1433
1434 TYPE(qs_environment_type), POINTER :: qs_env
1435 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(IN), &
1436 POINTER :: matrix_p
1437 TYPE(dbcsr_p_type), DIMENSION(:), INTENT(INOUT), &
1438 POINTER :: matrix_io
1439 REAL(kind=dp), INTENT(IN) :: eps_filter
1440
1441 CHARACTER(len=*), PARAMETER :: routinen = 'projector'
1442
1443 INTEGER :: handle, ispin, nspins
1444 TYPE(dbcsr_type) :: matrix_q, matrix_tmp
1445 TYPE(dft_control_type), POINTER :: dft_control
1446 TYPE(mp_para_env_type), POINTER :: para_env
1447
1448 CALL timeset(routinen, handle)
1449
1450 CALL get_qs_env(qs_env=qs_env, &
1451 dft_control=dft_control, &
1452 para_env=para_env)
1453 nspins = dft_control%nspins
1454
1455 CALL dbcsr_create(matrix_q, template=matrix_p(1)%matrix, &
1456 matrix_type=dbcsr_type_no_symmetry)
1457 CALL dbcsr_create(matrix_tmp, template=matrix_p(1)%matrix, &
1458 matrix_type=dbcsr_type_no_symmetry)
1459
1460 ! Q = (1 - P)
1461 CALL dbcsr_copy(matrix_q, matrix_p(1)%matrix)
1462 CALL dbcsr_scale(matrix_q, -1.0_dp)
1463 CALL dbcsr_add_on_diag(matrix_q, 1.0_dp)
1464 CALL dbcsr_finalize(matrix_q)
1465
1466 ! Proj(M) = P*M*Q + Q*M*P
1467 ! with P = D = CC^T
1468 ! and Q = (1 - P)
1469 DO ispin = 1, nspins
1470
1471 ! tmp1 = P*M
1472 CALL dbcsr_multiply("N", "N", 1.0_dp, matrix_p(ispin)%matrix, matrix_io(ispin)%matrix, &
1473 0.0_dp, matrix_tmp, filter_eps=eps_filter)
1474 ! m_io = P*M*Q
1475 CALL dbcsr_multiply("N", "N", 1.0_dp, matrix_tmp, matrix_q, &
1476 0.0_dp, matrix_io(ispin)%matrix, filter_eps=eps_filter)
1477
1478 ! tmp = (P^T*M^T*Q^T)^T = -(P*M*Q)^T
1479 CALL dbcsr_transposed(matrix_tmp, matrix_io(ispin)%matrix)
1480 CALL dbcsr_add(matrix_io(ispin)%matrix, matrix_tmp, 1.0_dp, -1.0_dp)
1481
1482 END DO
1483
1484 CALL dbcsr_release(matrix_tmp)
1485 CALL dbcsr_release(matrix_q)
1486
1487 CALL timestop(handle)
1488
1489 END SUBROUTINE projector
1490
1491! **************************************************************************************************
1492!> \brief performs a tranformation of a matrix back to/into orthonormal basis
1493!> in case of P a scaling of 0.5 has to be applied for closed shell case
1494!> \param matrix matrix to be transformed
1495!> \param matrix_trafo transformation matrix
1496!> \param eps_filter filtering threshold for sparse matrices
1497!> \par History
1498!> 2012.05 created [Florian Schiffmann]
1499!> \author Florian Schiffmann
1500!>
1501! **************************************************************************************************
1502
1503 SUBROUTINE transform_m_orth(matrix, matrix_trafo, eps_filter)
1504 TYPE(dbcsr_type) :: matrix, matrix_trafo
1505 REAL(kind=dp) :: eps_filter
1506
1507 CHARACTER(LEN=*), PARAMETER :: routinen = 'transform_m_orth'
1508
1509 INTEGER :: handle
1510 TYPE(dbcsr_type) :: matrix_tmp, matrix_work
1511
1512 CALL timeset(routinen, handle)
1513
1514 CALL dbcsr_create(matrix_work, template=matrix, matrix_type=dbcsr_type_no_symmetry)
1515 CALL dbcsr_create(matrix_tmp, template=matrix, matrix_type=dbcsr_type_no_symmetry)
1516
1517 CALL dbcsr_multiply("N", "N", 1.0_dp, matrix, matrix_trafo, &
1518 0.0_dp, matrix_work, filter_eps=eps_filter)
1519 CALL dbcsr_multiply("N", "N", 1.0_dp, matrix_trafo, matrix_work, &
1520 0.0_dp, matrix_tmp, filter_eps=eps_filter)
1521 ! symmetrize results (this is again needed to make sure everything is stable)
1522 CALL dbcsr_transposed(matrix_work, matrix_tmp)
1523 CALL dbcsr_add(matrix_tmp, matrix_work, 0.5_dp, 0.5_dp)
1524 CALL dbcsr_copy(matrix, matrix_tmp)
1525
1526 ! Avoid the buildup of noisy blocks
1527 CALL dbcsr_filter(matrix, eps_filter)
1528
1529 CALL dbcsr_release(matrix_tmp)
1530 CALL dbcsr_release(matrix_work)
1531 CALL timestop(handle)
1532
1533 END SUBROUTINE transform_m_orth
1534
1535END MODULE ec_orth_solver
cp_dbcsr_api::dbcsr_create
Definition cp_dbcsr_api.F:194
cp_dbcsr_operations::dbcsr_allocate_matrix_set
Definition cp_dbcsr_operations.F:77
cp_dbcsr_operations::dbcsr_deallocate_matrix_set
Definition cp_dbcsr_operations.F:85
pw_methods::pw_axpy
Definition pw_methods.F:165
pw_methods::pw_scale
Definition pw_methods.F:94
pw_methods::pw_transfer
Definition pw_methods.F:304
pw_methods::pw_zero
Definition pw_methods.F:83
pw_poisson_methods::pw_poisson_solve
Definition pw_poisson_methods.F:78
admm_types
Types and set/get functions for auxiliary density matrix methods.
Definition admm_types.F:15
admm_types::get_admm_env
subroutine, public get_admm_env(admm_env, mo_derivs_aux_fit, mos_aux_fit, sab_aux_fit, sab_aux_fit_asymm, sab_aux_fit_vs_orb, matrix_s_aux_fit, matrix_s_aux_fit_kp, matrix_s_aux_fit_vs_orb, matrix_s_aux_fit_vs_orb_kp, task_list_aux_fit, matrix_ks_aux_fit, matrix_ks_aux_fit_kp, matrix_ks_aux_fit_im, matrix_ks_aux_fit_dft, matrix_ks_aux_fit_hfx, matrix_ks_aux_fit_dft_kp, matrix_ks_aux_fit_hfx_kp, rho_aux_fit, rho_aux_fit_buffer, admm_dm)
Get routine for the ADMM env.
Definition admm_types.F:593
cp_control_types
Defines control structures, which contain the parameters and the settings for the DFT-based calculati...
Definition cp_control_types.F:12
cp_dbcsr_api
Definition cp_dbcsr_api.F:8
cp_dbcsr_api::dbcsr_transposed
subroutine, public dbcsr_transposed(transposed, normal, shallow_data_copy, transpose_distribution, use_distribution)
...
Definition cp_dbcsr_api.F:1232
cp_dbcsr_api::dbcsr_scale
subroutine, public dbcsr_scale(matrix, alpha_scalar)
...
Definition cp_dbcsr_api.F:1178
cp_dbcsr_api::dbcsr_desymmetrize
subroutine, public dbcsr_desymmetrize(matrix_a, matrix_b)
...
Definition cp_dbcsr_api.F:524
cp_dbcsr_api::dbcsr_copy
subroutine, public dbcsr_copy(matrix_b, matrix_a, name, keep_sparsity, keep_imaginary)
...
Definition cp_dbcsr_api.F:370
cp_dbcsr_api::dbcsr_multiply
subroutine, public dbcsr_multiply(transa, transb, alpha, matrix_a, matrix_b, beta, matrix_c, first_row, last_row, first_column, last_column, first_k, last_k, retain_sparsity, filter_eps, flop)
...
Definition cp_dbcsr_api.F:1086
cp_dbcsr_api::dbcsr_get_info
subroutine, public dbcsr_get_info(matrix, nblkrows_total, nblkcols_total, nfullrows_total, nfullcols_total, nblkrows_local, nblkcols_local, nfullrows_local, nfullcols_local, my_prow, my_pcol, local_rows, local_cols, proc_row_dist, proc_col_dist, row_blk_size, col_blk_size, row_blk_offset, col_blk_offset, distribution, name, matrix_type, group)
...
Definition cp_dbcsr_api.F:807
cp_dbcsr_api::dbcsr_filter
subroutine, public dbcsr_filter(matrix, eps)
...
Definition cp_dbcsr_api.F:657
cp_dbcsr_api::dbcsr_finalize
subroutine, public dbcsr_finalize(matrix)
...
Definition cp_dbcsr_api.F:672
cp_dbcsr_api::dbcsr_set
subroutine, public dbcsr_set(matrix, alpha)
...
Definition cp_dbcsr_api.F:1194
cp_dbcsr_api::dbcsr_release
subroutine, public dbcsr_release(matrix)
...
Definition cp_dbcsr_api.F:1132
cp_dbcsr_api::dbcsr_add
subroutine, public dbcsr_add(matrix_a, matrix_b, alpha_scalar, beta_scalar)
...
Definition cp_dbcsr_api.F:253
cp_dbcsr_contrib
Definition cp_dbcsr_contrib.F:8
cp_dbcsr_contrib::dbcsr_checksum
real(kind=dp) function, public dbcsr_checksum(matrix, pos)
Calculates the checksum of a DBCSR matrix.
Definition cp_dbcsr_contrib.F:628
cp_dbcsr_contrib::dbcsr_add_on_diag
subroutine, public dbcsr_add_on_diag(matrix, alpha)
Adds the given scalar to the diagonal of the matrix. Reserves any missing diagonal blocks.
Definition cp_dbcsr_contrib.F:331
cp_dbcsr_contrib::dbcsr_dot
subroutine, public dbcsr_dot(matrix_a, matrix_b, trace)
Computes the dot product of two matrices, also known as the trace of their matrix product.
Definition cp_dbcsr_contrib.F:367
cp_dbcsr_operations
DBCSR operations in CP2K.
Definition cp_dbcsr_operations.F:18
cp_external_control
Routines to handle the external control of CP2K.
Definition cp_external_control.F:15
cp_external_control::external_control
subroutine, public external_control(should_stop, flag, globenv, target_time, start_time, force_check)
External manipulations during a run : when the <PROJECT_NAME>.EXIT_$runtype command is sent the progr...
Definition cp_external_control.F:90
ec_methods
Routines used for Harris functional Kohn-Sham calculation.
Definition ec_methods.F:15
ec_methods::create_kernel
subroutine, public create_kernel(qs_env, vxc, vxc_tau, rho, rho1_r, rho1_g, tau1_r, xc_section, compute_virial, virial_xc)
Creation of second derivative xc-potential.
Definition ec_methods.F:88
ec_orth_solver
AO-based conjugate-gradient response solver routines.
Definition ec_orth_solver.F:15
ec_orth_solver::ec_response_ao
subroutine, public ec_response_ao(qs_env, p_env, matrix_hz, matrix_pz, matrix_wz, iounit, should_stop, silent)
AO-based conjugate gradient linear response solver. In goes the right hand side B of the equation AZ=...
Definition ec_orth_solver.F:366
input_constants
collects all constants needed in input so that they can be used without circular dependencies
Definition input_constants.F:17
input_constants::precond_mlp
integer, parameter, public precond_mlp
Definition input_constants.F:517
input_constants::kg_tnadd_embed_ri
integer, parameter, public kg_tnadd_embed_ri
Definition input_constants.F:1153
input_constants::kg_tnadd_embed
integer, parameter, public kg_tnadd_embed
Definition input_constants.F:1153
input_constants::do_admm_aux_exch_func_none
integer, parameter, public do_admm_aux_exch_func_none
Definition input_constants.F:875
input_constants::ls_s_sqrt_proot
integer, parameter, public ls_s_sqrt_proot
Definition input_constants.F:976
input_constants::ls_s_sqrt_ns
integer, parameter, public ls_s_sqrt_ns
Definition input_constants.F:976
input_section_types
objects that represent the structure of input sections and the data contained in an input section
Definition input_section_types.F:15
input_section_types::section_vals_get_subs_vals
recursive type(section_vals_type) function, pointer, public section_vals_get_subs_vals(section_vals, subsection_name, i_rep_section, can_return_null)
returns the values of the requested subsection
Definition input_section_types.F:731
input_section_types::section_vals_val_get
subroutine, public section_vals_val_get(section_vals, keyword_name, i_rep_section, i_rep_val, n_rep_val, val, l_val, i_val, r_val, c_val, l_vals, i_vals, r_vals, c_vals, explicit)
returns the requested value
Definition input_section_types.F:1047
iterate_matrix
Routines useful for iterative matrix calculations.
Definition iterate_matrix.F:13
iterate_matrix::matrix_sqrt_newton_schulz
subroutine, public matrix_sqrt_newton_schulz(matrix_sqrt, matrix_sqrt_inv, matrix, threshold, order, eps_lanczos, max_iter_lanczos, symmetrize, converged, iounit)
compute the sqrt of a matrix via the sign function and the corresponding Newton-Schulz iterations the...
Definition iterate_matrix.F:1626
iterate_matrix::matrix_sqrt_proot
subroutine, public matrix_sqrt_proot(matrix_sqrt, matrix_sqrt_inv, matrix, threshold, order, eps_lanczos, max_iter_lanczos, symmetrize, converged)
compute the sqrt of a matrix via the general algorithm for the p-th root of Richters et al....
Definition iterate_matrix.F:1878
kg_correction
Routines for a Kim-Gordon-like partitioning into molecular subunits.
Definition kg_correction.F:14
kg_correction::kg_ekin_subset
subroutine, public kg_ekin_subset(qs_env, ks_matrix, ekin_mol, calc_force, do_kernel, pmat_ext)
Calculates the subsystem Hohenberg-Kohn kinetic energy and the forces.
Definition kg_correction.F:88
kinds
Defines the basic variable types.
Definition kinds.F:23
kinds::dp
integer, parameter, public dp
Definition kinds.F:34
machine
Machine interface based on Fortran 2003 and POSIX.
Definition machine.F:17
machine::m_flush
subroutine, public m_flush(lunit)
flushes units if the &GLOBAL flag is set accordingly
Definition machine.F:136
machine::m_walltime
real(kind=dp) function, public m_walltime()
returns time from a real-time clock, protected against rolling early/easily
Definition machine.F:153
mathlib
Collection of simple mathematical functions and subroutines.
Definition mathlib.F:15
mathlib::abnormal_value
logical function, public abnormal_value(a)
determines if a value is not normal (e.g. for Inf and Nan) based on IO to work also under optimizatio...
Definition mathlib.F:159
message_passing
Interface to the message passing library MPI.
Definition message_passing.F:23
pw_env_types
container for various plainwaves related things
Definition pw_env_types.F:14
pw_env_types::pw_env_get
subroutine, public pw_env_get(pw_env, pw_pools, cube_info, gridlevel_info, auxbas_pw_pool, auxbas_grid, auxbas_rs_desc, auxbas_rs_grid, rs_descs, rs_grids, xc_pw_pool, vdw_pw_pool, poisson_env, interp_section)
returns the various attributes of the pw env
Definition pw_env_types.F:113
pw_methods
Definition pw_methods.F:25
pw_poisson_methods
Definition pw_poisson_methods.F:13
pw_poisson_types
functions related to the poisson solver on regular grids
Definition pw_poisson_types.F:22
pw_pool_types
Manages a pool of grids (to be used for example as tmp objects), but can also be used to instantiate ...
Definition pw_pool_types.F:24
pw_types
Definition pw_types.F:24
qs_environment_types
Definition qs_environment_types.F:14
qs_environment_types::get_qs_env
subroutine, public get_qs_env(qs_env, atomic_kind_set, qs_kind_set, cell, super_cell, cell_ref, use_ref_cell, kpoints, dft_control, mos, sab_orb, sab_all, qmmm, qmmm_periodic, mimic, sac_ae, sac_ppl, sac_lri, sap_ppnl, sab_vdw, sab_scp, sap_oce, sab_lrc, sab_se, sab_xtbe, sab_tbe, sab_core, sab_xb, sab_xtb_pp, sab_xtb_nonbond, sab_almo, sab_kp, sab_kp_nosym, sab_cneo, particle_set, energy, force, matrix_h, matrix_h_im, matrix_ks, matrix_ks_im, matrix_vxc, run_rtp, rtp, matrix_h_kp, matrix_h_im_kp, matrix_ks_kp, matrix_ks_im_kp, matrix_vxc_kp, kinetic_kp, matrix_s_kp, matrix_w_kp, matrix_s_ri_aux_kp, matrix_s, matrix_s_ri_aux, matrix_w, matrix_p_mp2, matrix_p_mp2_admm, rho, rho_xc, pw_env, ewald_env, ewald_pw, active_space, mpools, input, para_env, blacs_env, scf_control, rel_control, kinetic, qs_charges, vppl, xcint_weights, rho_core, rho_nlcc, rho_nlcc_g, ks_env, ks_qmmm_env, wf_history, scf_env, local_particles, local_molecules, distribution_2d, dbcsr_dist, molecule_kind_set, molecule_set, subsys, cp_subsys, oce, local_rho_set, rho_atom_set, task_list, task_list_soft, rho0_atom_set, rho0_mpole, rhoz_set, rhoz_cneo_set, ecoul_1c, rho0_s_rs, rho0_s_gs, rhoz_cneo_s_rs, rhoz_cneo_s_gs, do_kpoints, has_unit_metric, requires_mo_derivs, mo_derivs, mo_loc_history, nkind, natom, nelectron_total, nelectron_spin, efield, neighbor_list_id, linres_control, xas_env, virial, cp_ddapc_env, cp_ddapc_ewald, outer_scf_history, outer_scf_ihistory, x_data, et_coupling, dftb_potential, results, se_taper, se_store_int_env, se_nddo_mpole, se_nonbond_env, admm_env, lri_env, lri_density, exstate_env, ec_env, harris_env, dispersion_env, gcp_env, vee, rho_external, external_vxc, mask, mp2_env, bs_env, kg_env, wanniercentres, atprop, ls_scf_env, do_transport, transport_env, v_hartree_rspace, s_mstruct_changed, rho_changed, potential_changed, forces_up_to_date, mscfg_env, almo_scf_env, gradient_history, variable_history, embed_pot, spin_embed_pot, polar_env, mos_last_converged, eeq, rhs, do_rixs, tb_tblite)
Get the QUICKSTEP environment.
Definition qs_environment_types.F:530
qs_integrate_potential
Integrate single or product functions over a potential on a RS grid.
Definition qs_integrate_potential.F:22
qs_kpp1_env_types
basis types for the calculation of the perturbation of density theory.
Definition qs_kpp1_env_types.F:14
qs_linres_kernel
linres kernel functions
Definition qs_linres_kernel.F:14
qs_linres_kernel::apply_xc_admm
subroutine, public apply_xc_admm(qs_env, p_env)
...
Definition qs_linres_kernel.F:832
qs_linres_kernel::apply_hfx
subroutine, public apply_hfx(qs_env, p_env)
Update action of TDDFPT operator on trial vectors by adding exact-exchange term.
Definition qs_linres_kernel.F:668
qs_linres_types
Type definitiona for linear response calculations.
Definition qs_linres_types.F:12
qs_p_env_methods
Utility functions for the perturbation calculations.
Definition qs_p_env_methods.F:16
qs_p_env_methods::p_env_finish_kpp1
subroutine, public p_env_finish_kpp1(qs_env, p_env)
...
Definition qs_p_env_methods.F:856
qs_p_env_methods::p_env_update_rho
subroutine, public p_env_update_rho(p_env, qs_env)
...
Definition qs_p_env_methods.F:395
qs_p_env_methods::p_env_check_i_alloc
subroutine, public p_env_check_i_alloc(p_env, qs_env)
checks that the intenal storage is allocated, and allocs it if needed
Definition qs_p_env_methods.F:314
qs_p_env_types
basis types for the calculation of the perturbation of density theory.
Definition qs_p_env_types.F:14
qs_rho_types
superstucture that hold various representations of the density and keeps track of which ones are vali...
Definition qs_rho_types.F:18
qs_rho_types::qs_rho_get
subroutine, public qs_rho_get(rho_struct, rho_ao, rho_ao_im, rho_ao_kp, rho_ao_im_kp, rho_r, drho_r, rho_g, drho_g, tau_r, tau_g, rho_r_valid, drho_r_valid, rho_g_valid, drho_g_valid, tau_r_valid, tau_g_valid, tot_rho_r, tot_rho_g, rho_r_sccs, soft_valid, complex_rho_ao)
returns info about the density described by this object. If some representation is not available an e...
Definition qs_rho_types.F:229
xc
Exchange and Correlation functional calculations.
Definition xc.F:17
xc::xc_prep_2nd_deriv
subroutine, public xc_prep_2nd_deriv(deriv_set, rho_set, rho_r, pw_pool, weights, xc_section, tau_r)
Prepare objects for the calculation of the 2nd derivatives of the density functional....
Definition xc.F:5533
admm_types::admm_type
stores some data used in wavefunction fitting
Definition admm_types.F:120
cp_control_types::dft_control_type
Definition cp_control_types.F:631
cp_dbcsr_api::dbcsr_p_type
Definition cp_dbcsr_api.F:172
cp_dbcsr_api::dbcsr_type
Definition cp_dbcsr_api.F:176
input_section_types::section_vals_type
stores the values of a section
Definition input_section_types.F:127
message_passing::mp_para_env_type
stores all the informations relevant to an mpi environment
Definition message_passing.F:743
pw_env_types::pw_env_type
contained for different pw related things
Definition pw_env_types.F:70
pw_poisson_types::pw_poisson_type
environment for the poisson solver
Definition pw_poisson_types.F:123
pw_pool_types::pw_pool_p_type
to create arrays of pools
Definition pw_pool_types.F:135
pw_pool_types::pw_pool_type
Manages a pool of grids (to be used for example as tmp objects), but can also be used to instantiate ...
Definition pw_pool_types.F:83
pw_types::pw_c1d_gs_type
Definition pw_types.F:127
pw_types::pw_r3d_rs_type
Definition pw_types.F:62
qs_environment_types::qs_environment_type
Definition qs_environment_types.F:220
qs_kpp1_env_types::qs_kpp1_env_type
environment that keeps the informations and temporary val to build the kpp1 kernel matrix
Definition qs_kpp1_env_types.F:40
qs_linres_types::linres_control_type
General settings for linear response calculations.
Definition qs_linres_types.F:53
qs_p_env_types::qs_p_env_type
Represent a qs system that is perturbed. Can calculate the linear operator and the rhs of the system ...
Definition qs_p_env_types.F:58
qs_rho_types::qs_rho_type
keeps the density in various representations, keeping track of which ones are valid.
Definition qs_rho_types.F:62