This file is indexed.

/usr/share/code_saturne/user_examples/cs_user_cdo_extra_op-verif-diffusion.c is in code-saturne-data 4.2.0+repack-1build1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
/*============================================================================
 * Additional post-processing functions defined by user related to CDO schemes
 *============================================================================*/

/*
  This file is part of Code_Saturne, a general-purpose CFD tool.

  Copyright (C) 1998-2015 EDF S.A.

  This program is free software; you can redistribute it and/or modify it under
  the terms of the GNU General Public License as published by the Free Software
  Foundation; either version 2 of the License, or (at your option) any later
  version.

  This program is distributed in the hope that it will be useful, but WITHOUT
  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
  FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more
  details.

  You should have received a copy of the GNU General Public License along with
  this program; if not, write to the Free Software Foundation, Inc., 51 Franklin
  Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/

/*----------------------------------------------------------------------------*/

#include "cs_defs.h"

/*----------------------------------------------------------------------------
 * Standard C library headers
 *----------------------------------------------------------------------------*/

#include <errno.h>
#include <locale.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <math.h>

/*----------------------------------------------------------------------------
 * Local headers
 *----------------------------------------------------------------------------*/

#include <bft_mem.h>
#include <bft_printf.h>

#include "cs_mesh.h"
#include "cs_mesh_quantities.h"
#include "cs_mesh_location.h"
#include "cs_post.h"
#include "cs_field.h"
#include "cs_cdo.h"
#include "cs_reco.h"
#include "cs_quadrature.h"
#include "cs_cdo_toolbox.h"
#include "cs_param.h"
#include "cs_equation_priv.h"
#include "cs_equation.h"
#include "cs_evaluate.h"
#include "cs_hodge.h"
#include "cs_cdofb_scaleq.h"

/*----------------------------------------------------------------------------
 * Header for the current file
 *----------------------------------------------------------------------------*/

#include "cs_prototypes.h"

/*----------------------------------------------------------------------------*/

BEGIN_C_DECLS

/*=============================================================================
 * Additional doxygen documentation
 *============================================================================*/

/*!
  \file cs_user_cdo_extra_op.c

  \brief Additional user-defined post-processing and analysis functions
*/

/*! \cond DOXYGEN_SHOULD_SKIP_THIS */

/*=============================================================================
 * Local Macro definitions and structure definitions
 *============================================================================*/

/*! \endcond (end ignore by Doxygen) */

static FILE  *resume = NULL;
static cs_real_t  one_third = 1./3;

/*============================================================================
 * Private function prototypes
 *============================================================================*/

/* ---------------------------------------------------------------------------
 * Retrieve the analytical solution for a given problem
 * ---------------------------------------------------------------------------*/

static inline void
_get_sol(cs_real_t          time,
         const cs_real_3_t  xyz,
         cs_get_t          *retval)
{
  double  solution;

  const double  x = xyz[0], y = xyz[1], z = xyz[2];
  const double  pi = 4.0*atan(1.0);

  solution = 1+sin(pi*x)*sin(pi*(y+0.5))*sin(pi*(z+one_third));

  (*retval).val = solution;
}

static cs_analytic_func_t *get_sol = _get_sol;

/*----------------------------------------------------------------------------*/
/*!
 * \brief  Dump information into resume file to specify the parameters of the
 *         current simulation
 *
 * \param[in]  cdoq          pointer to a cs_cdo_quantities_t structure
 * \param[in]  space_scheme  scheme for the discretization in space
 * \param[in]  eqp           pointer to the setting structure related to an
 *                           equation
 */
/*----------------------------------------------------------------------------*/

static void
_dump_info(const cs_cdo_quantities_t   *cdoq,
           const cs_space_scheme_t      space_scheme,
           const cs_equation_param_t   *eqp)

{
  const cs_param_hodge_t  h_info = eqp->diffusion_hodge;

  fprintf(resume," -dim- n_vertices  %d\n", cdoq->n_vertices);
  fprintf(resume," -dim- n_edges     %d\n", cdoq->n_edges);
  fprintf(resume," -dim- n_faces     %d\n", cdoq->n_faces);
  fprintf(resume," -dim- n_cells     %d\n", cdoq->n_cells);
  fprintf(resume, "\n%s", msepline);

  if (eqp->flag & CS_EQUATION_DIFFUSION) {

    if (space_scheme == CS_SPACE_SCHEME_CDOFB) {
      switch (h_info.algo) {
      case CS_PARAM_HODGE_ALGO_COST:
        fprintf(resume, " -hdg- Hodge.Op         EDFP:COST\n");
        break;
      case CS_PARAM_HODGE_ALGO_VORONOI:
        fprintf(resume, " -hdg- Hodge.Op         EDFP:VORONOI\n");
        break;
      default:
        bft_error(__FILE__, __LINE__, 0,
                  _(" Invalid algorithm to define a discrete Hodge operator."));
      } // end of switch
    }
    else if (space_scheme == CS_SPACE_SCHEME_CDOVB) {
      switch (h_info.algo) {
      case CS_PARAM_HODGE_ALGO_COST:
        fprintf(resume, " -hdg- Hodge.Op         EPFD:COST\n");
        break;
      case CS_PARAM_HODGE_ALGO_VORONOI:
        fprintf(resume, " -hdg- Hodge.Op         EPFD:VORONOI\n");
        break;
      case CS_PARAM_HODGE_ALGO_WBS:
        fprintf(resume, " -hdg- Hodge.Op         EPFD:WHITNEY_BARY\n");
        break;
      default:
        bft_error(__FILE__, __LINE__, 0,
                  _(" Invalid algorithm to define a discrete Hodge operator."));
      } // end of switch
    }

    if (h_info.algo == CS_PARAM_HODGE_ALGO_COST)
      fprintf(resume, " -hdg- Beta.Coef        %5.3e\n", h_info.coef);

  } // Diffusion term is activated

  fprintf(resume," -bc-  Enforcement    %s",
          cs_param_get_bc_enforcement_name(eqp->bc->enforcement));

  fprintf(resume, "\n%s", msepline);
}

/*----------------------------------------------------------------------------*/
/*!
 * \brief  Compute the discrete L2 norm and energy norm of the error on the
 *         gradient in face-based scheme
 *
 *
 * \param[in] topo       pointer to the connectivity struct.
 * \param[in] cdoq       pointer to the additional quantities struct.
 * \param[in] h_info     information about the discrete Hodge op.
 * \param[in] cell_rpex  reduction of the exact solution at cell centers
 * \param[in] face_rpex  reduction of the exact solution at face centers
 * \param[in] cell_pdi   computed solution at cell centers
 * \param[in] face_pdi   computed solution at face centers
 */
/*----------------------------------------------------------------------------*/

static void
_compute_fb_errgrd(const cs_cdo_connect_t      *topo,
                   const cs_cdo_quantities_t   *cdoq,
                   const cs_param_hodge_t       h_info,
                   const double                 cell_rpex[],
                   const double                 face_rpex[],
                   const double                 cell_pdi[],
                   const double                 face_pdi[])
{
  short int sgn;
  int  i, j, k, ij, c_id, f_id;
  double  gcontrib, pvol, l2dgrd, enerd, gdic;
  cs_real_3_t  xc;
  cs_quant_t  fq;

  double  num_l2d = 0, denum_l2d = 0, num_end = 0, denum_end = 0;
  double  *gexc = NULL, *dgc = NULL;
  cs_locmat_t  *_h = NULL;
  cs_hodge_builder_t  *hb = cs_hodge_builder_init(topo, h_info);

  const double  zthreshold = cs_get_zero_threshold();
  const double  _over3 = 1./3.;

  /* Initialize local buffers */
  BFT_MALLOC(gexc, 2*topo->n_max_fbyc, double);
  dgc = gexc + topo->n_max_fbyc;
  for (i = 0; i < 2*topo->n_max_fbyc; i++) gexc[i] = 0;

  /* Loop on cells */
  for (c_id = 0; c_id < cdoq->n_cells; c_id++) {

    double  _nl2c = 0, _dl2c = 0, _denc = 0, _nenc = 0;

    /* Build a local discrete Hodge operator */
    _h = cs_hodge_build_local(c_id, topo, cdoq, hb);

    for (k = 0; k < 3; k++)
      xc[k] = cdoq->cell_centers[3*c_id+k];

    for (i = 0; i < _h->n_ent; i++) {

      cs_lnum_t  shift = topo->c2f->idx[c_id] + i;
      const cs_nvec3_t  deq = cdoq->dedge[shift]; /* Dual edge quantities */

      f_id = _h->ids[i];
      assert(f_id == topo->c2f->col_id[shift]);

      fq = cdoq->face[f_id];
      sgn = topo->c2f->sgn[shift];

      /* Compute pvol_{f,c} */
      pvol = _over3 * deq.meas * fq.meas * _dp3(deq.unitv, fq.unitv);
      gcontrib = pvol/(deq.meas*deq.meas);

      gexc[i] = sgn*(face_rpex[f_id] - cell_rpex[c_id]);
      gdic = sgn*(face_pdi[f_id] - cell_pdi[c_id]);
      dgc[i] = gexc[i] - gdic;

      _nl2c += gcontrib * dgc[i]*dgc[i];
      _dl2c += gcontrib * gexc[i]*gexc[i];

    } // Loop on cell faces

    for (i = 0; i < _h->n_ent; i++) {
      for (j = 0; j < _h->n_ent; j++) {
        ij = i*_h->n_ent+j;
        _nenc += dgc[i]*_h->mat[ij]*dgc[j];
        _denc += gexc[i]*_h->mat[ij]*gexc[j];
      }
    }

    num_l2d += _nl2c;
    denum_l2d += _dl2c;
    num_end += _nenc;
    denum_end += _denc;

  } /* Loop on cells */

  /* Compute the L2 discrete norm on gradient */
  if (fabs(denum_l2d) > zthreshold)
    l2dgrd = sqrt(num_l2d/denum_l2d);
  else
    l2dgrd = sqrt(num_l2d);
  printf(" >> l2dgrd       % 10.6e\n", l2dgrd);

  /* Compute the discrete energy norm */
  if (fabs(denum_end) > zthreshold)
    enerd = sqrt(num_end/denum_end);
  else
    enerd = sqrt(num_end);
  printf(" >> enerd        % 10.6e\n", enerd);

  /* Output results */
  fprintf(resume, " -cvg- l2dgrd              % 10.6e\n", l2dgrd);
  fprintf(resume, " -cvg- enerd               % 10.6e\n", enerd);

  /* Free buffers */
  hb = cs_hodge_builder_free(hb);
  BFT_FREE(gexc);
}

/*----------------------------------------------------------------------------*/
/*!
 * \brief  Compute the L2 norm for a potential in vertex-based scheme from
 *         a conforming reconstruction
 *
 * \param[in]   topo       pointer to the connectivity struct.
 * \param[in]   geom       pointer to the additional quantities struct.
 * \param[in]   time_step  pointer to a time step structure
 * \param[in]   pdi        pointer to the field of vtx-based DoFs
 * \param[out]  num        int_Omega (pex - Reco{vtxsp}(pdi))^2
 * \param[out]  denum      int_Omega pex^2
 */
/*----------------------------------------------------------------------------*/

static void
_compute_vb_l2pot(const cs_mesh_t             *m,
                  const cs_cdo_connect_t      *topo,
                  const cs_cdo_quantities_t   *geom,
                  const cs_time_step_t        *time_step,
                  const double                *pdi,
                  double                      *num,
                  double                      *denum)
{
  int  i, j, k, p, c_id;
  double  ddip_gpts, voltet, n_add, d_add;
  cs_real_3_t  xc, gpts[5];
  double  rpex_gpts[5], weights[5], pdi_gpts[5];

  double  _num = 0.0, _denum = 0.0;
  double  *pdi_recc = NULL, *pdi_recf = NULL;

  const double  tcur = time_step->t_cur;
  const double  _over_six = 1./6.;
  const double  *xyz = m->vtx_coord;

  /* Reconstruct potentials at face centers and cell centers */
  cs_reco_conf_vtx_dofs(topo, geom, pdi, &pdi_recc, &pdi_recf);

  /* Compute conforming norm */
  for (c_id = 0; c_id < geom->n_cells; c_id++) {

    cs_get_t  get;
    double  num_cell = 0.0, denum_cell = 0.0;

    const double  pc = pdi_recc[c_id];

    /* Get cell center */
    for (k = 0; k < 3; k++)
      xc[k] = geom->cell_centers[3*c_id+k];

    for (i = topo->c2f->idx[c_id]; i < topo->c2f->idx[c_id+1]; i++) {

      const cs_lnum_t  f_id = topo->c2f->col_id[i];
      const cs_quant_t  fq = geom->face[f_id]; // Face quantities
      const double  pf = pdi_recf[f_id];

      for (j = topo->f2e->idx[f_id]; j < topo->f2e->idx[f_id+1]; j++) {

        const cs_lnum_t  e_id = topo->f2e->col_id[j];
        const cs_lnum_t  v_id1 = topo->e2v->col_id[topo->e2v->idx[e_id]  ];
        const cs_lnum_t  v_id2 = topo->e2v->col_id[topo->e2v->idx[e_id]+1];
        const double  pv1 = pdi[v_id1], pv2 = pdi[v_id2];

        voltet = cs_voltet(&(xyz[3*v_id1]), &(xyz[3*v_id2]), fq.center, xc);

        /* Analytical function and integration with the highest
           available quadrature */
        cs_quadrature_tet_5pts(&(xyz[3*v_id1]), &(xyz[3*v_id2]), fq.center, xc,
                               voltet, gpts, weights);

        for (p = 0; p < 5; p++) {
          get_sol(tcur, gpts[p], &get);
          rpex_gpts[p] = get.val;
        }

        /* Should keep the same order */
        pdi_gpts[0] = _over_six *(pv1 + pv2 + pf) + 0.5*pc;
        pdi_gpts[1] = _over_six *(pv2 + pf + pc)  + 0.5*pv1;
        pdi_gpts[2] = _over_six *(pf + pc + pv1)  + 0.5*pv2;
        pdi_gpts[3] = _over_six *(pc + pv1 + pv2) + 0.5*pf;
        pdi_gpts[4] = 0.25*(pv1 + pv2 + pf + pc);

        n_add = 0, d_add = 0;
        for (p = 0; p < 5; p++) { /* Loop on Gauss points */
          ddip_gpts = rpex_gpts[p] - pdi_gpts[p];
          n_add += weights[p] * ddip_gpts*ddip_gpts;
          d_add += weights[p] * rpex_gpts[p]*rpex_gpts[p];
        }
        num_cell += n_add;
        denum_cell += d_add;

      } /* Loop on face edges */

    } /* Loop on cell faces */

    _denum += denum_cell;
    _num += num_cell;

  } /* Loop on cells */

  *num = _num;
  *denum = _denum;

  BFT_FREE(pdi_recc);
  BFT_FREE(pdi_recf);
}

/*----------------------------------------------------------------------------*/
/*!
 * \brief  Post-process the solution of a scalar convection/diffusion equation
 *         solved with a CDO vertex-based scheme.
 *
 * \param[in]  m          pointer to a cs_mesh_t structure
 * \param[in]  connect    pointer to a cs_cdo_connect_t structure
 * \param[in]  cdoq       pointer to a cs_cdo_quantities_t structure
 * \param[in]  time_step  pointer to a time step structure
 * \param[in]  eq         pointer to a cs_equation_t structure
 * \param[in]  anacomp    do an analytic comparison or not
 */
/*----------------------------------------------------------------------------*/

static void
_cdovb_post(const cs_mesh_t            *m,
            const cs_cdo_connect_t     *connect,
            const cs_cdo_quantities_t  *cdoq,
            const cs_time_step_t       *time_step,
            const cs_equation_t        *eq,
            bool                        anacomp)
{
  cs_data_info_t  dinfo;
  int  i, len, work_size;
  double  num, denum, l2pot, l2dpot, enerd, l2dgrd;
  cs_get_t  get;

  char  *postlabel = NULL;
  double  *ddip = NULL, *ddig = NULL, *rpex = NULL, *gdi = NULL, *rgex = NULL;
  double  *pvol = NULL, *work = NULL;

  const double  zthreshold = cs_get_zero_threshold();
  const double  tcur = time_step->t_cur;
  const cs_lnum_t  n_vertices = cdoq->n_vertices, n_edges = cdoq->n_edges;
  const cs_equation_param_t  *eqp = cs_equation_get_param(eq);
  const cs_param_hodge_t  h_info = eqp->diffusion_hodge;
  const cs_field_t  *field = cs_equation_get_field(eq);
  const cs_real_t  *pdi = field->val;

  /* Output a summary of results */
  dinfo = cs_analysis_data(n_vertices,  // n_elts
                           1,           // stride
                           CS_DOUBLE,   // cs_datatype_t
                           pdi,         // data
                           false);      // compute with absolute values

  fprintf(resume, " -bnd- Scal.Min   % 10.6e\n", dinfo.min.value);
  fprintf(resume, " -bnd- Scal.Max   % 10.6e\n", dinfo.max.value);
  fprintf(resume, " -bnd- Scal.Mean  % 10.6e\n", dinfo.mean);
  fprintf(resume, " -bnd- Scal.Sigma % 10.6e\n", dinfo.sigma);
  fprintf(resume, "%s", msepline);

  if (anacomp) { /* Comparison with an analytical solution */

    /* Work buffer */
    work_size = CS_MAX(3*cdoq->n_cells, n_vertices);
    work_size = CS_MAX(n_edges, work_size);
    BFT_MALLOC(work, work_size, double);

    /* pex = exact potential
       pdi = discrete potential (solution of the discrete system)
       rpex = Red(vtxsp)(Pex) = reduction on vertices of the exact potential
       ddip = rpex - pdi
    */

    BFT_MALLOC(rpex, n_vertices, double);
    BFT_MALLOC(ddip, n_vertices, double);
    for (i = 0; i < n_vertices; i++) {
      get_sol(tcur, &(m->vtx_coord[3*i]), &get);
      rpex[i] = get.val;
      ddip[i] = rpex[i] - pdi[i];
    }

    len = strlen(field->name) + 7 + 1;
    BFT_MALLOC(postlabel, len, char);
    sprintf(postlabel, "%s.Error", field->name);

    cs_post_write_vertex_var(-1,          /* id du maillage de post */
                             postlabel,
                             1,           /* dim */
                             false,       /* interlace */
                             true,        /* parent mesh */
                             CS_POST_TYPE_cs_real_t,
                             ddip,        /* values on vertices */
                             time_step);  /* time step structure */

    sprintf(postlabel, "%s.RefSol", field->name);
    cs_post_write_vertex_var(-1,          /* id du maillage de post */
                             postlabel,
                             1,           /* dim */
                             false,       /* interlace */
                             true,        /* parent mesh */
                             CS_POST_TYPE_cs_real_t,
                             rpex,        /* values on vertices */
                             time_step);  /* time step structure */

    /* Analyse the exact solution */
    dinfo = cs_analysis_data(n_vertices, // n_elts
                             1,                // stride
                             CS_DOUBLE,        // cs_datatype_t
                             rpex,             // data
                             false);           // compute with absolute values

    fprintf(resume, " -bnd- Ref.Min    % 10.6e\n", dinfo.min.value);
    fprintf(resume, " -bnd- Ref.Max    % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -bnd- Ref.Mean   % 10.6e\n", dinfo.mean);
    fprintf(resume, " -bnd- Ref.Sigma  % 10.6e\n", dinfo.sigma);
    fprintf(resume, "%s", msepline);

    for (i = 0; i < n_vertices; i++)
      work[i] = fabs(ddip[i]);

    sprintf(postlabel, "%s.AbsErr", field->name);
    cs_post_write_vertex_var(-1,          /* id du maillage de post */
                             postlabel,
                             1,           /* dim */
                             false,       /* interlace */
                             true,        /* parent mesh */
                             CS_POST_TYPE_cs_real_t,
                             work,        /* values on vertices */
                             time_step);  /* time step structure */

    dinfo = cs_analysis_data(n_vertices, // n_elts
                             1,                // stride
                             CS_DOUBLE,        // cs_datatype_t
                             work,             // data
                             false);           // compute with absolute values

    fprintf(resume, " -cvg- ErrAbs.Scal.Max  % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrAbs.Scal.Min  % 10.6e\n", dinfo.min.value);

    /* Compute pvol related to vertices */
    BFT_MALLOC(pvol, CS_MAX(n_vertices, n_edges), double);
    cs_compute_pvol_vtx(connect, cdoq, &pvol);

    /* Compute discrete L2 error norm on the potential */
    num = cs_sum(n_vertices, ddip, pvol, CS_TOOLBOX_WSUM2);
    denum = cs_sum(n_vertices, rpex, pvol, CS_TOOLBOX_WSUM2);
    if (fabs(denum) > zthreshold)
      l2dpot = sqrt(num/denum);
    else
      l2dpot = sqrt(num);

    /* Continuous L2 norm based on a conforming reconstruction of the potential
       ||*||_2^2 = int_Omega (pex - Reco{vtxsp}(pdi))^2

       Sudivide each cell into tetrehadron (xv, xe, xf, xc) and use Gauss
       quadrature to compute the elementary integrals.
    */

    _compute_vb_l2pot(m, connect, cdoq, time_step, pdi, &num, &denum);
    if (fabs(denum) > zthreshold)
      l2pot = sqrt(num/denum);
    else
      l2pot = sqrt(num);

    fprintf(resume, " -cvg- l2dpot           % 10.6e\n", l2dpot);
    fprintf(resume, " -cvg- l2pot            % 10.6e\n", l2pot);
    fprintf(resume, "%s", msepline);

    printf(" >> l2dpot = %7.4e\n", l2dpot);
    printf(" >> l2pot  = %7.4e\n", l2pot);

    /* Compute norm related to the gradient of the error
       - Discrete L2 norm on the discrete gradient
       - Energy norm

       gdi = the discrete gradient related to pdi
       rgex = the reduction of the exact gradient on each edge
       ddig = rgex - gdi
    */

    cs_sla_matvec(connect->e2v, pdi, &gdi, true);
    cs_sla_matvec(connect->e2v, rpex, &rgex, true);

    BFT_MALLOC(ddig, n_edges, double);
    for (i = 0; i < n_edges; i++)
      ddig[i] = rgex[i] - gdi[i];

    dinfo = cs_analysis_data(n_edges,      // n_elts
                             1,            // stride
                             CS_DOUBLE,    // cs_datatype_t
                             ddig,         // data
                             true);        // compute with absolute values

    /* Discrete L2 norm for the error on the gradient
       ||| * |||_{edgesp,2}^2 = Sum_e |pvol_e| (*_e/|e|)^2
       |pvol_e| volume related to each edge -> store in pvol  */
    cs_compute_pvol_edge(connect, cdoq, &pvol);
    for (i = 0; i < n_edges; i++)
      work[i] = ddig[i]/cdoq->edge[i].meas;
    num = cs_sum(n_edges, work, pvol, CS_TOOLBOX_WSUM2);
    for (i = 0; i < n_edges; i++)
      work[i] = rgex[i]/cdoq->edge[i].meas;
    denum = cs_sum(cdoq->n_edges, work, pvol, CS_TOOLBOX_WSUM2);
    if (fabs(denum) > zthreshold)
      l2dgrd = sqrt(num/denum);
    else
      l2dgrd = sqrt(num);
    printf(" >> l2dgrd = %7.4e\n", l2dgrd);

    /* Energy norm^2 = [ ddig, Hodge*ddig ]_EpFd / [ rgex, Hodge*rgex]_EpFd */
    if (h_info.algo == CS_PARAM_HODGE_ALGO_WBS) {

      // TODO
    }
    else {

      cs_sla_matrix_t  *H = cs_hodge_compute(connect, cdoq,
                                             eqp->diffusion_property,
                                             h_info);

      cs_sla_matvec(H, ddig, &work, true);
      num = cs_dp(n_edges, ddig, work);
      cs_sla_matvec(H, rgex, &work, true);
      denum = cs_dp(n_edges, rgex, work);

      H = cs_sla_matrix_free(H);
    }

    if (fabs(denum) > zthreshold)
      enerd = sqrt(num/denum);
    else
      enerd = sqrt(num);

    printf(" >> enerd  = %7.4e\n", enerd);

    /* Output results */
    fprintf(resume, " -cvg- ErrAbs.Grd.Max  % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrAbs.Grd.Min  % 10.6e\n", dinfo.min.value);
    fprintf(resume, " -cvg- l2dgrd          % 10.6e\n", l2dgrd);
    fprintf(resume, " -cvg- enerd           % 10.6e\n", enerd);
    fprintf(resume, "%s", msepline);

    /* Post-processing of reconstructed vector fields at each cell center */
    cs_reco_ccen_edge_dofs(connect, cdoq, gdi, &work);

    sprintf(postlabel, "%s.GrdRec", field->name);
    cs_post_write_var(-1,              // id du maillage de post
                      postlabel,
                      3,               // dim
                      true,            // interlace
                      true,            // true = original mesh
                      CS_POST_TYPE_cs_real_t,
                      work,            // values on cells
                      NULL,            // values at internal faces
                      NULL,            // values at border faces
                      time_step);      // time step management structure

    cs_reco_ccen_edge_dofs(connect, cdoq, ddig, &work);

    sprintf(postlabel, "%s.ErrGrd", field->name);
    cs_post_write_var(-1,              // id du maillage de post
                      postlabel,
                      3,               // dim
                      true,            // interlace
                      true,            // true = original mesh
                      CS_POST_TYPE_cs_real_t,
                      work,            // values on cells
                      NULL,            // values at internal faces
                      NULL,            // values at border faces
                      time_step);      // time step management structure

    dinfo = cs_analysis_data(cdoq->n_cells,  // n_elts
                             3,               // stride
                             CS_DOUBLE,       // cs_datatype_t
                             work,            // data
                             true);           // compute with absolute values

    fprintf(resume, " -cvg- ErrAbs.GrdReco.Max  % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrAbs.GrdReco.Min  % 10.6e\n", dinfo.min.value);

    /* Free */
    BFT_FREE(postlabel);
    BFT_FREE(pvol);
    BFT_FREE(work);
    BFT_FREE(ddip);
    BFT_FREE(rpex);
    BFT_FREE(ddig);
    BFT_FREE(rgex);
    BFT_FREE(gdi);

  }

}

/*----------------------------------------------------------------------------*/
/*!
 * \brief  Post-process the solution of a scalar convection/diffusion equation
 *         solved with a CDO face-based scheme.
 *
 * \param[in]  m          pointer to a cs_mesh_t structure
 * \param[in]  connect    pointer to a cs_cdo_connect_t structure
 * \param[in]  cdoq       pointer to a cs_cdo_quantities_t structure
 * \param[in]  time_step  pointer to a time step structure
 * \param[in]  eq         pointer to a cs_equation_t structure
 * \param[in]  anacomp    do an analytic comparison or not
 */
/*----------------------------------------------------------------------------*/

static void
_cdofb_post(const cs_mesh_t            *m,
            const cs_cdo_connect_t     *connect,
            const cs_cdo_quantities_t  *cdoq,
            const cs_time_step_t       *time_step,
            const cs_equation_t        *eq,
            bool                        anacomp)
{
  cs_data_info_t  dinfo;
  int  i, len, work_size;
  double  num, denum, l2dpotc, l2dpotf, l2r0potc;
  cs_get_t  get;

  char  *postlabel = NULL;
  double  *cell_dpdi = NULL, *face_dpdi = NULL;
  double  *cell_rpex = NULL, *face_rpex = NULL;
  double  *pvol = NULL, *work = NULL;

  const double  zthreshold = cs_get_zero_threshold();
  const double  tcur = time_step->t_cur;
  const cs_lnum_t  n_cells = cdoq->n_cells;
  const cs_lnum_t  n_faces = cdoq->n_faces;
  const cs_lnum_t  n_i_faces = m->n_i_faces;
  const cs_equation_param_t  *eqp = cs_equation_get_param(eq);
  const cs_param_hodge_t  h_info = eqp->diffusion_hodge;
  const cs_field_t  *field = cs_equation_get_field(eq);
  const cs_real_t  *cell_pdi = field->val;
  const cs_real_t  *face_pdi = cs_equation_get_face_values(eq);

  /* Output a summary of results */
  dinfo = cs_analysis_data(n_cells,    // n_elts
                           1,          // stride
                           CS_DOUBLE,  // cs_datatype_t
                           cell_pdi,   // data
                           false);     // compute with absolute values

  fprintf(resume, " -bnd- Scal.Cell.Min   % 10.6e\n", dinfo.min.value);
  fprintf(resume, " -bnd- Scal.Cell.Max   % 10.6e\n", dinfo.max.value);
  fprintf(resume, " -bnd- Scal.Cell.Mean  % 10.6e\n", dinfo.mean);
  fprintf(resume, " -bnd- Scal.Cell.Sigma % 10.6e\n", dinfo.sigma);
  fprintf(resume, "%s", msepline);

  dinfo = cs_analysis_data(n_faces,    // n_elts
                           1,          // stride
                           CS_DOUBLE,  // cs_datatype_t
                           face_pdi,   // data
                           false);     // compute with absolute values

  fprintf(resume, " -bnd- Scal.Face.Min   % 10.6e\n", dinfo.min.value);
  fprintf(resume, " -bnd- Scal.Face.Max   % 10.6e\n", dinfo.max.value);
  fprintf(resume, " -bnd- Scal.Cell.Mean  % 10.6e\n", dinfo.mean);
  fprintf(resume, " -bnd- Scal.Cell.Sigma % 10.6e\n", dinfo.sigma);
  fprintf(resume, "%s", msepline);

  if (anacomp) {  /* Comparison with an analytical solution */

    /* pex = exact potential
       pdi = discrete potential (solution of the discrete system)
       rpex = Red(vtxsp)(Pex) = reduction on vertices of the exact potential
       dpdi = rpex - pdi
    */

    BFT_MALLOC(cell_rpex, n_cells + n_faces, double);
    BFT_MALLOC(cell_dpdi, n_cells + n_faces, double);
    face_rpex = cell_rpex + n_cells;
    face_dpdi = cell_dpdi + n_cells;

    for (i = 0; i < n_cells; i++) {
      get_sol(tcur, &(cdoq->cell_centers[3*i]), &get);
      cell_rpex[i] = get.val;
      cell_dpdi[i] = cell_rpex[i] - cell_pdi[i];
    }

    /* Analyse the exact solution */
    dinfo = cs_analysis_data(n_cells,    // n_elts
                             1,          // stride
                             CS_DOUBLE,  // cs_datatype_t
                             cell_rpex,  // data
                             false);     // compute with absolute values

    fprintf(resume, " -bnd- Ref.Cell.Min    % 10.6e\n", dinfo.min.value);
    fprintf(resume, " -bnd- Ref.Cell.Max    % 10.6e\n", dinfo.max.value);

    for (i = 0; i < n_faces; i++) {

      cs_quant_t  fq = cdoq->face[i];

      get_sol(tcur, fq.center, &get);
      face_rpex[i] = get.val;
      face_dpdi[i] = face_rpex[i] - face_pdi[i];
    }

    len = strlen(field->name) + 8 + 1;
    BFT_MALLOC(postlabel, len, char);
    sprintf(postlabel, "%s.RefSolB", field->name);
    cs_post_write_var(-2,                    // id du maillage de post
                      postlabel,
                      1,                     // dim
                      false,                 // interlace
                      true,                  // true = original mesh
                      CS_POST_TYPE_cs_real_t,
                      NULL,                  // values on cells
                      NULL,                  // values at internal faces
                      face_rpex + n_i_faces, // values at border faces
                      time_step);            // time step management structure

    sprintf(postlabel, "%s.ErrBord", field->name);
    cs_post_write_var(-2,                    // id du maillage de post
                      postlabel,
                      1,                     // dim
                      false,                 // interlace
                      true,                  // true = original mesh
                      CS_POST_TYPE_cs_real_t,
                      NULL,                  // values on cells
                      NULL,                  // values at internal faces
                      face_dpdi + n_i_faces, // values at border faces
                      time_step);            // time step management structure

    /* Analyse the exact solution */
    dinfo = cs_analysis_data(n_cells,    // n_elts
                             1,          // stride
                             CS_DOUBLE,  // cs_datatype_t
                             cell_rpex,  // data
                             false);     // compute with absolute values

    fprintf(resume, " -bnd- Ref.Face.Min    % 10.6e\n", dinfo.min.value);
    fprintf(resume, " -bnd- Ref.Face.Max    % 10.6e\n", dinfo.max.value);
    fprintf(resume, "%s", msepline);

    /* Post-processing of the error */
    sprintf(postlabel, "%s.Error", field->name);
    cs_post_write_var(-1,              // id du maillage de post
                      postlabel,
                      1,               // dim
                      false,           // interlace
                      true,            // true = original mesh
                      CS_POST_TYPE_cs_real_t,
                      cell_dpdi,       // values on cells
                      NULL,            // values at interior faces
                      NULL,            // values at border faces
                      time_step);      // time step structure

    /* Compute an approximation of int_Omega (delta_p)^2 / int_Omega pex^2
       which approximates the normalized L2 norm */
    num = cs_sum(n_cells, cell_dpdi, cdoq->cell_vol, CS_TOOLBOX_WSUM2);
    denum = cs_sum(n_cells, cell_rpex, cdoq->cell_vol, CS_TOOLBOX_WSUM2);
    if (fabs(denum) > zthreshold)
      l2dpotc = sqrt(num/denum);
    else
      l2dpotc = sqrt(num);
    printf(" >> l2dpotc      % 10.6e\n", l2dpotc);

    dinfo = cs_analysis_data(n_cells,    // n_elts
                             1,          // stride
                             CS_DOUBLE,  // cs_datatype_t
                             cell_dpdi,  // data
                             false);     // compute with absolute values

    /* Output results to resume (for analysis) */
    fprintf(resume, " -cvg- ErrPot.Cell.Max     % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrPot.Cell.Min     % 10.6e\n", dinfo.min.value);

    /* Work buffer */
    work_size = CS_MAX(3*n_cells, n_cells + n_faces);
    BFT_MALLOC(work, work_size, double);
    for (i = 0; i < n_cells + n_faces; i++)
      work[i] = fabs(cell_dpdi[i]);

    sprintf(postlabel, "%s.AbsErr", field->name);
    cs_post_write_var(-1,              // id du maillage de post
                      postlabel,
                      1,               // dim
                      false,           // interlace
                      true,            // true = original mesh
                      CS_POST_TYPE_cs_real_t,
                      work,            // values on cells
                      NULL,            // values at interior faces
                      NULL,            // values at border faces
                      time_step);      // time step structure

    dinfo = cs_analysis_data(n_cells,   // n_elts
                             1,         // stride
                             CS_DOUBLE, // cs_datatype_t
                             work,      // data
                             false);    // compute with absolute values

    fprintf(resume, " -cvg- ErrAbsPot.Cell.Max  % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrAbsPot.Cell.Min  % 10.6e\n", dinfo.min.value);

    /* Compute pvol related to faces */
    BFT_MALLOC(pvol, n_faces, double);
    cs_compute_pvol_face(connect, cdoq, &pvol);

    /* Compute discrete L2 error norm on the potential */
    num = cs_sum(n_faces, face_dpdi, pvol, CS_TOOLBOX_WSUM2);
    denum = cs_sum(n_faces, face_rpex, pvol, CS_TOOLBOX_WSUM2);
    if (fabs(denum) > zthreshold)
      l2dpotf = sqrt(num/denum);
    else
      l2dpotf = sqrt(num);
    printf(" >> l2dpotf      % 10.6e\n", l2dpotf);

    dinfo = cs_analysis_data(n_faces,    // n_elts
                             1,          // stride
                             CS_DOUBLE,  // cs_datatype_t
                             face_dpdi,  // data
                             false);     // compute with absolute values

    /* Output results to resume (for analysis) */
    fprintf(resume, "%s", msepline);
    fprintf(resume, " -cvg- ErrPot.Face.Max     % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrPot.Face.Min     % 10.6e\n", dinfo.min.value);

    dinfo = cs_analysis_data(n_faces,         // n_elts
                             1,               // stride
                             CS_DOUBLE,       // cs_datatype_t
                             work + n_cells,  // data
                             false);          // compute with absolute values

    fprintf(resume, " -cvg- ErrAbsPot.Face.Max  % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrAbsPot.Face.Min  % 10.6e\n", dinfo.min.value);
    fprintf(resume, "%s", msepline);

    /* Compute the error in the norm that should be of order 2:
       -> compute the mean volume of the integral over each cell of the
       exact solution: L_cell*R_cell(solu) (using a piecewise constant
       reconstruction)
       -> dpdi = LcRc(solu) - L_vtxd^0(pdi) where L_vtxd^0(pdi) is a piecewise
       constant reconstruction over (primal) cells
       -> Compute the L^2 norm of dpdi
    */

    cs_def_t  def;
    def.analytic = get_sol;

    cs_flag_t  flag = CS_PARAM_FLAG_CELL | CS_PARAM_FLAG_PRIMAL
      | CS_PARAM_FLAG_SCAL;

    cs_evaluate(cdoq, connect, time_step,
                flag,  // DoF flag (where to compute the evaluation)
                cs_mesh_location_get_id_by_name("cells"),
                CS_PARAM_DEF_BY_ANALYTIC_FUNCTION,
                CS_QUADRATURE_BARY,
                true,   // use of subdivision into tetrahedra
                def,
                &work); // work --> int_cell solu(x,y,z)

    for (i = 0; i < n_cells; i++) {
      work[i] /= cdoq->cell_vol[i];
      cell_dpdi[i] = work[i] - cell_pdi[i];
    }

    num = cs_sum(n_cells, cell_dpdi, cdoq->cell_vol, CS_TOOLBOX_WSUM2);
    denum = cs_sum(n_cells, work, cdoq->cell_vol, CS_TOOLBOX_WSUM2);
    if (fabs(denum) > zthreshold)
      l2r0potc = sqrt(num/denum);
    else
      l2r0potc = sqrt(num);
    printf(" >> l2r0potc     % 10.6e\n", l2r0potc);

    dinfo = cs_analysis_data(n_cells,    // n_elts
                             1,          // stride
                             CS_DOUBLE,  // cs_datatype_t
                             cell_dpdi,  // data
                             true);      // compute with absolute values

    /* Output results to resume (for analysis) */
    fprintf(resume, " -cvg- ErrAbs.Pot.Lc0.Max  % 10.6e\n", dinfo.max.value);
    fprintf(resume, " -cvg- ErrAbs.Pot.Lc0.Min  % 10.6e\n", dinfo.min.value);
    fprintf(resume, "%s", msepline);
    fprintf(resume, " -cvg- l2dpotc             % 10.6e\n", l2dpotc);
    fprintf(resume, " -cvg- l2dpotf             % 10.6e\n", l2dpotf);
    fprintf(resume, " -cvg- l2r0potc            % 10.6e\n", l2r0potc);

    /* Compute the L2 discrete norm on the gradient error */
    _compute_fb_errgrd(connect, cdoq, h_info,
                       cell_rpex, face_rpex, cell_pdi, face_pdi);

    fprintf(resume, "%s", msepline);

    BFT_FREE(pvol);
    BFT_FREE(work);
    BFT_FREE(cell_dpdi);
    BFT_FREE(cell_rpex);
  }

  /* Free */
  BFT_FREE(postlabel);
}

/*============================================================================
 * Public function prototypes
 *============================================================================*/

/*----------------------------------------------------------------------------*/
/*!
 * \brief  Additional operations on results produced by CDO schemes.
 *         Define advanced post-processing and/or analysis for instance.
 *
 * \param[in]  domain   pointer to a cs_domain_t structure
 */
/*----------------------------------------------------------------------------*/

void
cs_user_cdo_extra_op(const cs_domain_t          *domain)
{
  return; /* REMOVE_LINE_FOR_USE_OF_SUBROUTINE */

  const cs_mesh_t  *m = domain->mesh;
  const cs_cdo_connect_t  *connect = domain->connect;
  const cs_cdo_quantities_t  *cdoq = domain->cdo_quantities;
  const cs_time_step_t  *time_step = domain->time_step;

  cs_equation_t  *eq = cs_domain_get_equation(domain, "FVCA6.1");
  const char *eqname = cs_equation_get_name(eq);
  const cs_equation_param_t  *eqp = cs_equation_get_param(eq);

  if (eq == NULL)
    bft_error(__FILE__, __LINE__, 0,
              " Invalid equation name. Stop extra operations.");

  /* Open a file */
  char  *filename = NULL;
  int len = strlen("Resume-.log")+strlen(eqname)+1;

  if (eqp->flag & CS_EQUATION_UNSTEADY) {
    if (time_step->nt_cur == 0)
      return;
    if (time_step->nt_cur % eqp->post_freq > 0)
      return;

    len += 9;
    BFT_MALLOC(filename, len, char);
    sprintf(filename, "Resume-%s-t%.f.log", eqname, time_step->t_cur);
  }
  else {
    if (time_step->nt_cur > 0)
        return;

    BFT_MALLOC(filename, len, char);
    sprintf(filename, "Resume-%s.log", eqname);
  }

  resume = fopen(filename, "w");

  bft_printf("\n%s", msepline);
  bft_printf("    Extra operations\n");
  bft_printf("%s", msepline);

  /* Extra-operation depends on the numerical scheme */
  cs_space_scheme_t  space_scheme = cs_equation_get_space_scheme(eq);

  _dump_info(cdoq, space_scheme, eqp);

  switch (space_scheme) {
  case CS_SPACE_SCHEME_CDOVB:
    _cdovb_post(m, connect, cdoq, time_step, eq, true);
    break;
  case CS_SPACE_SCHEME_CDOFB:
    _cdofb_post(m, connect, cdoq, time_step, eq, true);
    break;
  default:
    bft_error(__FILE__, __LINE__, 0,
              _("Invalid space scheme. Stop post-processing.\n"));
  }

  bft_printf("\n");
  bft_printf(" >> Equation %s (done)\n", eqname);
  printf("\n >> Extra operation for equation: %s\n", eqname);

  /* Free */
  BFT_FREE(filename);
  fclose(resume);
}

/*----------------------------------------------------------------------------*/

END_C_DECLS