src/gromacs/ewald/long-range-correction.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
   5  * Copyright (c) 2001-2004, The GROMACS development team.
   6  * Copyright (c) 2013,2014,2015,2016, by the GROMACS development team, led by
   7  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   8  * and including many others, as listed in the AUTHORS file in the
   9  * top-level source directory and at http://www.gromacs.org.
  10  *
  11  * GROMACS is free software; you can redistribute it and/or
  12  * modify it under the terms of the GNU Lesser General Public License
  13  * as published by the Free Software Foundation; either version 2.1
  14  * of the License, or (at your option) any later version.
  15  *
  16  * GROMACS is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  19  * Lesser General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU Lesser General Public
  22  * License along with GROMACS; if not, see
  23  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  24  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  25  *
  26  * If you want to redistribute modifications to GROMACS, please
  27  * consider that scientific software is very special. Version
  28  * control is crucial - bugs must be traceable. We will be happy to
  29  * consider code for inclusion in the official distribution, but
  30  * derived work must not be called official GROMACS. Details are found
  31  * in the README & COPYING files - if they are missing, get the
  32  * official version at http://www.gromacs.org.
  33  *
  34  * To help us fund GROMACS development, we humbly ask that you cite
  35  * the research papers on the package. Check out http://www.gromacs.org.
  36  */
  37 #include "gmxpre.h"
  38
  39 #include "long-range-correction.h"
  40
  41 #include <cmath>
  42
  43 #include "gromacs/math/functions.h"
  44 #include "gromacs/math/units.h"
  45 #include "gromacs/math/utilities.h"
  46 #include "gromacs/math/vec.h"
  47 #include "gromacs/mdtypes/commrec.h"
  48 #include "gromacs/mdtypes/forcerec.h"
  49 #include "gromacs/mdtypes/md_enums.h"
  50 #include "gromacs/utility/fatalerror.h"
  51 #include "gromacs/utility/gmxassert.h"
  52
  53 /* There's nothing special to do here if just masses are perturbed,
  54  * but if either charge or type is perturbed then the implementation
  55  * requires that B states are defined for both charge and type, and
  56  * does not optimize for the cases where only one changes.
  57  *
  58  * The parameter vectors for B states are left undefined in atoms2md()
  59  * when either FEP is inactive, or when there are no mass/charge/type
  60  * perturbations. The parameter vectors for LJ-PME are likewise
  61  * undefined when LJ-PME is not active. This works because
  62  * bHaveChargeOrTypePerturbed handles the control flow. */
  63 void ewald_LRcorrection(int numAtomsLocal,
  64                         t_commrec *cr,
  65                         int numThreads, int thread,
  66                         t_forcerec *fr,
  67                         real *chargeA, real *chargeB,
  68                         real *C6A, real *C6B,
  69                         real *sigmaA, real *sigmaB,
  70                         real *sigma3A, real *sigma3B,
  71                         gmx_bool bHaveChargeOrTypePerturbed,
  72                         gmx_bool calc_excl_corr,
  73                         t_blocka *excl, rvec x[],
  74                         matrix box, rvec mu_tot[],
  75                         int ewald_geometry, real epsilon_surface,
  76                         rvec *f, tensor vir_q, tensor vir_lj,
  77                         real *Vcorr_q, real *Vcorr_lj,
  78                         real lambda_q, real lambda_lj,
  79                         real *dvdlambda_q, real *dvdlambda_lj)
  80 {
  81     int numAtomsToBeCorrected;
  82     if (calc_excl_corr)
  83     {
  84         /* We need to correct all exclusion pairs (cutoff-scheme = group) */
  85         numAtomsToBeCorrected = excl->nr;
  86
  87         GMX_RELEASE_ASSERT(numAtomsToBeCorrected >= numAtomsLocal, "We might need to do self-corrections");
  88     }
  89     else
  90     {
  91         /* We need to correct only self interactions */
  92         numAtomsToBeCorrected = numAtomsLocal;
  93     }
  94     int         start =  (numAtomsToBeCorrected* thread     )/numThreads;
  95     int         end   =  (numAtomsToBeCorrected*(thread + 1))/numThreads;
  96
  97     int         i, i1, i2, j, k, m, iv, jv, q;
  98     int        *AA;
  99     double      Vexcl_q, dvdl_excl_q, dvdl_excl_lj; /* Necessary for precision */
 100     double      Vexcl_lj;
 101     real        one_4pi_eps;
 102     real        v, vc, qiA, qiB, dr2, rinv;
 103     real        Vself_q[2], Vself_lj[2], Vdipole[2], rinv2, ewc_q = fr->ewaldcoeff_q, ewcdr;
 104     real        ewc_lj = fr->ewaldcoeff_lj, ewc_lj2 = ewc_lj * ewc_lj;
 105     real        c6Ai   = 0, c6Bi = 0, c6A = 0, c6B = 0, ewcdr2, ewcdr4, c6L = 0, rinv6;
 106     rvec        df, dx, mutot[2], dipcorrA, dipcorrB;
 107     tensor      dxdf_q = {{0}}, dxdf_lj = {{0}};
 108     real        vol    = box[XX][XX]*box[YY][YY]*box[ZZ][ZZ];
 109     real        L1_q, L1_lj, dipole_coeff, qqA, qqB, qqL, vr0_q, vr0_lj = 0;
 110     gmx_bool    bMolPBC      = fr->bMolPBC;
 111     gmx_bool    bDoingLBRule = (fr->ljpme_combination_rule == eljpmeLB);
 112     gmx_bool    bNeedLongRangeCorrection;
 113
 114     /* This routine can be made faster by using tables instead of analytical interactions
 115      * However, that requires a thorough verification that they are correct in all cases.
 116      */
 117
 118     one_4pi_eps   = ONE_4PI_EPS0/fr->epsilon_r;
 119     vr0_q         = ewc_q*M_2_SQRTPI;
 120     if (EVDW_PME(fr->vdwtype))
 121     {
 122         vr0_lj    = -gmx::power6(ewc_lj)/6.0;
 123     }
 124
 125     AA           = excl->a;
 126     Vexcl_q      = 0;
 127     Vexcl_lj     = 0;
 128     dvdl_excl_q  = 0;
 129     dvdl_excl_lj = 0;
 130     Vdipole[0]   = 0;
 131     Vdipole[1]   = 0;
 132     L1_q         = 1.0-lambda_q;
 133     L1_lj        = 1.0-lambda_lj;
 134     /* Note that we have to transform back to gromacs units, since
 135      * mu_tot contains the dipole in debye units (for output).
 136      */
 137     for (i = 0; (i < DIM); i++)
 138     {
 139         mutot[0][i] = mu_tot[0][i]*DEBYE2ENM;
 140         mutot[1][i] = mu_tot[1][i]*DEBYE2ENM;
 141         dipcorrA[i] = 0;
 142         dipcorrB[i] = 0;
 143     }
 144     dipole_coeff = 0;
 145     switch (ewald_geometry)
 146     {
 147         case eewg3D:
 148             if (epsilon_surface != 0)
 149             {
 150                 dipole_coeff =
 151                     2*M_PI*ONE_4PI_EPS0/((2*epsilon_surface + fr->epsilon_r)*vol);
 152                 for (i = 0; (i < DIM); i++)
 153                 {
 154                     dipcorrA[i] = 2*dipole_coeff*mutot[0][i];
 155                     dipcorrB[i] = 2*dipole_coeff*mutot[1][i];
 156                 }
 157             }
 158             break;
 159         case eewg3DC:
 160             dipole_coeff = 2*M_PI*one_4pi_eps/vol;
 161             dipcorrA[ZZ] = 2*dipole_coeff*mutot[0][ZZ];
 162             dipcorrB[ZZ] = 2*dipole_coeff*mutot[1][ZZ];
 163             break;
 164         default:
 165             gmx_incons("Unsupported Ewald geometry");
 166             break;
 167     }
 168     if (debug)
 169     {
 170         fprintf(debug, "dipcorr = %8.3f  %8.3f  %8.3f\n",
 171                 dipcorrA[XX], dipcorrA[YY], dipcorrA[ZZ]);
 172         fprintf(debug, "mutot   = %8.3f  %8.3f  %8.3f\n",
 173                 mutot[0][XX], mutot[0][YY], mutot[0][ZZ]);
 174     }
 175     bNeedLongRangeCorrection = (calc_excl_corr || dipole_coeff != 0);
 176     if (bNeedLongRangeCorrection && !bHaveChargeOrTypePerturbed)
 177     {
 178         for (i = start; (i < end); i++)
 179         {
 180             /* Initiate local variables (for this i-particle) to 0 */
 181             qiA = chargeA[i]*one_4pi_eps;
 182             if (EVDW_PME(fr->vdwtype))
 183             {
 184                 c6Ai = C6A[i];
 185                 if (bDoingLBRule)
 186                 {
 187                     c6Ai *= sigma3A[i];
 188                 }
 189             }
 190             if (calc_excl_corr)
 191             {
 192                 i1  = excl->index[i];
 193                 i2  = excl->index[i+1];
 194
 195                 /* Loop over excluded neighbours */
 196                 for (j = i1; (j < i2); j++)
 197                 {
 198                     k = AA[j];
 199                     /*
 200                      * First we must test whether k <> i, and then,
 201                      * because the exclusions are all listed twice i->k
 202                      * and k->i we must select just one of the two.  As
 203                      * a minor optimization we only compute forces when
 204                      * the charges are non-zero.
 205                      */
 206                     if (k > i)
 207                     {
 208                         qqA = qiA*chargeA[k];
 209                         if (EVDW_PME(fr->vdwtype))
 210                         {
 211                             c6A  = c6Ai * C6A[k];
 212                             if (bDoingLBRule)
 213                             {
 214                                 c6A *= gmx::power6(0.5*(sigmaA[i]+sigmaA[k]))*sigma3A[k];
 215                             }
 216                         }
 217                         if (qqA != 0.0 || c6A != 0.0)
 218                         {
 219                             rvec_sub(x[i], x[k], dx);
 220                             if (bMolPBC)
 221                             {
 222                                 /* Cheap pbc_dx, assume excluded pairs are at short distance. */
 223                                 for (m = DIM-1; (m >= 0); m--)
 224                                 {
 225                                     if (dx[m] > 0.5*box[m][m])
 226                                     {
 227                                         rvec_dec(dx, box[m]);
 228                                     }
 229                                     else if (dx[m] < -0.5*box[m][m])
 230                                     {
 231                                         rvec_inc(dx, box[m]);
 232                                     }
 233                                 }
 234                             }
 235                             dr2 = norm2(dx);
 236                             /* Distance between two excluded particles
 237                              * may be zero in the case of shells
 238                              */
 239                             if (dr2 != 0)
 240                             {
 241                                 rinv              = gmx::invsqrt(dr2);
 242                                 rinv2             = rinv*rinv;
 243                                 if (qqA != 0.0)
 244                                 {
 245                                     real dr, fscal;
 246
 247                                     dr       = 1.0/rinv;
 248                                     ewcdr    = ewc_q*dr;
 249                                     vc       = qqA*std::erf(ewcdr)*rinv;
 250                                     Vexcl_q += vc;
 251 #if GMX_DOUBLE
 252                                     /* Relative accuracy at R_ERF_R_INACC of 3e-10 */
 253 #define       R_ERF_R_INACC 0.006
 254 #else
 255                                     /* Relative accuracy at R_ERF_R_INACC of 2e-5 */
 256 #define       R_ERF_R_INACC 0.1
 257 #endif
 258                                     /* fscal is the scalar force pre-multiplied by rinv,
 259                                      * to normalise the relative position vector dx */
 260                                     if (ewcdr > R_ERF_R_INACC)
 261                                     {
 262                                         fscal = rinv2*(vc - qqA*ewc_q*M_2_SQRTPI*exp(-ewcdr*ewcdr));
 263                                     }
 264                                     else
 265                                     {
 266                                         /* Use a fourth order series expansion for small ewcdr */
 267                                         fscal = ewc_q*ewc_q*qqA*vr0_q*(2.0/3.0 - 0.4*ewcdr*ewcdr);
 268                                     }
 269
 270                                     /* The force vector is obtained by multiplication with
 271                                      * the relative position vector
 272                                      */
 273                                     svmul(fscal, dx, df);
 274                                     rvec_inc(f[k], df);
 275                                     rvec_dec(f[i], df);
 276                                     for (iv = 0; (iv < DIM); iv++)
 277                                     {
 278                                         for (jv = 0; (jv < DIM); jv++)
 279                                         {
 280                                             dxdf_q[iv][jv] += dx[iv]*df[jv];
 281                                         }
 282                                     }
 283                                 }
 284
 285                                 if (c6A != 0.0)
 286                                 {
 287                                     real fscal;
 288
 289                                     rinv6     = rinv2*rinv2*rinv2;
 290                                     ewcdr2    = ewc_lj2*dr2;
 291                                     ewcdr4    = ewcdr2*ewcdr2;
 292                                     /* We get the excluded long-range contribution from -C6*(1-g(r))
 293                                      * g(r) is also defined in the manual under LJ-PME
 294                                      */
 295                                     vc        = -c6A*rinv6*(1.0 - exp(-ewcdr2)*(1 + ewcdr2 + 0.5*ewcdr4));
 296                                     Vexcl_lj += vc;
 297                                     /* The force is the derivative of the potential vc.
 298                                      * fscal is the scalar force pre-multiplied by rinv,
 299                                      * to normalise the relative position vector dx */
 300                                     fscal     = 6.0*vc*rinv2 + c6A*rinv6*exp(-ewcdr2)*ewc_lj2*ewcdr4;
 301
 302                                     /* The force vector is obtained by multiplication with
 303                                      * the relative position vector
 304                                      */
 305                                     svmul(fscal, dx, df);
 306                                     rvec_inc(f[k], df);
 307                                     rvec_dec(f[i], df);
 308                                     for (iv = 0; (iv < DIM); iv++)
 309                                     {
 310                                         for (jv = 0; (jv < DIM); jv++)
 311                                         {
 312                                             dxdf_lj[iv][jv] += dx[iv]*df[jv];
 313                                         }
 314                                     }
 315                                 }
 316                             }
 317                             else
 318                             {
 319                                 Vexcl_q  += qqA*vr0_q;
 320                                 Vexcl_lj += c6A*vr0_lj;
 321                             }
 322                         }
 323                     }
 324                 }
 325             }
 326             /* Dipole correction on force */
 327             if (dipole_coeff != 0 && i < numAtomsLocal)
 328             {
 329                 for (j = 0; (j < DIM); j++)
 330                 {
 331                     f[i][j] -= dipcorrA[j]*chargeA[i];
 332                 }
 333             }
 334         }
 335     }
 336     else if (bNeedLongRangeCorrection)
 337     {
 338         for (i = start; (i < end); i++)
 339         {
 340             /* Initiate local variables (for this i-particle) to 0 */
 341             qiA = chargeA[i]*one_4pi_eps;
 342             qiB = chargeB[i]*one_4pi_eps;
 343             if (EVDW_PME(fr->vdwtype))
 344             {
 345                 c6Ai = C6A[i];
 346                 c6Bi = C6B[i];
 347                 if (bDoingLBRule)
 348                 {
 349                     c6Ai *= sigma3A[i];
 350                     c6Bi *= sigma3B[i];
 351                 }
 352             }
 353             if (calc_excl_corr)
 354             {
 355                 i1  = excl->index[i];
 356                 i2  = excl->index[i+1];
 357
 358                 /* Loop over excluded neighbours */
 359                 for (j = i1; (j < i2); j++)
 360                 {
 361                     k = AA[j];
 362                     if (k > i)
 363                     {
 364                         qqA = qiA*chargeA[k];
 365                         qqB = qiB*chargeB[k];
 366                         if (EVDW_PME(fr->vdwtype))
 367                         {
 368                             c6A = c6Ai*C6A[k];
 369                             c6B = c6Bi*C6B[k];
 370                             if (bDoingLBRule)
 371                             {
 372                                 c6A *= gmx::power6(0.5*(sigmaA[i]+sigmaA[k]))*sigma3A[k];
 373                                 c6B *= gmx::power6(0.5*(sigmaB[i]+sigmaB[k]))*sigma3B[k];
 374                             }
 375                         }
 376                         if (qqA != 0.0 || qqB != 0.0 || c6A != 0.0 || c6B != 0.0)
 377                         {
 378                             real fscal;
 379
 380                             qqL   = L1_q*qqA + lambda_q*qqB;
 381                             if (EVDW_PME(fr->vdwtype))
 382                             {
 383                                 c6L = L1_lj*c6A + lambda_lj*c6B;
 384                             }
 385                             rvec_sub(x[i], x[k], dx);
 386                             if (bMolPBC)
 387                             {
 388                                 /* Cheap pbc_dx, assume excluded pairs are at short distance. */
 389                                 for (m = DIM-1; (m >= 0); m--)
 390                                 {
 391                                     if (dx[m] > 0.5*box[m][m])
 392                                     {
 393                                         rvec_dec(dx, box[m]);
 394                                     }
 395                                     else if (dx[m] < -0.5*box[m][m])
 396                                     {
 397                                         rvec_inc(dx, box[m]);
 398                                     }
 399                                 }
 400                             }
 401                             dr2 = norm2(dx);
 402                             if (dr2 != 0)
 403                             {
 404                                 rinv    = gmx::invsqrt(dr2);
 405                                 rinv2   = rinv*rinv;
 406                                 if (qqA != 0.0 || qqB != 0.0)
 407                                 {
 408                                     real dr;
 409
 410                                     dr           = 1.0/rinv;
 411                                     v            = std::erf(ewc_q*dr)*rinv;
 412                                     vc           = qqL*v;
 413                                     Vexcl_q     += vc;
 414                                     /* fscal is the scalar force pre-multiplied by rinv,
 415                                      * to normalise the relative position vector dx */
 416                                     fscal        = rinv2*(vc-qqL*ewc_q*M_2_SQRTPI*exp(-ewc_q*ewc_q*dr2));
 417                                     dvdl_excl_q += (qqB - qqA)*v;
 418
 419                                     /* The force vector is obtained by multiplication with
 420                                      * the relative position vector
 421                                      */
 422                                     svmul(fscal, dx, df);
 423                                     rvec_inc(f[k], df);
 424                                     rvec_dec(f[i], df);
 425                                     for (iv = 0; (iv < DIM); iv++)
 426                                     {
 427                                         for (jv = 0; (jv < DIM); jv++)
 428                                         {
 429                                             dxdf_q[iv][jv] += dx[iv]*df[jv];
 430                                         }
 431                                     }
 432                                 }
 433
 434                                 if ((c6A != 0.0 || c6B != 0.0) && EVDW_PME(fr->vdwtype))
 435                                 {
 436                                     rinv6         = rinv2*rinv2*rinv2;
 437                                     ewcdr2        = ewc_lj2*dr2;
 438                                     ewcdr4        = ewcdr2*ewcdr2;
 439                                     v             = -rinv6*(1.0 - exp(-ewcdr2)*(1 + ewcdr2 + 0.5*ewcdr4));
 440                                     vc            = c6L*v;
 441                                     Vexcl_lj     += vc;
 442                                     /* fscal is the scalar force pre-multiplied by rinv,
 443                                      * to normalise the relative position vector dx */
 444                                     fscal         = 6.0*vc*rinv2 + c6L*rinv6*exp(-ewcdr2)*ewc_lj2*ewcdr4;
 445                                     dvdl_excl_lj += (c6B - c6A)*v;
 446
 447                                     /* The force vector is obtained by multiplication with
 448                                      * the relative position vector
 449                                      */
 450                                     svmul(fscal, dx, df);
 451                                     rvec_inc(f[k], df);
 452                                     rvec_dec(f[i], df);
 453                                     for (iv = 0; (iv < DIM); iv++)
 454                                     {
 455                                         for (jv = 0; (jv < DIM); jv++)
 456                                         {
 457                                             dxdf_lj[iv][jv] += dx[iv]*df[jv];
 458                                         }
 459                                     }
 460                                 }
 461                             }
 462                             else
 463                             {
 464                                 Vexcl_q      += qqL*vr0_q;
 465                                 dvdl_excl_q  += (qqB - qqA)*vr0_q;
 466                                 Vexcl_lj     += c6L*vr0_lj;
 467                                 dvdl_excl_lj += (c6B - c6A)*vr0_lj;
 468                             }
 469                         }
 470                     }
 471                 }
 472             }
 473             /* Dipole correction on force */
 474             if (dipole_coeff != 0 && i < numAtomsLocal)
 475             {
 476                 for (j = 0; (j < DIM); j++)
 477                 {
 478                     f[i][j] -= L1_q*dipcorrA[j]*chargeA[i]
 479                         + lambda_q*dipcorrB[j]*chargeB[i];
 480                 }
 481             }
 482         }
 483     }
 484     for (iv = 0; (iv < DIM); iv++)
 485     {
 486         for (jv = 0; (jv < DIM); jv++)
 487         {
 488             vir_q[iv][jv]  += 0.5*dxdf_q[iv][jv];
 489             vir_lj[iv][jv] += 0.5*dxdf_lj[iv][jv];
 490         }
 491     }
 492
 493     Vself_q[0]  = 0;
 494     Vself_q[1]  = 0;
 495     Vself_lj[0] = 0;
 496     Vself_lj[1] = 0;
 497
 498     /* Global corrections only on master process */
 499     if (MASTER(cr) && thread == 0)
 500     {
 501         for (q = 0; q < (bHaveChargeOrTypePerturbed ? 2 : 1); q++)
 502         {
 503             if (calc_excl_corr)
 504             {
 505                 /* Self-energy correction */
 506                 Vself_q[q] = ewc_q*one_4pi_eps*fr->q2sum[q]*M_1_SQRTPI;
 507                 if (EVDW_PME(fr->vdwtype))
 508                 {
 509                     Vself_lj[q] =  fr->c6sum[q]*0.5*vr0_lj;
 510                 }
 511             }
 512
 513             /* Apply surface dipole correction:
 514              * correction = dipole_coeff * (dipole)^2
 515              */
 516             if (dipole_coeff != 0)
 517             {
 518                 if (ewald_geometry == eewg3D)
 519                 {
 520                     Vdipole[q] = dipole_coeff*iprod(mutot[q], mutot[q]);
 521                 }
 522                 else if (ewald_geometry == eewg3DC)
 523                 {
 524                     Vdipole[q] = dipole_coeff*mutot[q][ZZ]*mutot[q][ZZ];
 525                 }
 526             }
 527         }
 528     }
 529     if (!bHaveChargeOrTypePerturbed)
 530     {
 531         *Vcorr_q = Vdipole[0] - Vself_q[0] - Vexcl_q;
 532         if (EVDW_PME(fr->vdwtype))
 533         {
 534             *Vcorr_lj = -Vself_lj[0] - Vexcl_lj;
 535         }
 536     }
 537     else
 538     {
 539         *Vcorr_q = L1_q*(Vdipole[0] - Vself_q[0])
 540             + lambda_q*(Vdipole[1] - Vself_q[1])
 541             - Vexcl_q;
 542         *dvdlambda_q += Vdipole[1] - Vself_q[1]
 543             - (Vdipole[0] - Vself_q[0]) - dvdl_excl_q;
 544         if (EVDW_PME(fr->vdwtype))
 545         {
 546             *Vcorr_lj      = -(L1_lj*Vself_lj[0] + lambda_lj*Vself_lj[1]) - Vexcl_lj;
 547             *dvdlambda_lj += -Vself_lj[1] + Vself_lj[0] - dvdl_excl_lj;
 548         }
 549     }
 550
 551     if (debug)
 552     {
 553         fprintf(debug, "Long Range corrections for Ewald interactions:\n");
 554         fprintf(debug, "q2sum = %g, Vself_q=%g c6sum = %g, Vself_lj=%g\n",
 555                 L1_q*fr->q2sum[0]+lambda_q*fr->q2sum[1], L1_q*Vself_q[0]+lambda_q*Vself_q[1], L1_lj*fr->c6sum[0]+lambda_lj*fr->c6sum[1], L1_lj*Vself_lj[0]+lambda_lj*Vself_lj[1]);
 556         fprintf(debug, "Electrostatic Long Range correction: Vexcl=%g\n", Vexcl_q);
 557         fprintf(debug, "Lennard-Jones Long Range correction: Vexcl=%g\n", Vexcl_lj);
 558         if (MASTER(cr) && thread == 0)
 559         {
 560             if (epsilon_surface > 0 || ewald_geometry == eewg3DC)
 561             {
 562                 fprintf(debug, "Total dipole correction: Vdipole=%g\n",
 563                         L1_q*Vdipole[0]+lambda_q*Vdipole[1]);
 564             }
 565         }
 566     }
 567 }