src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecGB_VdwNone_GeomP1P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS c kernel generator.
  37  */
  38 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "types/simple.h"
  46 #include "gromacs/math/vec.h"
  47 #include "nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwNone_GeomP1P1_VF_c
  51  * Electrostatics interaction: GeneralizedBorn
  52  * VdW interaction:            None
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecGB_VdwNone_GeomP1P1_VF_c
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      t_forcerec                  * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     int              i_shift_offset,i_coord_offset,j_coord_offset;
  67     int              j_index_start,j_index_end;
  68     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  69     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  70     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  71     real             *shiftvec,*fshift,*x,*f;
  72     int              vdwioffset0;
  73     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  74     int              vdwjidx0;
  75     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  76     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  77     real             velec,felec,velecsum,facel,crf,krf,krf2;
  78     real             *charge;
  79     int              gbitab;
  80     real             vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,dvdaj,gbeps,dvdatmp;
  81     real             *invsqrta,*dvda,*gbtab;
  82     int              vfitab;
  83     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  84     real             *vftab;
  85
  86     x                = xx[0];
  87     f                = ff[0];
  88
  89     nri              = nlist->nri;
  90     iinr             = nlist->iinr;
  91     jindex           = nlist->jindex;
  92     jjnr             = nlist->jjnr;
  93     shiftidx         = nlist->shift;
  94     gid              = nlist->gid;
  95     shiftvec         = fr->shift_vec[0];
  96     fshift           = fr->fshift[0];
  97     facel            = fr->epsfac;
  98     charge           = mdatoms->chargeA;
  99
 100     invsqrta         = fr->invsqrta;
 101     dvda             = fr->dvda;
 102     gbtabscale       = fr->gbtab.scale;
 103     gbtab            = fr->gbtab.data;
 104     gbinvepsdiff     = (1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent);
 105
 106     outeriter        = 0;
 107     inneriter        = 0;
 108
 109     /* Start outer loop over neighborlists */
 110     for(iidx=0; iidx<nri; iidx++)
 111     {
 112         /* Load shift vector for this list */
 113         i_shift_offset   = DIM*shiftidx[iidx];
 114         shX              = shiftvec[i_shift_offset+XX];
 115         shY              = shiftvec[i_shift_offset+YY];
 116         shZ              = shiftvec[i_shift_offset+ZZ];
 117
 118         /* Load limits for loop over neighbors */
 119         j_index_start    = jindex[iidx];
 120         j_index_end      = jindex[iidx+1];
 121
 122         /* Get outer coordinate index */
 123         inr              = iinr[iidx];
 124         i_coord_offset   = DIM*inr;
 125
 126         /* Load i particle coords and add shift vector */
 127         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 128         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 129         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 130
 131         fix0             = 0.0;
 132         fiy0             = 0.0;
 133         fiz0             = 0.0;
 134
 135         /* Load parameters for i particles */
 136         iq0              = facel*charge[inr+0];
 137         isai0            = invsqrta[inr+0];
 138
 139         /* Reset potential sums */
 140         velecsum         = 0.0;
 141         vgbsum           = 0.0;
 142         dvdasum          = 0.0;
 143
 144         /* Start inner kernel loop */
 145         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 146         {
 147             /* Get j neighbor index, and coordinate index */
 148             jnr              = jjnr[jidx];
 149             j_coord_offset   = DIM*jnr;
 150
 151             /* load j atom coordinates */
 152             jx0              = x[j_coord_offset+DIM*0+XX];
 153             jy0              = x[j_coord_offset+DIM*0+YY];
 154             jz0              = x[j_coord_offset+DIM*0+ZZ];
 155
 156             /* Calculate displacement vector */
 157             dx00             = ix0 - jx0;
 158             dy00             = iy0 - jy0;
 159             dz00             = iz0 - jz0;
 160
 161             /* Calculate squared distance and things based on it */
 162             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 163
 164             rinv00           = gmx_invsqrt(rsq00);
 165
 166             /* Load parameters for j particles */
 167             jq0              = charge[jnr+0];
 168             isaj0           = invsqrta[jnr+0];
 169
 170             /**************************
 171              * CALCULATE INTERACTIONS *
 172              **************************/
 173
 174             r00              = rsq00*rinv00;
 175
 176             qq00             = iq0*jq0;
 177
 178             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 179             isaprod          = isai0*isaj0;
 180             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 181             gbscale          = isaprod*gbtabscale;
 182             dvdaj            = dvda[jnr+0];
 183
 184             /* Calculate generalized born table index - this is a separate table from the normal one,
 185              * but we use the same procedure by multiplying r with scale and truncating to integer.
 186              */
 187             rt               = r00*gbscale;
 188             gbitab           = rt;
 189             gbeps            = rt-gbitab;
 190             gbitab           = 4*gbitab;
 191
 192             Y                = gbtab[gbitab];
 193             F                = gbtab[gbitab+1];
 194             Geps             = gbeps*gbtab[gbitab+2];
 195             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 196             Fp               = F+Geps+Heps2;
 197             VV               = Y+gbeps*Fp;
 198             vgb              = gbqqfactor*VV;
 199
 200             FF               = Fp+Geps+2.0*Heps2;
 201             fgb              = gbqqfactor*FF*gbscale;
 202             dvdatmp          = -0.5*(vgb+fgb*r00);
 203             dvdasum          = dvdasum + dvdatmp;
 204             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 205             velec            = qq00*rinv00;
 206             felec            = (velec*rinv00-fgb)*rinv00;
 207
 208             /* Update potential sums from outer loop */
 209             velecsum        += velec;
 210             vgbsum          += vgb;
 211
 212             fscal            = felec;
 213
 214             /* Calculate temporary vectorial force */
 215             tx               = fscal*dx00;
 216             ty               = fscal*dy00;
 217             tz               = fscal*dz00;
 218
 219             /* Update vectorial force */
 220             fix0            += tx;
 221             fiy0            += ty;
 222             fiz0            += tz;
 223             f[j_coord_offset+DIM*0+XX] -= tx;
 224             f[j_coord_offset+DIM*0+YY] -= ty;
 225             f[j_coord_offset+DIM*0+ZZ] -= tz;
 226
 227             /* Inner loop uses 58 flops */
 228         }
 229         /* End of innermost loop */
 230
 231         tx = ty = tz = 0;
 232         f[i_coord_offset+DIM*0+XX] += fix0;
 233         f[i_coord_offset+DIM*0+YY] += fiy0;
 234         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 235         tx                         += fix0;
 236         ty                         += fiy0;
 237         tz                         += fiz0;
 238         fshift[i_shift_offset+XX]  += tx;
 239         fshift[i_shift_offset+YY]  += ty;
 240         fshift[i_shift_offset+ZZ]  += tz;
 241
 242         ggid                        = gid[iidx];
 243         /* Update potential energies */
 244         kernel_data->energygrp_elec[ggid] += velecsum;
 245         kernel_data->energygrp_polarization[ggid] += vgbsum;
 246         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 247
 248         /* Increment number of inner iterations */
 249         inneriter                  += j_index_end - j_index_start;
 250
 251         /* Outer loop uses 15 flops */
 252     }
 253
 254     /* Increment number of outer iterations */
 255     outeriter        += nri;
 256
 257     /* Update outer/inner flops */
 258
 259     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VF,outeriter*15 + inneriter*58);
 260 }
 261 /*
 262  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwNone_GeomP1P1_F_c
 263  * Electrostatics interaction: GeneralizedBorn
 264  * VdW interaction:            None
 265  * Geometry:                   Particle-Particle
 266  * Calculate force/pot:        Force
 267  */
 268 void
 269 nb_kernel_ElecGB_VdwNone_GeomP1P1_F_c
 270                     (t_nblist                    * gmx_restrict       nlist,
 271                      rvec                        * gmx_restrict          xx,
 272                      rvec                        * gmx_restrict          ff,
 273                      t_forcerec                  * gmx_restrict          fr,
 274                      t_mdatoms                   * gmx_restrict     mdatoms,
 275                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 276                      t_nrnb                      * gmx_restrict        nrnb)
 277 {
 278     int              i_shift_offset,i_coord_offset,j_coord_offset;
 279     int              j_index_start,j_index_end;
 280     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 281     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 282     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 283     real             *shiftvec,*fshift,*x,*f;
 284     int              vdwioffset0;
 285     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 286     int              vdwjidx0;
 287     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 288     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 289     real             velec,felec,velecsum,facel,crf,krf,krf2;
 290     real             *charge;
 291     int              gbitab;
 292     real             vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,dvdaj,gbeps,dvdatmp;
 293     real             *invsqrta,*dvda,*gbtab;
 294     int              vfitab;
 295     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 296     real             *vftab;
 297
 298     x                = xx[0];
 299     f                = ff[0];
 300
 301     nri              = nlist->nri;
 302     iinr             = nlist->iinr;
 303     jindex           = nlist->jindex;
 304     jjnr             = nlist->jjnr;
 305     shiftidx         = nlist->shift;
 306     gid              = nlist->gid;
 307     shiftvec         = fr->shift_vec[0];
 308     fshift           = fr->fshift[0];
 309     facel            = fr->epsfac;
 310     charge           = mdatoms->chargeA;
 311
 312     invsqrta         = fr->invsqrta;
 313     dvda             = fr->dvda;
 314     gbtabscale       = fr->gbtab.scale;
 315     gbtab            = fr->gbtab.data;
 316     gbinvepsdiff     = (1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent);
 317
 318     outeriter        = 0;
 319     inneriter        = 0;
 320
 321     /* Start outer loop over neighborlists */
 322     for(iidx=0; iidx<nri; iidx++)
 323     {
 324         /* Load shift vector for this list */
 325         i_shift_offset   = DIM*shiftidx[iidx];
 326         shX              = shiftvec[i_shift_offset+XX];
 327         shY              = shiftvec[i_shift_offset+YY];
 328         shZ              = shiftvec[i_shift_offset+ZZ];
 329
 330         /* Load limits for loop over neighbors */
 331         j_index_start    = jindex[iidx];
 332         j_index_end      = jindex[iidx+1];
 333
 334         /* Get outer coordinate index */
 335         inr              = iinr[iidx];
 336         i_coord_offset   = DIM*inr;
 337
 338         /* Load i particle coords and add shift vector */
 339         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 340         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 341         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 342
 343         fix0             = 0.0;
 344         fiy0             = 0.0;
 345         fiz0             = 0.0;
 346
 347         /* Load parameters for i particles */
 348         iq0              = facel*charge[inr+0];
 349         isai0            = invsqrta[inr+0];
 350
 351         dvdasum          = 0.0;
 352
 353         /* Start inner kernel loop */
 354         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 355         {
 356             /* Get j neighbor index, and coordinate index */
 357             jnr              = jjnr[jidx];
 358             j_coord_offset   = DIM*jnr;
 359
 360             /* load j atom coordinates */
 361             jx0              = x[j_coord_offset+DIM*0+XX];
 362             jy0              = x[j_coord_offset+DIM*0+YY];
 363             jz0              = x[j_coord_offset+DIM*0+ZZ];
 364
 365             /* Calculate displacement vector */
 366             dx00             = ix0 - jx0;
 367             dy00             = iy0 - jy0;
 368             dz00             = iz0 - jz0;
 369
 370             /* Calculate squared distance and things based on it */
 371             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 372
 373             rinv00           = gmx_invsqrt(rsq00);
 374
 375             /* Load parameters for j particles */
 376             jq0              = charge[jnr+0];
 377             isaj0           = invsqrta[jnr+0];
 378
 379             /**************************
 380              * CALCULATE INTERACTIONS *
 381              **************************/
 382
 383             r00              = rsq00*rinv00;
 384
 385             qq00             = iq0*jq0;
 386
 387             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 388             isaprod          = isai0*isaj0;
 389             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 390             gbscale          = isaprod*gbtabscale;
 391             dvdaj            = dvda[jnr+0];
 392
 393             /* Calculate generalized born table index - this is a separate table from the normal one,
 394              * but we use the same procedure by multiplying r with scale and truncating to integer.
 395              */
 396             rt               = r00*gbscale;
 397             gbitab           = rt;
 398             gbeps            = rt-gbitab;
 399             gbitab           = 4*gbitab;
 400
 401             Y                = gbtab[gbitab];
 402             F                = gbtab[gbitab+1];
 403             Geps             = gbeps*gbtab[gbitab+2];
 404             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 405             Fp               = F+Geps+Heps2;
 406             VV               = Y+gbeps*Fp;
 407             vgb              = gbqqfactor*VV;
 408
 409             FF               = Fp+Geps+2.0*Heps2;
 410             fgb              = gbqqfactor*FF*gbscale;
 411             dvdatmp          = -0.5*(vgb+fgb*r00);
 412             dvdasum          = dvdasum + dvdatmp;
 413             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 414             velec            = qq00*rinv00;
 415             felec            = (velec*rinv00-fgb)*rinv00;
 416
 417             fscal            = felec;
 418
 419             /* Calculate temporary vectorial force */
 420             tx               = fscal*dx00;
 421             ty               = fscal*dy00;
 422             tz               = fscal*dz00;
 423
 424             /* Update vectorial force */
 425             fix0            += tx;
 426             fiy0            += ty;
 427             fiz0            += tz;
 428             f[j_coord_offset+DIM*0+XX] -= tx;
 429             f[j_coord_offset+DIM*0+YY] -= ty;
 430             f[j_coord_offset+DIM*0+ZZ] -= tz;
 431
 432             /* Inner loop uses 56 flops */
 433         }
 434         /* End of innermost loop */
 435
 436         tx = ty = tz = 0;
 437         f[i_coord_offset+DIM*0+XX] += fix0;
 438         f[i_coord_offset+DIM*0+YY] += fiy0;
 439         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 440         tx                         += fix0;
 441         ty                         += fiy0;
 442         tz                         += fiz0;
 443         fshift[i_shift_offset+XX]  += tx;
 444         fshift[i_shift_offset+YY]  += ty;
 445         fshift[i_shift_offset+ZZ]  += tz;
 446
 447         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 448
 449         /* Increment number of inner iterations */
 450         inneriter                  += j_index_end - j_index_start;
 451
 452         /* Outer loop uses 13 flops */
 453     }
 454
 455     /* Increment number of outer iterations */
 456     outeriter        += nri;
 457
 458     /* Update outer/inner flops */
 459
 460     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_F,outeriter*13 + inneriter*56);
 461 }