src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwLJ_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 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJ_GeomP1P1_VF_c
  49  * Electrostatics interaction: Ewald
  50  * VdW interaction:            LennardJones
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecEw_VdwLJ_GeomP1P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      t_forcerec                  * gmx_restrict          fr,
  60                      t_mdatoms                   * gmx_restrict     mdatoms,
  61                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  62                      t_nrnb                      * gmx_restrict        nrnb)
  63 {
  64     int              i_shift_offset,i_coord_offset,j_coord_offset;
  65     int              j_index_start,j_index_end;
  66     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  67     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  68     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  69     real             *shiftvec,*fshift,*x,*f;
  70     int              vdwioffset0;
  71     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     int              vdwjidx0;
  73     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  74     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  75     real             velec,felec,velecsum,facel,crf,krf,krf2;
  76     real             *charge;
  77     int              nvdwtype;
  78     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  79     int              *vdwtype;
  80     real             *vdwparam;
  81     int              ewitab;
  82     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  83     real             *ewtab;
  84
  85     x                = xx[0];
  86     f                = ff[0];
  87
  88     nri              = nlist->nri;
  89     iinr             = nlist->iinr;
  90     jindex           = nlist->jindex;
  91     jjnr             = nlist->jjnr;
  92     shiftidx         = nlist->shift;
  93     gid              = nlist->gid;
  94     shiftvec         = fr->shift_vec[0];
  95     fshift           = fr->fshift[0];
  96     facel            = fr->epsfac;
  97     charge           = mdatoms->chargeA;
  98     nvdwtype         = fr->ntype;
  99     vdwparam         = fr->nbfp;
 100     vdwtype          = mdatoms->typeA;
 101
 102     sh_ewald         = fr->ic->sh_ewald;
 103     ewtab            = fr->ic->tabq_coul_FDV0;
 104     ewtabscale       = fr->ic->tabq_scale;
 105     ewtabhalfspace   = 0.5/ewtabscale;
 106
 107     outeriter        = 0;
 108     inneriter        = 0;
 109
 110     /* Start outer loop over neighborlists */
 111     for(iidx=0; iidx<nri; iidx++)
 112     {
 113         /* Load shift vector for this list */
 114         i_shift_offset   = DIM*shiftidx[iidx];
 115         shX              = shiftvec[i_shift_offset+XX];
 116         shY              = shiftvec[i_shift_offset+YY];
 117         shZ              = shiftvec[i_shift_offset+ZZ];
 118
 119         /* Load limits for loop over neighbors */
 120         j_index_start    = jindex[iidx];
 121         j_index_end      = jindex[iidx+1];
 122
 123         /* Get outer coordinate index */
 124         inr              = iinr[iidx];
 125         i_coord_offset   = DIM*inr;
 126
 127         /* Load i particle coords and add shift vector */
 128         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 129         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 130         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 131
 132         fix0             = 0.0;
 133         fiy0             = 0.0;
 134         fiz0             = 0.0;
 135
 136         /* Load parameters for i particles */
 137         iq0              = facel*charge[inr+0];
 138         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 139
 140         /* Reset potential sums */
 141         velecsum         = 0.0;
 142         vvdwsum          = 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             rinvsq00         = rinv00*rinv00;
 167
 168             /* Load parameters for j particles */
 169             jq0              = charge[jnr+0];
 170             vdwjidx0         = 2*vdwtype[jnr+0];
 171
 172             /**************************
 173              * CALCULATE INTERACTIONS *
 174              **************************/
 175
 176             r00              = rsq00*rinv00;
 177
 178             qq00             = iq0*jq0;
 179             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 180             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 181
 182             /* EWALD ELECTROSTATICS */
 183
 184             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 185             ewrt             = r00*ewtabscale;
 186             ewitab           = ewrt;
 187             eweps            = ewrt-ewitab;
 188             ewitab           = 4*ewitab;
 189             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 190             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 191             felec            = qq00*rinv00*(rinvsq00-felec);
 192
 193             /* LENNARD-JONES DISPERSION/REPULSION */
 194
 195             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 196             vvdw6            = c6_00*rinvsix;
 197             vvdw12           = c12_00*rinvsix*rinvsix;
 198             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 199             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 200
 201             /* Update potential sums from outer loop */
 202             velecsum        += velec;
 203             vvdwsum         += vvdw;
 204
 205             fscal            = felec+fvdw;
 206
 207             /* Calculate temporary vectorial force */
 208             tx               = fscal*dx00;
 209             ty               = fscal*dy00;
 210             tz               = fscal*dz00;
 211
 212             /* Update vectorial force */
 213             fix0            += tx;
 214             fiy0            += ty;
 215             fiz0            += tz;
 216             f[j_coord_offset+DIM*0+XX] -= tx;
 217             f[j_coord_offset+DIM*0+YY] -= ty;
 218             f[j_coord_offset+DIM*0+ZZ] -= tz;
 219
 220             /* Inner loop uses 53 flops */
 221         }
 222         /* End of innermost loop */
 223
 224         tx = ty = tz = 0;
 225         f[i_coord_offset+DIM*0+XX] += fix0;
 226         f[i_coord_offset+DIM*0+YY] += fiy0;
 227         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 228         tx                         += fix0;
 229         ty                         += fiy0;
 230         tz                         += fiz0;
 231         fshift[i_shift_offset+XX]  += tx;
 232         fshift[i_shift_offset+YY]  += ty;
 233         fshift[i_shift_offset+ZZ]  += tz;
 234
 235         ggid                        = gid[iidx];
 236         /* Update potential energies */
 237         kernel_data->energygrp_elec[ggid] += velecsum;
 238         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 239
 240         /* Increment number of inner iterations */
 241         inneriter                  += j_index_end - j_index_start;
 242
 243         /* Outer loop uses 15 flops */
 244     }
 245
 246     /* Increment number of outer iterations */
 247     outeriter        += nri;
 248
 249     /* Update outer/inner flops */
 250
 251     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*53);
 252 }
 253 /*
 254  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJ_GeomP1P1_F_c
 255  * Electrostatics interaction: Ewald
 256  * VdW interaction:            LennardJones
 257  * Geometry:                   Particle-Particle
 258  * Calculate force/pot:        Force
 259  */
 260 void
 261 nb_kernel_ElecEw_VdwLJ_GeomP1P1_F_c
 262                     (t_nblist                    * gmx_restrict       nlist,
 263                      rvec                        * gmx_restrict          xx,
 264                      rvec                        * gmx_restrict          ff,
 265                      t_forcerec                  * gmx_restrict          fr,
 266                      t_mdatoms                   * gmx_restrict     mdatoms,
 267                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 268                      t_nrnb                      * gmx_restrict        nrnb)
 269 {
 270     int              i_shift_offset,i_coord_offset,j_coord_offset;
 271     int              j_index_start,j_index_end;
 272     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 273     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 274     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 275     real             *shiftvec,*fshift,*x,*f;
 276     int              vdwioffset0;
 277     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 278     int              vdwjidx0;
 279     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 280     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 281     real             velec,felec,velecsum,facel,crf,krf,krf2;
 282     real             *charge;
 283     int              nvdwtype;
 284     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 285     int              *vdwtype;
 286     real             *vdwparam;
 287     int              ewitab;
 288     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 289     real             *ewtab;
 290
 291     x                = xx[0];
 292     f                = ff[0];
 293
 294     nri              = nlist->nri;
 295     iinr             = nlist->iinr;
 296     jindex           = nlist->jindex;
 297     jjnr             = nlist->jjnr;
 298     shiftidx         = nlist->shift;
 299     gid              = nlist->gid;
 300     shiftvec         = fr->shift_vec[0];
 301     fshift           = fr->fshift[0];
 302     facel            = fr->epsfac;
 303     charge           = mdatoms->chargeA;
 304     nvdwtype         = fr->ntype;
 305     vdwparam         = fr->nbfp;
 306     vdwtype          = mdatoms->typeA;
 307
 308     sh_ewald         = fr->ic->sh_ewald;
 309     ewtab            = fr->ic->tabq_coul_F;
 310     ewtabscale       = fr->ic->tabq_scale;
 311     ewtabhalfspace   = 0.5/ewtabscale;
 312
 313     outeriter        = 0;
 314     inneriter        = 0;
 315
 316     /* Start outer loop over neighborlists */
 317     for(iidx=0; iidx<nri; iidx++)
 318     {
 319         /* Load shift vector for this list */
 320         i_shift_offset   = DIM*shiftidx[iidx];
 321         shX              = shiftvec[i_shift_offset+XX];
 322         shY              = shiftvec[i_shift_offset+YY];
 323         shZ              = shiftvec[i_shift_offset+ZZ];
 324
 325         /* Load limits for loop over neighbors */
 326         j_index_start    = jindex[iidx];
 327         j_index_end      = jindex[iidx+1];
 328
 329         /* Get outer coordinate index */
 330         inr              = iinr[iidx];
 331         i_coord_offset   = DIM*inr;
 332
 333         /* Load i particle coords and add shift vector */
 334         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 335         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 336         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 337
 338         fix0             = 0.0;
 339         fiy0             = 0.0;
 340         fiz0             = 0.0;
 341
 342         /* Load parameters for i particles */
 343         iq0              = facel*charge[inr+0];
 344         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 345
 346         /* Start inner kernel loop */
 347         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 348         {
 349             /* Get j neighbor index, and coordinate index */
 350             jnr              = jjnr[jidx];
 351             j_coord_offset   = DIM*jnr;
 352
 353             /* load j atom coordinates */
 354             jx0              = x[j_coord_offset+DIM*0+XX];
 355             jy0              = x[j_coord_offset+DIM*0+YY];
 356             jz0              = x[j_coord_offset+DIM*0+ZZ];
 357
 358             /* Calculate displacement vector */
 359             dx00             = ix0 - jx0;
 360             dy00             = iy0 - jy0;
 361             dz00             = iz0 - jz0;
 362
 363             /* Calculate squared distance and things based on it */
 364             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 365
 366             rinv00           = gmx_invsqrt(rsq00);
 367
 368             rinvsq00         = rinv00*rinv00;
 369
 370             /* Load parameters for j particles */
 371             jq0              = charge[jnr+0];
 372             vdwjidx0         = 2*vdwtype[jnr+0];
 373
 374             /**************************
 375              * CALCULATE INTERACTIONS *
 376              **************************/
 377
 378             r00              = rsq00*rinv00;
 379
 380             qq00             = iq0*jq0;
 381             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 382             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 383
 384             /* EWALD ELECTROSTATICS */
 385
 386             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 387             ewrt             = r00*ewtabscale;
 388             ewitab           = ewrt;
 389             eweps            = ewrt-ewitab;
 390             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 391             felec            = qq00*rinv00*(rinvsq00-felec);
 392
 393             /* LENNARD-JONES DISPERSION/REPULSION */
 394
 395             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 396             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 397
 398             fscal            = felec+fvdw;
 399
 400             /* Calculate temporary vectorial force */
 401             tx               = fscal*dx00;
 402             ty               = fscal*dy00;
 403             tz               = fscal*dz00;
 404
 405             /* Update vectorial force */
 406             fix0            += tx;
 407             fiy0            += ty;
 408             fiz0            += tz;
 409             f[j_coord_offset+DIM*0+XX] -= tx;
 410             f[j_coord_offset+DIM*0+YY] -= ty;
 411             f[j_coord_offset+DIM*0+ZZ] -= tz;
 412
 413             /* Inner loop uses 41 flops */
 414         }
 415         /* End of innermost loop */
 416
 417         tx = ty = tz = 0;
 418         f[i_coord_offset+DIM*0+XX] += fix0;
 419         f[i_coord_offset+DIM*0+YY] += fiy0;
 420         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 421         tx                         += fix0;
 422         ty                         += fiy0;
 423         tz                         += fiz0;
 424         fshift[i_shift_offset+XX]  += tx;
 425         fshift[i_shift_offset+YY]  += ty;
 426         fshift[i_shift_offset+ZZ]  += tz;
 427
 428         /* Increment number of inner iterations */
 429         inneriter                  += j_index_end - j_index_start;
 430
 431         /* Outer loop uses 13 flops */
 432     }
 433
 434     /* Increment number of outer iterations */
 435     outeriter        += nri;
 436
 437     /* Update outer/inner flops */
 438
 439     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*41);
 440 }