src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwLJEw_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_ElecEw_VdwLJEw_GeomP1P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            LJEwald
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEw_VdwLJEw_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              nvdwtype;
  80     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  81     int              *vdwtype;
  82     real             *vdwparam;
  83     real             c6grid_00;
  84     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
  85     real             *vdwgridparam;
  86     int              ewitab;
  87     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  88     real             *ewtab;
  89
  90     x                = xx[0];
  91     f                = ff[0];
  92
  93     nri              = nlist->nri;
  94     iinr             = nlist->iinr;
  95     jindex           = nlist->jindex;
  96     jjnr             = nlist->jjnr;
  97     shiftidx         = nlist->shift;
  98     gid              = nlist->gid;
  99     shiftvec         = fr->shift_vec[0];
 100     fshift           = fr->fshift[0];
 101     facel            = fr->epsfac;
 102     charge           = mdatoms->chargeA;
 103     nvdwtype         = fr->ntype;
 104     vdwparam         = fr->nbfp;
 105     vdwtype          = mdatoms->typeA;
 106     vdwgridparam     = fr->ljpme_c6grid;
 107     ewclj            = fr->ewaldcoeff_lj;
 108     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 109     ewclj2           = ewclj*ewclj;
 110     ewclj6           = ewclj2*ewclj2*ewclj2;
 111
 112     sh_ewald         = fr->ic->sh_ewald;
 113     ewtab            = fr->ic->tabq_coul_FDV0;
 114     ewtabscale       = fr->ic->tabq_scale;
 115     ewtabhalfspace   = 0.5/ewtabscale;
 116
 117     outeriter        = 0;
 118     inneriter        = 0;
 119
 120     /* Start outer loop over neighborlists */
 121     for(iidx=0; iidx<nri; iidx++)
 122     {
 123         /* Load shift vector for this list */
 124         i_shift_offset   = DIM*shiftidx[iidx];
 125         shX              = shiftvec[i_shift_offset+XX];
 126         shY              = shiftvec[i_shift_offset+YY];
 127         shZ              = shiftvec[i_shift_offset+ZZ];
 128
 129         /* Load limits for loop over neighbors */
 130         j_index_start    = jindex[iidx];
 131         j_index_end      = jindex[iidx+1];
 132
 133         /* Get outer coordinate index */
 134         inr              = iinr[iidx];
 135         i_coord_offset   = DIM*inr;
 136
 137         /* Load i particle coords and add shift vector */
 138         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 139         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 140         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 141
 142         fix0             = 0.0;
 143         fiy0             = 0.0;
 144         fiz0             = 0.0;
 145
 146         /* Load parameters for i particles */
 147         iq0              = facel*charge[inr+0];
 148         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 149
 150         /* Reset potential sums */
 151         velecsum         = 0.0;
 152         vvdwsum          = 0.0;
 153
 154         /* Start inner kernel loop */
 155         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 156         {
 157             /* Get j neighbor index, and coordinate index */
 158             jnr              = jjnr[jidx];
 159             j_coord_offset   = DIM*jnr;
 160
 161             /* load j atom coordinates */
 162             jx0              = x[j_coord_offset+DIM*0+XX];
 163             jy0              = x[j_coord_offset+DIM*0+YY];
 164             jz0              = x[j_coord_offset+DIM*0+ZZ];
 165
 166             /* Calculate displacement vector */
 167             dx00             = ix0 - jx0;
 168             dy00             = iy0 - jy0;
 169             dz00             = iz0 - jz0;
 170
 171             /* Calculate squared distance and things based on it */
 172             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 173
 174             rinv00           = gmx_invsqrt(rsq00);
 175
 176             rinvsq00         = rinv00*rinv00;
 177
 178             /* Load parameters for j particles */
 179             jq0              = charge[jnr+0];
 180             vdwjidx0         = 2*vdwtype[jnr+0];
 181
 182             /**************************
 183              * CALCULATE INTERACTIONS *
 184              **************************/
 185
 186             r00              = rsq00*rinv00;
 187
 188             qq00             = iq0*jq0;
 189             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 190             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 191             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 192
 193             /* EWALD ELECTROSTATICS */
 194
 195             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 196             ewrt             = r00*ewtabscale;
 197             ewitab           = ewrt;
 198             eweps            = ewrt-ewitab;
 199             ewitab           = 4*ewitab;
 200             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 201             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 202             felec            = qq00*rinv00*(rinvsq00-felec);
 203
 204             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 205             ewcljrsq         = ewclj2*rsq00;
 206             exponent         = exp(-ewcljrsq);
 207             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 208             vvdw6            = (c6_00-c6grid_00*(1.0-poly))*rinvsix;
 209             vvdw12           = c12_00*rinvsix*rinvsix;
 210             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 211             fvdw             = (vvdw12 - vvdw6 - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 212
 213             /* Update potential sums from outer loop */
 214             velecsum        += velec;
 215             vvdwsum         += vvdw;
 216
 217             fscal            = felec+fvdw;
 218
 219             /* Calculate temporary vectorial force */
 220             tx               = fscal*dx00;
 221             ty               = fscal*dy00;
 222             tz               = fscal*dz00;
 223
 224             /* Update vectorial force */
 225             fix0            += tx;
 226             fiy0            += ty;
 227             fiz0            += tz;
 228             f[j_coord_offset+DIM*0+XX] -= tx;
 229             f[j_coord_offset+DIM*0+YY] -= ty;
 230             f[j_coord_offset+DIM*0+ZZ] -= tz;
 231
 232             /* Inner loop uses 67 flops */
 233         }
 234         /* End of innermost loop */
 235
 236         tx = ty = tz = 0;
 237         f[i_coord_offset+DIM*0+XX] += fix0;
 238         f[i_coord_offset+DIM*0+YY] += fiy0;
 239         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 240         tx                         += fix0;
 241         ty                         += fiy0;
 242         tz                         += fiz0;
 243         fshift[i_shift_offset+XX]  += tx;
 244         fshift[i_shift_offset+YY]  += ty;
 245         fshift[i_shift_offset+ZZ]  += tz;
 246
 247         ggid                        = gid[iidx];
 248         /* Update potential energies */
 249         kernel_data->energygrp_elec[ggid] += velecsum;
 250         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 251
 252         /* Increment number of inner iterations */
 253         inneriter                  += j_index_end - j_index_start;
 254
 255         /* Outer loop uses 15 flops */
 256     }
 257
 258     /* Increment number of outer iterations */
 259     outeriter        += nri;
 260
 261     /* Update outer/inner flops */
 262
 263     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*67);
 264 }
 265 /*
 266  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomP1P1_F_c
 267  * Electrostatics interaction: Ewald
 268  * VdW interaction:            LJEwald
 269  * Geometry:                   Particle-Particle
 270  * Calculate force/pot:        Force
 271  */
 272 void
 273 nb_kernel_ElecEw_VdwLJEw_GeomP1P1_F_c
 274                     (t_nblist                    * gmx_restrict       nlist,
 275                      rvec                        * gmx_restrict          xx,
 276                      rvec                        * gmx_restrict          ff,
 277                      t_forcerec                  * gmx_restrict          fr,
 278                      t_mdatoms                   * gmx_restrict     mdatoms,
 279                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 280                      t_nrnb                      * gmx_restrict        nrnb)
 281 {
 282     int              i_shift_offset,i_coord_offset,j_coord_offset;
 283     int              j_index_start,j_index_end;
 284     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 285     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 286     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 287     real             *shiftvec,*fshift,*x,*f;
 288     int              vdwioffset0;
 289     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 290     int              vdwjidx0;
 291     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 292     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 293     real             velec,felec,velecsum,facel,crf,krf,krf2;
 294     real             *charge;
 295     int              nvdwtype;
 296     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 297     int              *vdwtype;
 298     real             *vdwparam;
 299     real             c6grid_00;
 300     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
 301     real             *vdwgridparam;
 302     int              ewitab;
 303     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 304     real             *ewtab;
 305
 306     x                = xx[0];
 307     f                = ff[0];
 308
 309     nri              = nlist->nri;
 310     iinr             = nlist->iinr;
 311     jindex           = nlist->jindex;
 312     jjnr             = nlist->jjnr;
 313     shiftidx         = nlist->shift;
 314     gid              = nlist->gid;
 315     shiftvec         = fr->shift_vec[0];
 316     fshift           = fr->fshift[0];
 317     facel            = fr->epsfac;
 318     charge           = mdatoms->chargeA;
 319     nvdwtype         = fr->ntype;
 320     vdwparam         = fr->nbfp;
 321     vdwtype          = mdatoms->typeA;
 322     vdwgridparam     = fr->ljpme_c6grid;
 323     ewclj            = fr->ewaldcoeff_lj;
 324     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 325     ewclj2           = ewclj*ewclj;
 326     ewclj6           = ewclj2*ewclj2*ewclj2;
 327
 328     sh_ewald         = fr->ic->sh_ewald;
 329     ewtab            = fr->ic->tabq_coul_F;
 330     ewtabscale       = fr->ic->tabq_scale;
 331     ewtabhalfspace   = 0.5/ewtabscale;
 332
 333     outeriter        = 0;
 334     inneriter        = 0;
 335
 336     /* Start outer loop over neighborlists */
 337     for(iidx=0; iidx<nri; iidx++)
 338     {
 339         /* Load shift vector for this list */
 340         i_shift_offset   = DIM*shiftidx[iidx];
 341         shX              = shiftvec[i_shift_offset+XX];
 342         shY              = shiftvec[i_shift_offset+YY];
 343         shZ              = shiftvec[i_shift_offset+ZZ];
 344
 345         /* Load limits for loop over neighbors */
 346         j_index_start    = jindex[iidx];
 347         j_index_end      = jindex[iidx+1];
 348
 349         /* Get outer coordinate index */
 350         inr              = iinr[iidx];
 351         i_coord_offset   = DIM*inr;
 352
 353         /* Load i particle coords and add shift vector */
 354         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 355         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 356         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 357
 358         fix0             = 0.0;
 359         fiy0             = 0.0;
 360         fiz0             = 0.0;
 361
 362         /* Load parameters for i particles */
 363         iq0              = facel*charge[inr+0];
 364         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 365
 366         /* Start inner kernel loop */
 367         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 368         {
 369             /* Get j neighbor index, and coordinate index */
 370             jnr              = jjnr[jidx];
 371             j_coord_offset   = DIM*jnr;
 372
 373             /* load j atom coordinates */
 374             jx0              = x[j_coord_offset+DIM*0+XX];
 375             jy0              = x[j_coord_offset+DIM*0+YY];
 376             jz0              = x[j_coord_offset+DIM*0+ZZ];
 377
 378             /* Calculate displacement vector */
 379             dx00             = ix0 - jx0;
 380             dy00             = iy0 - jy0;
 381             dz00             = iz0 - jz0;
 382
 383             /* Calculate squared distance and things based on it */
 384             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 385
 386             rinv00           = gmx_invsqrt(rsq00);
 387
 388             rinvsq00         = rinv00*rinv00;
 389
 390             /* Load parameters for j particles */
 391             jq0              = charge[jnr+0];
 392             vdwjidx0         = 2*vdwtype[jnr+0];
 393
 394             /**************************
 395              * CALCULATE INTERACTIONS *
 396              **************************/
 397
 398             r00              = rsq00*rinv00;
 399
 400             qq00             = iq0*jq0;
 401             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 402             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 403             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 404
 405             /* EWALD ELECTROSTATICS */
 406
 407             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 408             ewrt             = r00*ewtabscale;
 409             ewitab           = ewrt;
 410             eweps            = ewrt-ewitab;
 411             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 412             felec            = qq00*rinv00*(rinvsq00-felec);
 413
 414             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 415             ewcljrsq         = ewclj2*rsq00;
 416             exponent         = exp(-ewcljrsq);
 417             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 418             fvdw             = (((c12_00*rinvsix - c6_00 + c6grid_00*(1.0-poly))*rinvsix) - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 419
 420             fscal            = felec+fvdw;
 421
 422             /* Calculate temporary vectorial force */
 423             tx               = fscal*dx00;
 424             ty               = fscal*dy00;
 425             tz               = fscal*dz00;
 426
 427             /* Update vectorial force */
 428             fix0            += tx;
 429             fiy0            += ty;
 430             fiz0            += tz;
 431             f[j_coord_offset+DIM*0+XX] -= tx;
 432             f[j_coord_offset+DIM*0+YY] -= ty;
 433             f[j_coord_offset+DIM*0+ZZ] -= tz;
 434
 435             /* Inner loop uses 55 flops */
 436         }
 437         /* End of innermost loop */
 438
 439         tx = ty = tz = 0;
 440         f[i_coord_offset+DIM*0+XX] += fix0;
 441         f[i_coord_offset+DIM*0+YY] += fiy0;
 442         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 443         tx                         += fix0;
 444         ty                         += fiy0;
 445         tz                         += fiz0;
 446         fshift[i_shift_offset+XX]  += tx;
 447         fshift[i_shift_offset+YY]  += ty;
 448         fshift[i_shift_offset+ZZ]  += tz;
 449
 450         /* Increment number of inner iterations */
 451         inneriter                  += j_index_end - j_index_start;
 452
 453         /* Outer loop uses 13 flops */
 454     }
 455
 456     /* Increment number of outer iterations */
 457     outeriter        += nri;
 458
 459     /* Update outer/inner flops */
 460
 461     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*55);
 462 }