src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSh_VdwLJSh_GeomP1P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013, 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 "vec.h"
  47 #include "nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJSh_GeomP1P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEwSh_VdwLJSh_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     int              ewitab;
  84     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  85     real             *ewtab;
  86
  87     x                = xx[0];
  88     f                = ff[0];
  89
  90     nri              = nlist->nri;
  91     iinr             = nlist->iinr;
  92     jindex           = nlist->jindex;
  93     jjnr             = nlist->jjnr;
  94     shiftidx         = nlist->shift;
  95     gid              = nlist->gid;
  96     shiftvec         = fr->shift_vec[0];
  97     fshift           = fr->fshift[0];
  98     facel            = fr->epsfac;
  99     charge           = mdatoms->chargeA;
 100     nvdwtype         = fr->ntype;
 101     vdwparam         = fr->nbfp;
 102     vdwtype          = mdatoms->typeA;
 103
 104     sh_ewald         = fr->ic->sh_ewald;
 105     ewtab            = fr->ic->tabq_coul_FDV0;
 106     ewtabscale       = fr->ic->tabq_scale;
 107     ewtabhalfspace   = 0.5/ewtabscale;
 108
 109     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 110     rcutoff          = fr->rcoulomb;
 111     rcutoff2         = rcutoff*rcutoff;
 112
 113     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 114     rvdw             = fr->rvdw;
 115
 116     outeriter        = 0;
 117     inneriter        = 0;
 118
 119     /* Start outer loop over neighborlists */
 120     for(iidx=0; iidx<nri; iidx++)
 121     {
 122         /* Load shift vector for this list */
 123         i_shift_offset   = DIM*shiftidx[iidx];
 124         shX              = shiftvec[i_shift_offset+XX];
 125         shY              = shiftvec[i_shift_offset+YY];
 126         shZ              = shiftvec[i_shift_offset+ZZ];
 127
 128         /* Load limits for loop over neighbors */
 129         j_index_start    = jindex[iidx];
 130         j_index_end      = jindex[iidx+1];
 131
 132         /* Get outer coordinate index */
 133         inr              = iinr[iidx];
 134         i_coord_offset   = DIM*inr;
 135
 136         /* Load i particle coords and add shift vector */
 137         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 138         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 139         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 140
 141         fix0             = 0.0;
 142         fiy0             = 0.0;
 143         fiz0             = 0.0;
 144
 145         /* Load parameters for i particles */
 146         iq0              = facel*charge[inr+0];
 147         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 148
 149         /* Reset potential sums */
 150         velecsum         = 0.0;
 151         vvdwsum          = 0.0;
 152
 153         /* Start inner kernel loop */
 154         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 155         {
 156             /* Get j neighbor index, and coordinate index */
 157             jnr              = jjnr[jidx];
 158             j_coord_offset   = DIM*jnr;
 159
 160             /* load j atom coordinates */
 161             jx0              = x[j_coord_offset+DIM*0+XX];
 162             jy0              = x[j_coord_offset+DIM*0+YY];
 163             jz0              = x[j_coord_offset+DIM*0+ZZ];
 164
 165             /* Calculate displacement vector */
 166             dx00             = ix0 - jx0;
 167             dy00             = iy0 - jy0;
 168             dz00             = iz0 - jz0;
 169
 170             /* Calculate squared distance and things based on it */
 171             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 172
 173             rinv00           = gmx_invsqrt(rsq00);
 174
 175             rinvsq00         = rinv00*rinv00;
 176
 177             /* Load parameters for j particles */
 178             jq0              = charge[jnr+0];
 179             vdwjidx0         = 2*vdwtype[jnr+0];
 180
 181             /**************************
 182              * CALCULATE INTERACTIONS *
 183              **************************/
 184
 185             if (rsq00<rcutoff2)
 186             {
 187
 188             r00              = rsq00*rinv00;
 189
 190             qq00             = iq0*jq0;
 191             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 192             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 193
 194             /* EWALD ELECTROSTATICS */
 195
 196             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 197             ewrt             = r00*ewtabscale;
 198             ewitab           = ewrt;
 199             eweps            = ewrt-ewitab;
 200             ewitab           = 4*ewitab;
 201             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 202             velec            = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 203             felec            = qq00*rinv00*(rinvsq00-felec);
 204
 205             /* LENNARD-JONES DISPERSION/REPULSION */
 206
 207             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 208             vvdw6            = c6_00*rinvsix;
 209             vvdw12           = c12_00*rinvsix*rinvsix;
 210             vvdw             = (vvdw12 - c12_00*sh_vdw_invrcut6*sh_vdw_invrcut6)*(1.0/12.0) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
 211             fvdw             = (vvdw12-vvdw6)*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             }
 233
 234             /* Inner loop uses 59 flops */
 235         }
 236         /* End of innermost loop */
 237
 238         tx = ty = tz = 0;
 239         f[i_coord_offset+DIM*0+XX] += fix0;
 240         f[i_coord_offset+DIM*0+YY] += fiy0;
 241         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 242         tx                         += fix0;
 243         ty                         += fiy0;
 244         tz                         += fiz0;
 245         fshift[i_shift_offset+XX]  += tx;
 246         fshift[i_shift_offset+YY]  += ty;
 247         fshift[i_shift_offset+ZZ]  += tz;
 248
 249         ggid                        = gid[iidx];
 250         /* Update potential energies */
 251         kernel_data->energygrp_elec[ggid] += velecsum;
 252         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 253
 254         /* Increment number of inner iterations */
 255         inneriter                  += j_index_end - j_index_start;
 256
 257         /* Outer loop uses 15 flops */
 258     }
 259
 260     /* Increment number of outer iterations */
 261     outeriter        += nri;
 262
 263     /* Update outer/inner flops */
 264
 265     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*59);
 266 }
 267 /*
 268  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJSh_GeomP1P1_F_c
 269  * Electrostatics interaction: Ewald
 270  * VdW interaction:            LennardJones
 271  * Geometry:                   Particle-Particle
 272  * Calculate force/pot:        Force
 273  */
 274 void
 275 nb_kernel_ElecEwSh_VdwLJSh_GeomP1P1_F_c
 276                     (t_nblist                    * gmx_restrict       nlist,
 277                      rvec                        * gmx_restrict          xx,
 278                      rvec                        * gmx_restrict          ff,
 279                      t_forcerec                  * gmx_restrict          fr,
 280                      t_mdatoms                   * gmx_restrict     mdatoms,
 281                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 282                      t_nrnb                      * gmx_restrict        nrnb)
 283 {
 284     int              i_shift_offset,i_coord_offset,j_coord_offset;
 285     int              j_index_start,j_index_end;
 286     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 287     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 288     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 289     real             *shiftvec,*fshift,*x,*f;
 290     int              vdwioffset0;
 291     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 292     int              vdwjidx0;
 293     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 294     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 295     real             velec,felec,velecsum,facel,crf,krf,krf2;
 296     real             *charge;
 297     int              nvdwtype;
 298     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 299     int              *vdwtype;
 300     real             *vdwparam;
 301     int              ewitab;
 302     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 303     real             *ewtab;
 304
 305     x                = xx[0];
 306     f                = ff[0];
 307
 308     nri              = nlist->nri;
 309     iinr             = nlist->iinr;
 310     jindex           = nlist->jindex;
 311     jjnr             = nlist->jjnr;
 312     shiftidx         = nlist->shift;
 313     gid              = nlist->gid;
 314     shiftvec         = fr->shift_vec[0];
 315     fshift           = fr->fshift[0];
 316     facel            = fr->epsfac;
 317     charge           = mdatoms->chargeA;
 318     nvdwtype         = fr->ntype;
 319     vdwparam         = fr->nbfp;
 320     vdwtype          = mdatoms->typeA;
 321
 322     sh_ewald         = fr->ic->sh_ewald;
 323     ewtab            = fr->ic->tabq_coul_F;
 324     ewtabscale       = fr->ic->tabq_scale;
 325     ewtabhalfspace   = 0.5/ewtabscale;
 326
 327     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 328     rcutoff          = fr->rcoulomb;
 329     rcutoff2         = rcutoff*rcutoff;
 330
 331     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 332     rvdw             = fr->rvdw;
 333
 334     outeriter        = 0;
 335     inneriter        = 0;
 336
 337     /* Start outer loop over neighborlists */
 338     for(iidx=0; iidx<nri; iidx++)
 339     {
 340         /* Load shift vector for this list */
 341         i_shift_offset   = DIM*shiftidx[iidx];
 342         shX              = shiftvec[i_shift_offset+XX];
 343         shY              = shiftvec[i_shift_offset+YY];
 344         shZ              = shiftvec[i_shift_offset+ZZ];
 345
 346         /* Load limits for loop over neighbors */
 347         j_index_start    = jindex[iidx];
 348         j_index_end      = jindex[iidx+1];
 349
 350         /* Get outer coordinate index */
 351         inr              = iinr[iidx];
 352         i_coord_offset   = DIM*inr;
 353
 354         /* Load i particle coords and add shift vector */
 355         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 356         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 357         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 358
 359         fix0             = 0.0;
 360         fiy0             = 0.0;
 361         fiz0             = 0.0;
 362
 363         /* Load parameters for i particles */
 364         iq0              = facel*charge[inr+0];
 365         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 366
 367         /* Start inner kernel loop */
 368         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 369         {
 370             /* Get j neighbor index, and coordinate index */
 371             jnr              = jjnr[jidx];
 372             j_coord_offset   = DIM*jnr;
 373
 374             /* load j atom coordinates */
 375             jx0              = x[j_coord_offset+DIM*0+XX];
 376             jy0              = x[j_coord_offset+DIM*0+YY];
 377             jz0              = x[j_coord_offset+DIM*0+ZZ];
 378
 379             /* Calculate displacement vector */
 380             dx00             = ix0 - jx0;
 381             dy00             = iy0 - jy0;
 382             dz00             = iz0 - jz0;
 383
 384             /* Calculate squared distance and things based on it */
 385             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 386
 387             rinv00           = gmx_invsqrt(rsq00);
 388
 389             rinvsq00         = rinv00*rinv00;
 390
 391             /* Load parameters for j particles */
 392             jq0              = charge[jnr+0];
 393             vdwjidx0         = 2*vdwtype[jnr+0];
 394
 395             /**************************
 396              * CALCULATE INTERACTIONS *
 397              **************************/
 398
 399             if (rsq00<rcutoff2)
 400             {
 401
 402             r00              = rsq00*rinv00;
 403
 404             qq00             = iq0*jq0;
 405             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 406             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 407
 408             /* EWALD ELECTROSTATICS */
 409
 410             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 411             ewrt             = r00*ewtabscale;
 412             ewitab           = ewrt;
 413             eweps            = ewrt-ewitab;
 414             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 415             felec            = qq00*rinv00*(rinvsq00-felec);
 416
 417             /* LENNARD-JONES DISPERSION/REPULSION */
 418
 419             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 420             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 421
 422             fscal            = felec+fvdw;
 423
 424             /* Calculate temporary vectorial force */
 425             tx               = fscal*dx00;
 426             ty               = fscal*dy00;
 427             tz               = fscal*dz00;
 428
 429             /* Update vectorial force */
 430             fix0            += tx;
 431             fiy0            += ty;
 432             fiz0            += tz;
 433             f[j_coord_offset+DIM*0+XX] -= tx;
 434             f[j_coord_offset+DIM*0+YY] -= ty;
 435             f[j_coord_offset+DIM*0+ZZ] -= tz;
 436
 437             }
 438
 439             /* Inner loop uses 41 flops */
 440         }
 441         /* End of innermost loop */
 442
 443         tx = ty = tz = 0;
 444         f[i_coord_offset+DIM*0+XX] += fix0;
 445         f[i_coord_offset+DIM*0+YY] += fiy0;
 446         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 447         tx                         += fix0;
 448         ty                         += fiy0;
 449         tz                         += fiz0;
 450         fshift[i_shift_offset+XX]  += tx;
 451         fshift[i_shift_offset+YY]  += ty;
 452         fshift[i_shift_offset+ZZ]  += tz;
 453
 454         /* Increment number of inner iterations */
 455         inneriter                  += j_index_end - j_index_start;
 456
 457         /* Outer loop uses 13 flops */
 458     }
 459
 460     /* Increment number of outer iterations */
 461     outeriter        += nri;
 462
 463     /* Update outer/inner flops */
 464
 465     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*41);
 466 }