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 #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_VdwLJEw_GeomP1P1_VF_c
  49  * Electrostatics interaction: Ewald
  50  * VdW interaction:            LJEwald
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecEw_VdwLJEw_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     real             c6grid_00;
  82     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
  83     real             *vdwgridparam;
  84     int              ewitab;
  85     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  86     real             *ewtab;
  87
  88     x                = xx[0];
  89     f                = ff[0];
  90
  91     nri              = nlist->nri;
  92     iinr             = nlist->iinr;
  93     jindex           = nlist->jindex;
  94     jjnr             = nlist->jjnr;
  95     shiftidx         = nlist->shift;
  96     gid              = nlist->gid;
  97     shiftvec         = fr->shift_vec[0];
  98     fshift           = fr->fshift[0];
  99     facel            = fr->epsfac;
 100     charge           = mdatoms->chargeA;
 101     nvdwtype         = fr->ntype;
 102     vdwparam         = fr->nbfp;
 103     vdwtype          = mdatoms->typeA;
 104     vdwgridparam     = fr->ljpme_c6grid;
 105     ewclj            = fr->ewaldcoeff_lj;
 106     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 107     ewclj2           = ewclj*ewclj;
 108     ewclj6           = ewclj2*ewclj2*ewclj2;
 109
 110     sh_ewald         = fr->ic->sh_ewald;
 111     ewtab            = fr->ic->tabq_coul_FDV0;
 112     ewtabscale       = fr->ic->tabq_scale;
 113     ewtabhalfspace   = 0.5/ewtabscale;
 114
 115     outeriter        = 0;
 116     inneriter        = 0;
 117
 118     /* Start outer loop over neighborlists */
 119     for(iidx=0; iidx<nri; iidx++)
 120     {
 121         /* Load shift vector for this list */
 122         i_shift_offset   = DIM*shiftidx[iidx];
 123         shX              = shiftvec[i_shift_offset+XX];
 124         shY              = shiftvec[i_shift_offset+YY];
 125         shZ              = shiftvec[i_shift_offset+ZZ];
 126
 127         /* Load limits for loop over neighbors */
 128         j_index_start    = jindex[iidx];
 129         j_index_end      = jindex[iidx+1];
 130
 131         /* Get outer coordinate index */
 132         inr              = iinr[iidx];
 133         i_coord_offset   = DIM*inr;
 134
 135         /* Load i particle coords and add shift vector */
 136         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 137         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 138         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 139
 140         fix0             = 0.0;
 141         fiy0             = 0.0;
 142         fiz0             = 0.0;
 143
 144         /* Load parameters for i particles */
 145         iq0              = facel*charge[inr+0];
 146         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 147
 148         /* Reset potential sums */
 149         velecsum         = 0.0;
 150         vvdwsum          = 0.0;
 151
 152         /* Start inner kernel loop */
 153         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 154         {
 155             /* Get j neighbor index, and coordinate index */
 156             jnr              = jjnr[jidx];
 157             j_coord_offset   = DIM*jnr;
 158
 159             /* load j atom coordinates */
 160             jx0              = x[j_coord_offset+DIM*0+XX];
 161             jy0              = x[j_coord_offset+DIM*0+YY];
 162             jz0              = x[j_coord_offset+DIM*0+ZZ];
 163
 164             /* Calculate displacement vector */
 165             dx00             = ix0 - jx0;
 166             dy00             = iy0 - jy0;
 167             dz00             = iz0 - jz0;
 168
 169             /* Calculate squared distance and things based on it */
 170             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 171
 172             rinv00           = gmx_invsqrt(rsq00);
 173
 174             rinvsq00         = rinv00*rinv00;
 175
 176             /* Load parameters for j particles */
 177             jq0              = charge[jnr+0];
 178             vdwjidx0         = 2*vdwtype[jnr+0];
 179
 180             /**************************
 181              * CALCULATE INTERACTIONS *
 182              **************************/
 183
 184             r00              = rsq00*rinv00;
 185
 186             qq00             = iq0*jq0;
 187             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 188             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 189             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 190
 191             /* EWALD ELECTROSTATICS */
 192
 193             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 194             ewrt             = r00*ewtabscale;
 195             ewitab           = ewrt;
 196             eweps            = ewrt-ewitab;
 197             ewitab           = 4*ewitab;
 198             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 199             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 200             felec            = qq00*rinv00*(rinvsq00-felec);
 201
 202             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 203             ewcljrsq         = ewclj2*rsq00;
 204             exponent         = exp(-ewcljrsq);
 205             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 206             vvdw6            = (c6_00-c6grid_00*(1.0-poly))*rinvsix;
 207             vvdw12           = c12_00*rinvsix*rinvsix;
 208             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 209             fvdw             = (vvdw12 - vvdw6 - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 210
 211             /* Update potential sums from outer loop */
 212             velecsum        += velec;
 213             vvdwsum         += vvdw;
 214
 215             fscal            = felec+fvdw;
 216
 217             /* Calculate temporary vectorial force */
 218             tx               = fscal*dx00;
 219             ty               = fscal*dy00;
 220             tz               = fscal*dz00;
 221
 222             /* Update vectorial force */
 223             fix0            += tx;
 224             fiy0            += ty;
 225             fiz0            += tz;
 226             f[j_coord_offset+DIM*0+XX] -= tx;
 227             f[j_coord_offset+DIM*0+YY] -= ty;
 228             f[j_coord_offset+DIM*0+ZZ] -= tz;
 229
 230             /* Inner loop uses 67 flops */
 231         }
 232         /* End of innermost loop */
 233
 234         tx = ty = tz = 0;
 235         f[i_coord_offset+DIM*0+XX] += fix0;
 236         f[i_coord_offset+DIM*0+YY] += fiy0;
 237         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 238         tx                         += fix0;
 239         ty                         += fiy0;
 240         tz                         += fiz0;
 241         fshift[i_shift_offset+XX]  += tx;
 242         fshift[i_shift_offset+YY]  += ty;
 243         fshift[i_shift_offset+ZZ]  += tz;
 244
 245         ggid                        = gid[iidx];
 246         /* Update potential energies */
 247         kernel_data->energygrp_elec[ggid] += velecsum;
 248         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 249
 250         /* Increment number of inner iterations */
 251         inneriter                  += j_index_end - j_index_start;
 252
 253         /* Outer loop uses 15 flops */
 254     }
 255
 256     /* Increment number of outer iterations */
 257     outeriter        += nri;
 258
 259     /* Update outer/inner flops */
 260
 261     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*67);
 262 }
 263 /*
 264  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomP1P1_F_c
 265  * Electrostatics interaction: Ewald
 266  * VdW interaction:            LJEwald
 267  * Geometry:                   Particle-Particle
 268  * Calculate force/pot:        Force
 269  */
 270 void
 271 nb_kernel_ElecEw_VdwLJEw_GeomP1P1_F_c
 272                     (t_nblist                    * gmx_restrict       nlist,
 273                      rvec                        * gmx_restrict          xx,
 274                      rvec                        * gmx_restrict          ff,
 275                      t_forcerec                  * gmx_restrict          fr,
 276                      t_mdatoms                   * gmx_restrict     mdatoms,
 277                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 278                      t_nrnb                      * gmx_restrict        nrnb)
 279 {
 280     int              i_shift_offset,i_coord_offset,j_coord_offset;
 281     int              j_index_start,j_index_end;
 282     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 283     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 284     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 285     real             *shiftvec,*fshift,*x,*f;
 286     int              vdwioffset0;
 287     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 288     int              vdwjidx0;
 289     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 290     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 291     real             velec,felec,velecsum,facel,crf,krf,krf2;
 292     real             *charge;
 293     int              nvdwtype;
 294     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 295     int              *vdwtype;
 296     real             *vdwparam;
 297     real             c6grid_00;
 298     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
 299     real             *vdwgridparam;
 300     int              ewitab;
 301     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 302     real             *ewtab;
 303
 304     x                = xx[0];
 305     f                = ff[0];
 306
 307     nri              = nlist->nri;
 308     iinr             = nlist->iinr;
 309     jindex           = nlist->jindex;
 310     jjnr             = nlist->jjnr;
 311     shiftidx         = nlist->shift;
 312     gid              = nlist->gid;
 313     shiftvec         = fr->shift_vec[0];
 314     fshift           = fr->fshift[0];
 315     facel            = fr->epsfac;
 316     charge           = mdatoms->chargeA;
 317     nvdwtype         = fr->ntype;
 318     vdwparam         = fr->nbfp;
 319     vdwtype          = mdatoms->typeA;
 320     vdwgridparam     = fr->ljpme_c6grid;
 321     ewclj            = fr->ewaldcoeff_lj;
 322     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 323     ewclj2           = ewclj*ewclj;
 324     ewclj6           = ewclj2*ewclj2*ewclj2;
 325
 326     sh_ewald         = fr->ic->sh_ewald;
 327     ewtab            = fr->ic->tabq_coul_F;
 328     ewtabscale       = fr->ic->tabq_scale;
 329     ewtabhalfspace   = 0.5/ewtabscale;
 330
 331     outeriter        = 0;
 332     inneriter        = 0;
 333
 334     /* Start outer loop over neighborlists */
 335     for(iidx=0; iidx<nri; iidx++)
 336     {
 337         /* Load shift vector for this list */
 338         i_shift_offset   = DIM*shiftidx[iidx];
 339         shX              = shiftvec[i_shift_offset+XX];
 340         shY              = shiftvec[i_shift_offset+YY];
 341         shZ              = shiftvec[i_shift_offset+ZZ];
 342
 343         /* Load limits for loop over neighbors */
 344         j_index_start    = jindex[iidx];
 345         j_index_end      = jindex[iidx+1];
 346
 347         /* Get outer coordinate index */
 348         inr              = iinr[iidx];
 349         i_coord_offset   = DIM*inr;
 350
 351         /* Load i particle coords and add shift vector */
 352         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 353         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 354         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 355
 356         fix0             = 0.0;
 357         fiy0             = 0.0;
 358         fiz0             = 0.0;
 359
 360         /* Load parameters for i particles */
 361         iq0              = facel*charge[inr+0];
 362         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 363
 364         /* Start inner kernel loop */
 365         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 366         {
 367             /* Get j neighbor index, and coordinate index */
 368             jnr              = jjnr[jidx];
 369             j_coord_offset   = DIM*jnr;
 370
 371             /* load j atom coordinates */
 372             jx0              = x[j_coord_offset+DIM*0+XX];
 373             jy0              = x[j_coord_offset+DIM*0+YY];
 374             jz0              = x[j_coord_offset+DIM*0+ZZ];
 375
 376             /* Calculate displacement vector */
 377             dx00             = ix0 - jx0;
 378             dy00             = iy0 - jy0;
 379             dz00             = iz0 - jz0;
 380
 381             /* Calculate squared distance and things based on it */
 382             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 383
 384             rinv00           = gmx_invsqrt(rsq00);
 385
 386             rinvsq00         = rinv00*rinv00;
 387
 388             /* Load parameters for j particles */
 389             jq0              = charge[jnr+0];
 390             vdwjidx0         = 2*vdwtype[jnr+0];
 391
 392             /**************************
 393              * CALCULATE INTERACTIONS *
 394              **************************/
 395
 396             r00              = rsq00*rinv00;
 397
 398             qq00             = iq0*jq0;
 399             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 400             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 401             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 402
 403             /* EWALD ELECTROSTATICS */
 404
 405             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 406             ewrt             = r00*ewtabscale;
 407             ewitab           = ewrt;
 408             eweps            = ewrt-ewitab;
 409             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 410             felec            = qq00*rinv00*(rinvsq00-felec);
 411
 412             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 413             ewcljrsq         = ewclj2*rsq00;
 414             exponent         = exp(-ewcljrsq);
 415             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 416             fvdw             = (((c12_00*rinvsix - c6_00 + c6grid_00*(1.0-poly))*rinvsix) - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 417
 418             fscal            = felec+fvdw;
 419
 420             /* Calculate temporary vectorial force */
 421             tx               = fscal*dx00;
 422             ty               = fscal*dy00;
 423             tz               = fscal*dz00;
 424
 425             /* Update vectorial force */
 426             fix0            += tx;
 427             fiy0            += ty;
 428             fiz0            += tz;
 429             f[j_coord_offset+DIM*0+XX] -= tx;
 430             f[j_coord_offset+DIM*0+YY] -= ty;
 431             f[j_coord_offset+DIM*0+ZZ] -= tz;
 432
 433             /* Inner loop uses 55 flops */
 434         }
 435         /* End of innermost loop */
 436
 437         tx = ty = tz = 0;
 438         f[i_coord_offset+DIM*0+XX] += fix0;
 439         f[i_coord_offset+DIM*0+YY] += fiy0;
 440         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 441         tx                         += fix0;
 442         ty                         += fiy0;
 443         tz                         += fiz0;
 444         fshift[i_shift_offset+XX]  += tx;
 445         fshift[i_shift_offset+YY]  += ty;
 446         fshift[i_shift_offset+ZZ]  += tz;
 447
 448         /* Increment number of inner iterations */
 449         inneriter                  += j_index_end - j_index_start;
 450
 451         /* Outer loop uses 13 flops */
 452     }
 453
 454     /* Increment number of outer iterations */
 455     outeriter        += nri;
 456
 457     /* Update outer/inner flops */
 458
 459     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*55);
 460 }