src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_VdwLJ_GeomP1P1_c.c

   1 /*
   2  * Note: this file was generated by the Gromacs c kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 /*
  34  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwLJ_GeomP1P1_VF_c
  35  * Electrostatics interaction: CubicSplineTable
  36  * VdW interaction:            LennardJones
  37  * Geometry:                   Particle-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecCSTab_VdwLJ_GeomP1P1_VF_c
  42                     (t_nblist * gmx_restrict                nlist,
  43                      rvec * gmx_restrict                    xx,
  44                      rvec * gmx_restrict                    ff,
  45                      t_forcerec * gmx_restrict              fr,
  46                      t_mdatoms * gmx_restrict               mdatoms,
  47                      nb_kernel_data_t * gmx_restrict        kernel_data,
  48                      t_nrnb * gmx_restrict                  nrnb)
  49 {
  50     int              i_shift_offset,i_coord_offset,j_coord_offset;
  51     int              j_index_start,j_index_end;
  52     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  53     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  54     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  55     real             *shiftvec,*fshift,*x,*f;
  56     int              vdwioffset0;
  57     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  58     int              vdwjidx0;
  59     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  60     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  61     real             velec,felec,velecsum,facel,crf,krf,krf2;
  62     real             *charge;
  63     int              nvdwtype;
  64     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  65     int              *vdwtype;
  66     real             *vdwparam;
  67     int              vfitab;
  68     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  69     real             *vftab;
  70
  71     x                = xx[0];
  72     f                = ff[0];
  73
  74     nri              = nlist->nri;
  75     iinr             = nlist->iinr;
  76     jindex           = nlist->jindex;
  77     jjnr             = nlist->jjnr;
  78     shiftidx         = nlist->shift;
  79     gid              = nlist->gid;
  80     shiftvec         = fr->shift_vec[0];
  81     fshift           = fr->fshift[0];
  82     facel            = fr->epsfac;
  83     charge           = mdatoms->chargeA;
  84     nvdwtype         = fr->ntype;
  85     vdwparam         = fr->nbfp;
  86     vdwtype          = mdatoms->typeA;
  87
  88     vftab            = kernel_data->table_elec->data;
  89     vftabscale       = kernel_data->table_elec->scale;
  90
  91     outeriter        = 0;
  92     inneriter        = 0;
  93
  94     /* Start outer loop over neighborlists */
  95     for(iidx=0; iidx<nri; iidx++)
  96     {
  97         /* Load shift vector for this list */
  98         i_shift_offset   = DIM*shiftidx[iidx];
  99         shX              = shiftvec[i_shift_offset+XX];
 100         shY              = shiftvec[i_shift_offset+YY];
 101         shZ              = shiftvec[i_shift_offset+ZZ];
 102
 103         /* Load limits for loop over neighbors */
 104         j_index_start    = jindex[iidx];
 105         j_index_end      = jindex[iidx+1];
 106
 107         /* Get outer coordinate index */
 108         inr              = iinr[iidx];
 109         i_coord_offset   = DIM*inr;
 110
 111         /* Load i particle coords and add shift vector */
 112         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 113         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 114         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 115
 116         fix0             = 0.0;
 117         fiy0             = 0.0;
 118         fiz0             = 0.0;
 119
 120         /* Load parameters for i particles */
 121         iq0              = facel*charge[inr+0];
 122         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 123
 124         /* Reset potential sums */
 125         velecsum         = 0.0;
 126         vvdwsum          = 0.0;
 127
 128         /* Start inner kernel loop */
 129         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 130         {
 131             /* Get j neighbor index, and coordinate index */
 132             jnr              = jjnr[jidx];
 133             j_coord_offset   = DIM*jnr;
 134
 135             /* load j atom coordinates */
 136             jx0              = x[j_coord_offset+DIM*0+XX];
 137             jy0              = x[j_coord_offset+DIM*0+YY];
 138             jz0              = x[j_coord_offset+DIM*0+ZZ];
 139
 140             /* Calculate displacement vector */
 141             dx00             = ix0 - jx0;
 142             dy00             = iy0 - jy0;
 143             dz00             = iz0 - jz0;
 144
 145             /* Calculate squared distance and things based on it */
 146             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 147
 148             rinv00           = gmx_invsqrt(rsq00);
 149
 150             rinvsq00         = rinv00*rinv00;
 151
 152             /* Load parameters for j particles */
 153             jq0              = charge[jnr+0];
 154             vdwjidx0         = 2*vdwtype[jnr+0];
 155
 156             /**************************
 157              * CALCULATE INTERACTIONS *
 158              **************************/
 159
 160             r00              = rsq00*rinv00;
 161
 162             qq00             = iq0*jq0;
 163             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 164             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 165
 166             /* Calculate table index by multiplying r with table scale and truncate to integer */
 167             rt               = r00*vftabscale;
 168             vfitab           = rt;
 169             vfeps            = rt-vfitab;
 170             vfitab           = 1*4*vfitab;
 171
 172             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 173             Y                = vftab[vfitab];
 174             F                = vftab[vfitab+1];
 175             Geps             = vfeps*vftab[vfitab+2];
 176             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 177             Fp               = F+Geps+Heps2;
 178             VV               = Y+vfeps*Fp;
 179             velec            = qq00*VV;
 180             FF               = Fp+Geps+2.0*Heps2;
 181             felec            = -qq00*FF*vftabscale*rinv00;
 182
 183             /* LENNARD-JONES DISPERSION/REPULSION */
 184
 185             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 186             vvdw6            = c6_00*rinvsix;
 187             vvdw12           = c12_00*rinvsix*rinvsix;
 188             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 189             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 190
 191             /* Update potential sums from outer loop */
 192             velecsum        += velec;
 193             vvdwsum         += vvdw;
 194
 195             fscal            = felec+fvdw;
 196
 197             /* Calculate temporary vectorial force */
 198             tx               = fscal*dx00;
 199             ty               = fscal*dy00;
 200             tz               = fscal*dz00;
 201
 202             /* Update vectorial force */
 203             fix0            += tx;
 204             fiy0            += ty;
 205             fiz0            += tz;
 206             f[j_coord_offset+DIM*0+XX] -= tx;
 207             f[j_coord_offset+DIM*0+YY] -= ty;
 208             f[j_coord_offset+DIM*0+ZZ] -= tz;
 209
 210             /* Inner loop uses 55 flops */
 211         }
 212         /* End of innermost loop */
 213
 214         tx = ty = tz = 0;
 215         f[i_coord_offset+DIM*0+XX] += fix0;
 216         f[i_coord_offset+DIM*0+YY] += fiy0;
 217         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 218         tx                         += fix0;
 219         ty                         += fiy0;
 220         tz                         += fiz0;
 221         fshift[i_shift_offset+XX]  += tx;
 222         fshift[i_shift_offset+YY]  += ty;
 223         fshift[i_shift_offset+ZZ]  += tz;
 224
 225         ggid                        = gid[iidx];
 226         /* Update potential energies */
 227         kernel_data->energygrp_elec[ggid] += velecsum;
 228         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 229
 230         /* Increment number of inner iterations */
 231         inneriter                  += j_index_end - j_index_start;
 232
 233         /* Outer loop uses 15 flops */
 234     }
 235
 236     /* Increment number of outer iterations */
 237     outeriter        += nri;
 238
 239     /* Update outer/inner flops */
 240
 241     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*55);
 242 }
 243 /*
 244  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwLJ_GeomP1P1_F_c
 245  * Electrostatics interaction: CubicSplineTable
 246  * VdW interaction:            LennardJones
 247  * Geometry:                   Particle-Particle
 248  * Calculate force/pot:        Force
 249  */
 250 void
 251 nb_kernel_ElecCSTab_VdwLJ_GeomP1P1_F_c
 252                     (t_nblist * gmx_restrict                nlist,
 253                      rvec * gmx_restrict                    xx,
 254                      rvec * gmx_restrict                    ff,
 255                      t_forcerec * gmx_restrict              fr,
 256                      t_mdatoms * gmx_restrict               mdatoms,
 257                      nb_kernel_data_t * gmx_restrict        kernel_data,
 258                      t_nrnb * gmx_restrict                  nrnb)
 259 {
 260     int              i_shift_offset,i_coord_offset,j_coord_offset;
 261     int              j_index_start,j_index_end;
 262     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 263     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 264     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 265     real             *shiftvec,*fshift,*x,*f;
 266     int              vdwioffset0;
 267     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 268     int              vdwjidx0;
 269     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 270     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 271     real             velec,felec,velecsum,facel,crf,krf,krf2;
 272     real             *charge;
 273     int              nvdwtype;
 274     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 275     int              *vdwtype;
 276     real             *vdwparam;
 277     int              vfitab;
 278     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 279     real             *vftab;
 280
 281     x                = xx[0];
 282     f                = ff[0];
 283
 284     nri              = nlist->nri;
 285     iinr             = nlist->iinr;
 286     jindex           = nlist->jindex;
 287     jjnr             = nlist->jjnr;
 288     shiftidx         = nlist->shift;
 289     gid              = nlist->gid;
 290     shiftvec         = fr->shift_vec[0];
 291     fshift           = fr->fshift[0];
 292     facel            = fr->epsfac;
 293     charge           = mdatoms->chargeA;
 294     nvdwtype         = fr->ntype;
 295     vdwparam         = fr->nbfp;
 296     vdwtype          = mdatoms->typeA;
 297
 298     vftab            = kernel_data->table_elec->data;
 299     vftabscale       = kernel_data->table_elec->scale;
 300
 301     outeriter        = 0;
 302     inneriter        = 0;
 303
 304     /* Start outer loop over neighborlists */
 305     for(iidx=0; iidx<nri; iidx++)
 306     {
 307         /* Load shift vector for this list */
 308         i_shift_offset   = DIM*shiftidx[iidx];
 309         shX              = shiftvec[i_shift_offset+XX];
 310         shY              = shiftvec[i_shift_offset+YY];
 311         shZ              = shiftvec[i_shift_offset+ZZ];
 312
 313         /* Load limits for loop over neighbors */
 314         j_index_start    = jindex[iidx];
 315         j_index_end      = jindex[iidx+1];
 316
 317         /* Get outer coordinate index */
 318         inr              = iinr[iidx];
 319         i_coord_offset   = DIM*inr;
 320
 321         /* Load i particle coords and add shift vector */
 322         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 323         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 324         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 325
 326         fix0             = 0.0;
 327         fiy0             = 0.0;
 328         fiz0             = 0.0;
 329
 330         /* Load parameters for i particles */
 331         iq0              = facel*charge[inr+0];
 332         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 333
 334         /* Start inner kernel loop */
 335         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 336         {
 337             /* Get j neighbor index, and coordinate index */
 338             jnr              = jjnr[jidx];
 339             j_coord_offset   = DIM*jnr;
 340
 341             /* load j atom coordinates */
 342             jx0              = x[j_coord_offset+DIM*0+XX];
 343             jy0              = x[j_coord_offset+DIM*0+YY];
 344             jz0              = x[j_coord_offset+DIM*0+ZZ];
 345
 346             /* Calculate displacement vector */
 347             dx00             = ix0 - jx0;
 348             dy00             = iy0 - jy0;
 349             dz00             = iz0 - jz0;
 350
 351             /* Calculate squared distance and things based on it */
 352             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 353
 354             rinv00           = gmx_invsqrt(rsq00);
 355
 356             rinvsq00         = rinv00*rinv00;
 357
 358             /* Load parameters for j particles */
 359             jq0              = charge[jnr+0];
 360             vdwjidx0         = 2*vdwtype[jnr+0];
 361
 362             /**************************
 363              * CALCULATE INTERACTIONS *
 364              **************************/
 365
 366             r00              = rsq00*rinv00;
 367
 368             qq00             = iq0*jq0;
 369             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 370             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 371
 372             /* Calculate table index by multiplying r with table scale and truncate to integer */
 373             rt               = r00*vftabscale;
 374             vfitab           = rt;
 375             vfeps            = rt-vfitab;
 376             vfitab           = 1*4*vfitab;
 377
 378             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 379             F                = vftab[vfitab+1];
 380             Geps             = vfeps*vftab[vfitab+2];
 381             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 382             Fp               = F+Geps+Heps2;
 383             FF               = Fp+Geps+2.0*Heps2;
 384             felec            = -qq00*FF*vftabscale*rinv00;
 385
 386             /* LENNARD-JONES DISPERSION/REPULSION */
 387
 388             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 389             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 390
 391             fscal            = felec+fvdw;
 392
 393             /* Calculate temporary vectorial force */
 394             tx               = fscal*dx00;
 395             ty               = fscal*dy00;
 396             tz               = fscal*dz00;
 397
 398             /* Update vectorial force */
 399             fix0            += tx;
 400             fiy0            += ty;
 401             fiz0            += tz;
 402             f[j_coord_offset+DIM*0+XX] -= tx;
 403             f[j_coord_offset+DIM*0+YY] -= ty;
 404             f[j_coord_offset+DIM*0+ZZ] -= tz;
 405
 406             /* Inner loop uses 46 flops */
 407         }
 408         /* End of innermost loop */
 409
 410         tx = ty = tz = 0;
 411         f[i_coord_offset+DIM*0+XX] += fix0;
 412         f[i_coord_offset+DIM*0+YY] += fiy0;
 413         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 414         tx                         += fix0;
 415         ty                         += fiy0;
 416         tz                         += fiz0;
 417         fshift[i_shift_offset+XX]  += tx;
 418         fshift[i_shift_offset+YY]  += ty;
 419         fshift[i_shift_offset+ZZ]  += tz;
 420
 421         /* Increment number of inner iterations */
 422         inneriter                  += j_index_end - j_index_start;
 423
 424         /* Outer loop uses 13 flops */
 425     }
 426
 427     /* Increment number of outer iterations */
 428     outeriter        += nri;
 429
 430     /* Update outer/inner flops */
 431
 432     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*46);
 433 }