src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_VdwBham_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_VdwBham_GeomP1P1_VF_c
  35  * Electrostatics interaction: CubicSplineTable
  36  * VdW interaction:            Buckingham
  37  * Geometry:                   Particle-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecCSTab_VdwBham_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      = 3*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         = 3*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             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 165             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 166
 167             /* Calculate table index by multiplying r with table scale and truncate to integer */
 168             rt               = r00*vftabscale;
 169             vfitab           = rt;
 170             vfeps            = rt-vfitab;
 171             vfitab           = 1*4*vfitab;
 172
 173             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 174             Y                = vftab[vfitab];
 175             F                = vftab[vfitab+1];
 176             Geps             = vfeps*vftab[vfitab+2];
 177             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 178             Fp               = F+Geps+Heps2;
 179             VV               = Y+vfeps*Fp;
 180             velec            = qq00*VV;
 181             FF               = Fp+Geps+2.0*Heps2;
 182             felec            = -qq00*FF*vftabscale*rinv00;
 183
 184             /* BUCKINGHAM DISPERSION/REPULSION */
 185             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 186             vvdw6            = c6_00*rinvsix;
 187             br               = cexp2_00*r00;
 188             vvdwexp          = cexp1_00*exp(-br);
 189             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 190             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 191
 192             /* Update potential sums from outer loop */
 193             velecsum        += velec;
 194             vvdwsum         += vvdw;
 195
 196             fscal            = felec+fvdw;
 197
 198             /* Calculate temporary vectorial force */
 199             tx               = fscal*dx00;
 200             ty               = fscal*dy00;
 201             tz               = fscal*dz00;
 202
 203             /* Update vectorial force */
 204             fix0            += tx;
 205             fiy0            += ty;
 206             fiz0            += tz;
 207             f[j_coord_offset+DIM*0+XX] -= tx;
 208             f[j_coord_offset+DIM*0+YY] -= ty;
 209             f[j_coord_offset+DIM*0+ZZ] -= tz;
 210
 211             /* Inner loop uses 81 flops */
 212         }
 213         /* End of innermost loop */
 214
 215         tx = ty = tz = 0;
 216         f[i_coord_offset+DIM*0+XX] += fix0;
 217         f[i_coord_offset+DIM*0+YY] += fiy0;
 218         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 219         tx                         += fix0;
 220         ty                         += fiy0;
 221         tz                         += fiz0;
 222         fshift[i_shift_offset+XX]  += tx;
 223         fshift[i_shift_offset+YY]  += ty;
 224         fshift[i_shift_offset+ZZ]  += tz;
 225
 226         ggid                        = gid[iidx];
 227         /* Update potential energies */
 228         kernel_data->energygrp_elec[ggid] += velecsum;
 229         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 230
 231         /* Increment number of inner iterations */
 232         inneriter                  += j_index_end - j_index_start;
 233
 234         /* Outer loop uses 15 flops */
 235     }
 236
 237     /* Increment number of outer iterations */
 238     outeriter        += nri;
 239
 240     /* Update outer/inner flops */
 241
 242     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*81);
 243 }
 244 /*
 245  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwBham_GeomP1P1_F_c
 246  * Electrostatics interaction: CubicSplineTable
 247  * VdW interaction:            Buckingham
 248  * Geometry:                   Particle-Particle
 249  * Calculate force/pot:        Force
 250  */
 251 void
 252 nb_kernel_ElecCSTab_VdwBham_GeomP1P1_F_c
 253                     (t_nblist * gmx_restrict                nlist,
 254                      rvec * gmx_restrict                    xx,
 255                      rvec * gmx_restrict                    ff,
 256                      t_forcerec * gmx_restrict              fr,
 257                      t_mdatoms * gmx_restrict               mdatoms,
 258                      nb_kernel_data_t * gmx_restrict        kernel_data,
 259                      t_nrnb * gmx_restrict                  nrnb)
 260 {
 261     int              i_shift_offset,i_coord_offset,j_coord_offset;
 262     int              j_index_start,j_index_end;
 263     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 264     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 265     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 266     real             *shiftvec,*fshift,*x,*f;
 267     int              vdwioffset0;
 268     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 269     int              vdwjidx0;
 270     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 271     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 272     real             velec,felec,velecsum,facel,crf,krf,krf2;
 273     real             *charge;
 274     int              nvdwtype;
 275     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 276     int              *vdwtype;
 277     real             *vdwparam;
 278     int              vfitab;
 279     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 280     real             *vftab;
 281
 282     x                = xx[0];
 283     f                = ff[0];
 284
 285     nri              = nlist->nri;
 286     iinr             = nlist->iinr;
 287     jindex           = nlist->jindex;
 288     jjnr             = nlist->jjnr;
 289     shiftidx         = nlist->shift;
 290     gid              = nlist->gid;
 291     shiftvec         = fr->shift_vec[0];
 292     fshift           = fr->fshift[0];
 293     facel            = fr->epsfac;
 294     charge           = mdatoms->chargeA;
 295     nvdwtype         = fr->ntype;
 296     vdwparam         = fr->nbfp;
 297     vdwtype          = mdatoms->typeA;
 298
 299     vftab            = kernel_data->table_elec->data;
 300     vftabscale       = kernel_data->table_elec->scale;
 301
 302     outeriter        = 0;
 303     inneriter        = 0;
 304
 305     /* Start outer loop over neighborlists */
 306     for(iidx=0; iidx<nri; iidx++)
 307     {
 308         /* Load shift vector for this list */
 309         i_shift_offset   = DIM*shiftidx[iidx];
 310         shX              = shiftvec[i_shift_offset+XX];
 311         shY              = shiftvec[i_shift_offset+YY];
 312         shZ              = shiftvec[i_shift_offset+ZZ];
 313
 314         /* Load limits for loop over neighbors */
 315         j_index_start    = jindex[iidx];
 316         j_index_end      = jindex[iidx+1];
 317
 318         /* Get outer coordinate index */
 319         inr              = iinr[iidx];
 320         i_coord_offset   = DIM*inr;
 321
 322         /* Load i particle coords and add shift vector */
 323         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 324         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 325         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 326
 327         fix0             = 0.0;
 328         fiy0             = 0.0;
 329         fiz0             = 0.0;
 330
 331         /* Load parameters for i particles */
 332         iq0              = facel*charge[inr+0];
 333         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 334
 335         /* Start inner kernel loop */
 336         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 337         {
 338             /* Get j neighbor index, and coordinate index */
 339             jnr              = jjnr[jidx];
 340             j_coord_offset   = DIM*jnr;
 341
 342             /* load j atom coordinates */
 343             jx0              = x[j_coord_offset+DIM*0+XX];
 344             jy0              = x[j_coord_offset+DIM*0+YY];
 345             jz0              = x[j_coord_offset+DIM*0+ZZ];
 346
 347             /* Calculate displacement vector */
 348             dx00             = ix0 - jx0;
 349             dy00             = iy0 - jy0;
 350             dz00             = iz0 - jz0;
 351
 352             /* Calculate squared distance and things based on it */
 353             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 354
 355             rinv00           = gmx_invsqrt(rsq00);
 356
 357             rinvsq00         = rinv00*rinv00;
 358
 359             /* Load parameters for j particles */
 360             jq0              = charge[jnr+0];
 361             vdwjidx0         = 3*vdwtype[jnr+0];
 362
 363             /**************************
 364              * CALCULATE INTERACTIONS *
 365              **************************/
 366
 367             r00              = rsq00*rinv00;
 368
 369             qq00             = iq0*jq0;
 370             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 371             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 372             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 373
 374             /* Calculate table index by multiplying r with table scale and truncate to integer */
 375             rt               = r00*vftabscale;
 376             vfitab           = rt;
 377             vfeps            = rt-vfitab;
 378             vfitab           = 1*4*vfitab;
 379
 380             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 381             Y                = vftab[vfitab];
 382             F                = vftab[vfitab+1];
 383             Geps             = vfeps*vftab[vfitab+2];
 384             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 385             Fp               = F+Geps+Heps2;
 386             FF               = Fp+Geps+2.0*Heps2;
 387             felec            = -qq00*FF*vftabscale*rinv00;
 388
 389             /* BUCKINGHAM DISPERSION/REPULSION */
 390             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 391             vvdw6            = c6_00*rinvsix;
 392             br               = cexp2_00*r00;
 393             vvdwexp          = cexp1_00*exp(-br);
 394             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 395
 396             fscal            = felec+fvdw;
 397
 398             /* Calculate temporary vectorial force */
 399             tx               = fscal*dx00;
 400             ty               = fscal*dy00;
 401             tz               = fscal*dz00;
 402
 403             /* Update vectorial force */
 404             fix0            += tx;
 405             fiy0            += ty;
 406             fiz0            += tz;
 407             f[j_coord_offset+DIM*0+XX] -= tx;
 408             f[j_coord_offset+DIM*0+YY] -= ty;
 409             f[j_coord_offset+DIM*0+ZZ] -= tz;
 410
 411             /* Inner loop uses 74 flops */
 412         }
 413         /* End of innermost loop */
 414
 415         tx = ty = tz = 0;
 416         f[i_coord_offset+DIM*0+XX] += fix0;
 417         f[i_coord_offset+DIM*0+YY] += fiy0;
 418         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 419         tx                         += fix0;
 420         ty                         += fiy0;
 421         tz                         += fiz0;
 422         fshift[i_shift_offset+XX]  += tx;
 423         fshift[i_shift_offset+YY]  += ty;
 424         fshift[i_shift_offset+ZZ]  += tz;
 425
 426         /* Increment number of inner iterations */
 427         inneriter                  += j_index_end - j_index_start;
 428
 429         /* Outer loop uses 13 flops */
 430     }
 431
 432     /* Increment number of outer iterations */
 433     outeriter        += nri;
 434
 435     /* Update outer/inner flops */
 436
 437     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*74);
 438 }