src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRF_VdwCSTab_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_ElecRF_VdwCSTab_GeomP1P1_VF_c
  35  * Electrostatics interaction: ReactionField
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Particle-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecRF_VdwCSTab_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     krf              = fr->ic->k_rf;
  85     krf2             = krf*2.0;
  86     crf              = fr->ic->c_rf;
  87     nvdwtype         = fr->ntype;
  88     vdwparam         = fr->nbfp;
  89     vdwtype          = mdatoms->typeA;
  90
  91     vftab            = kernel_data->table_vdw->data;
  92     vftabscale       = kernel_data->table_vdw->scale;
  93
  94     outeriter        = 0;
  95     inneriter        = 0;
  96
  97     /* Start outer loop over neighborlists */
  98     for(iidx=0; iidx<nri; iidx++)
  99     {
 100         /* Load shift vector for this list */
 101         i_shift_offset   = DIM*shiftidx[iidx];
 102         shX              = shiftvec[i_shift_offset+XX];
 103         shY              = shiftvec[i_shift_offset+YY];
 104         shZ              = shiftvec[i_shift_offset+ZZ];
 105
 106         /* Load limits for loop over neighbors */
 107         j_index_start    = jindex[iidx];
 108         j_index_end      = jindex[iidx+1];
 109
 110         /* Get outer coordinate index */
 111         inr              = iinr[iidx];
 112         i_coord_offset   = DIM*inr;
 113
 114         /* Load i particle coords and add shift vector */
 115         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 116         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 117         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 118
 119         fix0             = 0.0;
 120         fiy0             = 0.0;
 121         fiz0             = 0.0;
 122
 123         /* Load parameters for i particles */
 124         iq0              = facel*charge[inr+0];
 125         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 126
 127         /* Reset potential sums */
 128         velecsum         = 0.0;
 129         vvdwsum          = 0.0;
 130
 131         /* Start inner kernel loop */
 132         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 133         {
 134             /* Get j neighbor index, and coordinate index */
 135             jnr              = jjnr[jidx];
 136             j_coord_offset   = DIM*jnr;
 137
 138             /* load j atom coordinates */
 139             jx0              = x[j_coord_offset+DIM*0+XX];
 140             jy0              = x[j_coord_offset+DIM*0+YY];
 141             jz0              = x[j_coord_offset+DIM*0+ZZ];
 142
 143             /* Calculate displacement vector */
 144             dx00             = ix0 - jx0;
 145             dy00             = iy0 - jy0;
 146             dz00             = iz0 - jz0;
 147
 148             /* Calculate squared distance and things based on it */
 149             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 150
 151             rinv00           = gmx_invsqrt(rsq00);
 152
 153             rinvsq00         = rinv00*rinv00;
 154
 155             /* Load parameters for j particles */
 156             jq0              = charge[jnr+0];
 157             vdwjidx0         = 2*vdwtype[jnr+0];
 158
 159             /**************************
 160              * CALCULATE INTERACTIONS *
 161              **************************/
 162
 163             r00              = rsq00*rinv00;
 164
 165             qq00             = iq0*jq0;
 166             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 167             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 168
 169             /* Calculate table index by multiplying r with table scale and truncate to integer */
 170             rt               = r00*vftabscale;
 171             vfitab           = rt;
 172             vfeps            = rt-vfitab;
 173             vfitab           = 2*4*vfitab;
 174
 175             /* REACTION-FIELD ELECTROSTATICS */
 176             velec            = qq00*(rinv00+krf*rsq00-crf);
 177             felec            = qq00*(rinv00*rinvsq00-krf2);
 178
 179             /* CUBIC SPLINE TABLE DISPERSION */
 180             vfitab          += 0;
 181             Y                = vftab[vfitab];
 182             F                = vftab[vfitab+1];
 183             Geps             = vfeps*vftab[vfitab+2];
 184             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 185             Fp               = F+Geps+Heps2;
 186             VV               = Y+vfeps*Fp;
 187             vvdw6            = c6_00*VV;
 188             FF               = Fp+Geps+2.0*Heps2;
 189             fvdw6            = c6_00*FF;
 190
 191             /* CUBIC SPLINE TABLE REPULSION */
 192             Y                = vftab[vfitab+4];
 193             F                = vftab[vfitab+5];
 194             Geps             = vfeps*vftab[vfitab+6];
 195             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 196             Fp               = F+Geps+Heps2;
 197             VV               = Y+vfeps*Fp;
 198             vvdw12           = c12_00*VV;
 199             FF               = Fp+Geps+2.0*Heps2;
 200             fvdw12           = c12_00*FF;
 201             vvdw             = vvdw12+vvdw6;
 202             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 203
 204             /* Update potential sums from outer loop */
 205             velecsum        += velec;
 206             vvdwsum         += vvdw;
 207
 208             fscal            = felec+fvdw;
 209
 210             /* Calculate temporary vectorial force */
 211             tx               = fscal*dx00;
 212             ty               = fscal*dy00;
 213             tz               = fscal*dz00;
 214
 215             /* Update vectorial force */
 216             fix0            += tx;
 217             fiy0            += ty;
 218             fiz0            += tz;
 219             f[j_coord_offset+DIM*0+XX] -= tx;
 220             f[j_coord_offset+DIM*0+YY] -= ty;
 221             f[j_coord_offset+DIM*0+ZZ] -= tz;
 222
 223             /* Inner loop uses 66 flops */
 224         }
 225         /* End of innermost loop */
 226
 227         tx = ty = tz = 0;
 228         f[i_coord_offset+DIM*0+XX] += fix0;
 229         f[i_coord_offset+DIM*0+YY] += fiy0;
 230         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 231         tx                         += fix0;
 232         ty                         += fiy0;
 233         tz                         += fiz0;
 234         fshift[i_shift_offset+XX]  += tx;
 235         fshift[i_shift_offset+YY]  += ty;
 236         fshift[i_shift_offset+ZZ]  += tz;
 237
 238         ggid                        = gid[iidx];
 239         /* Update potential energies */
 240         kernel_data->energygrp_elec[ggid] += velecsum;
 241         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 242
 243         /* Increment number of inner iterations */
 244         inneriter                  += j_index_end - j_index_start;
 245
 246         /* Outer loop uses 15 flops */
 247     }
 248
 249     /* Increment number of outer iterations */
 250     outeriter        += nri;
 251
 252     /* Update outer/inner flops */
 253
 254     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*66);
 255 }
 256 /*
 257  * Gromacs nonbonded kernel:   nb_kernel_ElecRF_VdwCSTab_GeomP1P1_F_c
 258  * Electrostatics interaction: ReactionField
 259  * VdW interaction:            CubicSplineTable
 260  * Geometry:                   Particle-Particle
 261  * Calculate force/pot:        Force
 262  */
 263 void
 264 nb_kernel_ElecRF_VdwCSTab_GeomP1P1_F_c
 265                     (t_nblist * gmx_restrict                nlist,
 266                      rvec * gmx_restrict                    xx,
 267                      rvec * gmx_restrict                    ff,
 268                      t_forcerec * gmx_restrict              fr,
 269                      t_mdatoms * gmx_restrict               mdatoms,
 270                      nb_kernel_data_t * gmx_restrict        kernel_data,
 271                      t_nrnb * gmx_restrict                  nrnb)
 272 {
 273     int              i_shift_offset,i_coord_offset,j_coord_offset;
 274     int              j_index_start,j_index_end;
 275     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 276     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 277     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 278     real             *shiftvec,*fshift,*x,*f;
 279     int              vdwioffset0;
 280     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 281     int              vdwjidx0;
 282     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 283     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 284     real             velec,felec,velecsum,facel,crf,krf,krf2;
 285     real             *charge;
 286     int              nvdwtype;
 287     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 288     int              *vdwtype;
 289     real             *vdwparam;
 290     int              vfitab;
 291     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 292     real             *vftab;
 293
 294     x                = xx[0];
 295     f                = ff[0];
 296
 297     nri              = nlist->nri;
 298     iinr             = nlist->iinr;
 299     jindex           = nlist->jindex;
 300     jjnr             = nlist->jjnr;
 301     shiftidx         = nlist->shift;
 302     gid              = nlist->gid;
 303     shiftvec         = fr->shift_vec[0];
 304     fshift           = fr->fshift[0];
 305     facel            = fr->epsfac;
 306     charge           = mdatoms->chargeA;
 307     krf              = fr->ic->k_rf;
 308     krf2             = krf*2.0;
 309     crf              = fr->ic->c_rf;
 310     nvdwtype         = fr->ntype;
 311     vdwparam         = fr->nbfp;
 312     vdwtype          = mdatoms->typeA;
 313
 314     vftab            = kernel_data->table_vdw->data;
 315     vftabscale       = kernel_data->table_vdw->scale;
 316
 317     outeriter        = 0;
 318     inneriter        = 0;
 319
 320     /* Start outer loop over neighborlists */
 321     for(iidx=0; iidx<nri; iidx++)
 322     {
 323         /* Load shift vector for this list */
 324         i_shift_offset   = DIM*shiftidx[iidx];
 325         shX              = shiftvec[i_shift_offset+XX];
 326         shY              = shiftvec[i_shift_offset+YY];
 327         shZ              = shiftvec[i_shift_offset+ZZ];
 328
 329         /* Load limits for loop over neighbors */
 330         j_index_start    = jindex[iidx];
 331         j_index_end      = jindex[iidx+1];
 332
 333         /* Get outer coordinate index */
 334         inr              = iinr[iidx];
 335         i_coord_offset   = DIM*inr;
 336
 337         /* Load i particle coords and add shift vector */
 338         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 339         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 340         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 341
 342         fix0             = 0.0;
 343         fiy0             = 0.0;
 344         fiz0             = 0.0;
 345
 346         /* Load parameters for i particles */
 347         iq0              = facel*charge[inr+0];
 348         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 349
 350         /* Start inner kernel loop */
 351         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 352         {
 353             /* Get j neighbor index, and coordinate index */
 354             jnr              = jjnr[jidx];
 355             j_coord_offset   = DIM*jnr;
 356
 357             /* load j atom coordinates */
 358             jx0              = x[j_coord_offset+DIM*0+XX];
 359             jy0              = x[j_coord_offset+DIM*0+YY];
 360             jz0              = x[j_coord_offset+DIM*0+ZZ];
 361
 362             /* Calculate displacement vector */
 363             dx00             = ix0 - jx0;
 364             dy00             = iy0 - jy0;
 365             dz00             = iz0 - jz0;
 366
 367             /* Calculate squared distance and things based on it */
 368             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 369
 370             rinv00           = gmx_invsqrt(rsq00);
 371
 372             rinvsq00         = rinv00*rinv00;
 373
 374             /* Load parameters for j particles */
 375             jq0              = charge[jnr+0];
 376             vdwjidx0         = 2*vdwtype[jnr+0];
 377
 378             /**************************
 379              * CALCULATE INTERACTIONS *
 380              **************************/
 381
 382             r00              = rsq00*rinv00;
 383
 384             qq00             = iq0*jq0;
 385             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 386             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 387
 388             /* Calculate table index by multiplying r with table scale and truncate to integer */
 389             rt               = r00*vftabscale;
 390             vfitab           = rt;
 391             vfeps            = rt-vfitab;
 392             vfitab           = 2*4*vfitab;
 393
 394             /* REACTION-FIELD ELECTROSTATICS */
 395             felec            = qq00*(rinv00*rinvsq00-krf2);
 396
 397             /* CUBIC SPLINE TABLE DISPERSION */
 398             vfitab          += 0;
 399             F                = vftab[vfitab+1];
 400             Geps             = vfeps*vftab[vfitab+2];
 401             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 402             Fp               = F+Geps+Heps2;
 403             FF               = Fp+Geps+2.0*Heps2;
 404             fvdw6            = c6_00*FF;
 405
 406             /* CUBIC SPLINE TABLE REPULSION */
 407             F                = vftab[vfitab+5];
 408             Geps             = vfeps*vftab[vfitab+6];
 409             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 410             Fp               = F+Geps+Heps2;
 411             FF               = Fp+Geps+2.0*Heps2;
 412             fvdw12           = c12_00*FF;
 413             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 414
 415             fscal            = felec+fvdw;
 416
 417             /* Calculate temporary vectorial force */
 418             tx               = fscal*dx00;
 419             ty               = fscal*dy00;
 420             tz               = fscal*dz00;
 421
 422             /* Update vectorial force */
 423             fix0            += tx;
 424             fiy0            += ty;
 425             fiz0            += tz;
 426             f[j_coord_offset+DIM*0+XX] -= tx;
 427             f[j_coord_offset+DIM*0+YY] -= ty;
 428             f[j_coord_offset+DIM*0+ZZ] -= tz;
 429
 430             /* Inner loop uses 53 flops */
 431         }
 432         /* End of innermost loop */
 433
 434         tx = ty = tz = 0;
 435         f[i_coord_offset+DIM*0+XX] += fix0;
 436         f[i_coord_offset+DIM*0+YY] += fiy0;
 437         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 438         tx                         += fix0;
 439         ty                         += fiy0;
 440         tz                         += fiz0;
 441         fshift[i_shift_offset+XX]  += tx;
 442         fshift[i_shift_offset+YY]  += ty;
 443         fshift[i_shift_offset+ZZ]  += tz;
 444
 445         /* Increment number of inner iterations */
 446         inneriter                  += j_index_end - j_index_start;
 447
 448         /* Outer loop uses 13 flops */
 449     }
 450
 451     /* Increment number of outer iterations */
 452     outeriter        += nri;
 453
 454     /* Update outer/inner flops */
 455
 456     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*53);
 457 }