src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_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_ElecCSTab_VdwCSTab_GeomP1P1_VF_c
  35  * Electrostatics interaction: CubicSplineTable
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Particle-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecCSTab_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     nvdwtype         = fr->ntype;
  85     vdwparam         = fr->nbfp;
  86     vdwtype          = mdatoms->typeA;
  87
  88     vftab            = kernel_data->table_elec_vdw->data;
  89     vftabscale       = kernel_data->table_elec_vdw->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             /* Load parameters for j particles */
 151             jq0              = charge[jnr+0];
 152             vdwjidx0         = 2*vdwtype[jnr+0];
 153
 154             /**************************
 155              * CALCULATE INTERACTIONS *
 156              **************************/
 157
 158             r00              = rsq00*rinv00;
 159
 160             qq00             = iq0*jq0;
 161             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 162             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 163
 164             /* Calculate table index by multiplying r with table scale and truncate to integer */
 165             rt               = r00*vftabscale;
 166             vfitab           = rt;
 167             vfeps            = rt-vfitab;
 168             vfitab           = 3*4*vfitab;
 169
 170             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 171             Y                = vftab[vfitab];
 172             F                = vftab[vfitab+1];
 173             Geps             = vfeps*vftab[vfitab+2];
 174             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 175             Fp               = F+Geps+Heps2;
 176             VV               = Y+vfeps*Fp;
 177             velec            = qq00*VV;
 178             FF               = Fp+Geps+2.0*Heps2;
 179             felec            = -qq00*FF*vftabscale*rinv00;
 180
 181             /* CUBIC SPLINE TABLE DISPERSION */
 182             vfitab          += 4;
 183             Y                = vftab[vfitab];
 184             F                = vftab[vfitab+1];
 185             Geps             = vfeps*vftab[vfitab+2];
 186             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 187             Fp               = F+Geps+Heps2;
 188             VV               = Y+vfeps*Fp;
 189             vvdw6            = c6_00*VV;
 190             FF               = Fp+Geps+2.0*Heps2;
 191             fvdw6            = c6_00*FF;
 192
 193             /* CUBIC SPLINE TABLE REPULSION */
 194             Y                = vftab[vfitab+4];
 195             F                = vftab[vfitab+5];
 196             Geps             = vfeps*vftab[vfitab+6];
 197             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 198             Fp               = F+Geps+Heps2;
 199             VV               = Y+vfeps*Fp;
 200             vvdw12           = c12_00*VV;
 201             FF               = Fp+Geps+2.0*Heps2;
 202             fvdw12           = c12_00*FF;
 203             vvdw             = vvdw12+vvdw6;
 204             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 205
 206             /* Update potential sums from outer loop */
 207             velecsum        += velec;
 208             vvdwsum         += vvdw;
 209
 210             fscal            = felec+fvdw;
 211
 212             /* Calculate temporary vectorial force */
 213             tx               = fscal*dx00;
 214             ty               = fscal*dy00;
 215             tz               = fscal*dz00;
 216
 217             /* Update vectorial force */
 218             fix0            += tx;
 219             fiy0            += ty;
 220             fiz0            += tz;
 221             f[j_coord_offset+DIM*0+XX] -= tx;
 222             f[j_coord_offset+DIM*0+YY] -= ty;
 223             f[j_coord_offset+DIM*0+ZZ] -= tz;
 224
 225             /* Inner loop uses 73 flops */
 226         }
 227         /* End of innermost loop */
 228
 229         tx = ty = tz = 0;
 230         f[i_coord_offset+DIM*0+XX] += fix0;
 231         f[i_coord_offset+DIM*0+YY] += fiy0;
 232         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 233         tx                         += fix0;
 234         ty                         += fiy0;
 235         tz                         += fiz0;
 236         fshift[i_shift_offset+XX]  += tx;
 237         fshift[i_shift_offset+YY]  += ty;
 238         fshift[i_shift_offset+ZZ]  += tz;
 239
 240         ggid                        = gid[iidx];
 241         /* Update potential energies */
 242         kernel_data->energygrp_elec[ggid] += velecsum;
 243         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 244
 245         /* Increment number of inner iterations */
 246         inneriter                  += j_index_end - j_index_start;
 247
 248         /* Outer loop uses 15 flops */
 249     }
 250
 251     /* Increment number of outer iterations */
 252     outeriter        += nri;
 253
 254     /* Update outer/inner flops */
 255
 256     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*73);
 257 }
 258 /*
 259  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwCSTab_GeomP1P1_F_c
 260  * Electrostatics interaction: CubicSplineTable
 261  * VdW interaction:            CubicSplineTable
 262  * Geometry:                   Particle-Particle
 263  * Calculate force/pot:        Force
 264  */
 265 void
 266 nb_kernel_ElecCSTab_VdwCSTab_GeomP1P1_F_c
 267                     (t_nblist * gmx_restrict                nlist,
 268                      rvec * gmx_restrict                    xx,
 269                      rvec * gmx_restrict                    ff,
 270                      t_forcerec * gmx_restrict              fr,
 271                      t_mdatoms * gmx_restrict               mdatoms,
 272                      nb_kernel_data_t * gmx_restrict        kernel_data,
 273                      t_nrnb * gmx_restrict                  nrnb)
 274 {
 275     int              i_shift_offset,i_coord_offset,j_coord_offset;
 276     int              j_index_start,j_index_end;
 277     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 278     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 279     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 280     real             *shiftvec,*fshift,*x,*f;
 281     int              vdwioffset0;
 282     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 283     int              vdwjidx0;
 284     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 285     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 286     real             velec,felec,velecsum,facel,crf,krf,krf2;
 287     real             *charge;
 288     int              nvdwtype;
 289     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 290     int              *vdwtype;
 291     real             *vdwparam;
 292     int              vfitab;
 293     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 294     real             *vftab;
 295
 296     x                = xx[0];
 297     f                = ff[0];
 298
 299     nri              = nlist->nri;
 300     iinr             = nlist->iinr;
 301     jindex           = nlist->jindex;
 302     jjnr             = nlist->jjnr;
 303     shiftidx         = nlist->shift;
 304     gid              = nlist->gid;
 305     shiftvec         = fr->shift_vec[0];
 306     fshift           = fr->fshift[0];
 307     facel            = fr->epsfac;
 308     charge           = mdatoms->chargeA;
 309     nvdwtype         = fr->ntype;
 310     vdwparam         = fr->nbfp;
 311     vdwtype          = mdatoms->typeA;
 312
 313     vftab            = kernel_data->table_elec_vdw->data;
 314     vftabscale       = kernel_data->table_elec_vdw->scale;
 315
 316     outeriter        = 0;
 317     inneriter        = 0;
 318
 319     /* Start outer loop over neighborlists */
 320     for(iidx=0; iidx<nri; iidx++)
 321     {
 322         /* Load shift vector for this list */
 323         i_shift_offset   = DIM*shiftidx[iidx];
 324         shX              = shiftvec[i_shift_offset+XX];
 325         shY              = shiftvec[i_shift_offset+YY];
 326         shZ              = shiftvec[i_shift_offset+ZZ];
 327
 328         /* Load limits for loop over neighbors */
 329         j_index_start    = jindex[iidx];
 330         j_index_end      = jindex[iidx+1];
 331
 332         /* Get outer coordinate index */
 333         inr              = iinr[iidx];
 334         i_coord_offset   = DIM*inr;
 335
 336         /* Load i particle coords and add shift vector */
 337         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 338         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 339         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 340
 341         fix0             = 0.0;
 342         fiy0             = 0.0;
 343         fiz0             = 0.0;
 344
 345         /* Load parameters for i particles */
 346         iq0              = facel*charge[inr+0];
 347         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 348
 349         /* Start inner kernel loop */
 350         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 351         {
 352             /* Get j neighbor index, and coordinate index */
 353             jnr              = jjnr[jidx];
 354             j_coord_offset   = DIM*jnr;
 355
 356             /* load j atom coordinates */
 357             jx0              = x[j_coord_offset+DIM*0+XX];
 358             jy0              = x[j_coord_offset+DIM*0+YY];
 359             jz0              = x[j_coord_offset+DIM*0+ZZ];
 360
 361             /* Calculate displacement vector */
 362             dx00             = ix0 - jx0;
 363             dy00             = iy0 - jy0;
 364             dz00             = iz0 - jz0;
 365
 366             /* Calculate squared distance and things based on it */
 367             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 368
 369             rinv00           = gmx_invsqrt(rsq00);
 370
 371             /* Load parameters for j particles */
 372             jq0              = charge[jnr+0];
 373             vdwjidx0         = 2*vdwtype[jnr+0];
 374
 375             /**************************
 376              * CALCULATE INTERACTIONS *
 377              **************************/
 378
 379             r00              = rsq00*rinv00;
 380
 381             qq00             = iq0*jq0;
 382             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 383             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 384
 385             /* Calculate table index by multiplying r with table scale and truncate to integer */
 386             rt               = r00*vftabscale;
 387             vfitab           = rt;
 388             vfeps            = rt-vfitab;
 389             vfitab           = 3*4*vfitab;
 390
 391             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 392             Y                = vftab[vfitab];
 393             F                = vftab[vfitab+1];
 394             Geps             = vfeps*vftab[vfitab+2];
 395             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 396             Fp               = F+Geps+Heps2;
 397             FF               = Fp+Geps+2.0*Heps2;
 398             felec            = -qq00*FF*vftabscale*rinv00;
 399
 400             /* CUBIC SPLINE TABLE DISPERSION */
 401             vfitab          += 4;
 402             Y                = vftab[vfitab];
 403             F                = vftab[vfitab+1];
 404             Geps             = vfeps*vftab[vfitab+2];
 405             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 406             Fp               = F+Geps+Heps2;
 407             FF               = Fp+Geps+2.0*Heps2;
 408             fvdw6            = c6_00*FF;
 409
 410             /* CUBIC SPLINE TABLE REPULSION */
 411             Y                = vftab[vfitab+4];
 412             F                = vftab[vfitab+5];
 413             Geps             = vfeps*vftab[vfitab+6];
 414             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 415             Fp               = F+Geps+Heps2;
 416             FF               = Fp+Geps+2.0*Heps2;
 417             fvdw12           = c12_00*FF;
 418             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 419
 420             fscal            = felec+fvdw;
 421
 422             /* Calculate temporary vectorial force */
 423             tx               = fscal*dx00;
 424             ty               = fscal*dy00;
 425             tz               = fscal*dz00;
 426
 427             /* Update vectorial force */
 428             fix0            += tx;
 429             fiy0            += ty;
 430             fiz0            += tz;
 431             f[j_coord_offset+DIM*0+XX] -= tx;
 432             f[j_coord_offset+DIM*0+YY] -= ty;
 433             f[j_coord_offset+DIM*0+ZZ] -= tz;
 434
 435             /* Inner loop uses 61 flops */
 436         }
 437         /* End of innermost loop */
 438
 439         tx = ty = tz = 0;
 440         f[i_coord_offset+DIM*0+XX] += fix0;
 441         f[i_coord_offset+DIM*0+YY] += fiy0;
 442         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 443         tx                         += fix0;
 444         ty                         += fiy0;
 445         tz                         += fiz0;
 446         fshift[i_shift_offset+XX]  += tx;
 447         fshift[i_shift_offset+YY]  += ty;
 448         fshift[i_shift_offset+ZZ]  += tz;
 449
 450         /* Increment number of inner iterations */
 451         inneriter                  += j_index_end - j_index_start;
 452
 453         /* Outer loop uses 13 flops */
 454     }
 455
 456     /* Increment number of outer iterations */
 457     outeriter        += nri;
 458
 459     /* Update outer/inner flops */
 460
 461     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*61);
 462 }