src/gromacs/gmxlib/nonbonded/nb_kernel_avx_128_fma_single/nb_kernel_ElecCoul_VdwLJ_GeomW4P1_avx_128_fma_single.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014,2015,2017, 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 avx_128_fma_single kernel generator.
  37  */
  38 #include "gmxpre.h"
  39
  40 #include "config.h"
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "gromacs/gmxlib/nrnb.h"
  46
  47 #include "kernelutil_x86_avx_128_fma_single.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwLJ_GeomW4P1_VF_avx_128_fma_single
  51  * Electrostatics interaction: Coulomb
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecCoul_VdwLJ_GeomW4P1_VF_avx_128_fma_single
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      struct t_forcerec           * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  67      * just 0 for non-waters.
  68      * Suffixes A,B,C,D refer to j loop unrolling done with AVX_128, e.g. for the four different
  69      * jnr indices corresponding to data put in the four positions in the SIMD register.
  70      */
  71     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  72     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  73     int              jnrA,jnrB,jnrC,jnrD;
  74     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  75     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  76     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  77     real             rcutoff_scalar;
  78     real             *shiftvec,*fshift,*x,*f;
  79     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  80     real             scratch[4*DIM];
  81     __m128           fscal,rcutoff,rcutoff2,jidxall;
  82     int              vdwioffset0;
  83     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  84     int              vdwioffset1;
  85     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  86     int              vdwioffset2;
  87     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  88     int              vdwioffset3;
  89     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  90     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  91     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  92     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  93     __m128           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  94     __m128           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  95     __m128           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  96     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
  97     real             *charge;
  98     int              nvdwtype;
  99     __m128           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 100     int              *vdwtype;
 101     real             *vdwparam;
 102     __m128           one_sixth   = _mm_set1_ps(1.0/6.0);
 103     __m128           one_twelfth = _mm_set1_ps(1.0/12.0);
 104     __m128           dummy_mask,cutoff_mask;
 105     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 106     __m128           one     = _mm_set1_ps(1.0);
 107     __m128           two     = _mm_set1_ps(2.0);
 108     x                = xx[0];
 109     f                = ff[0];
 110
 111     nri              = nlist->nri;
 112     iinr             = nlist->iinr;
 113     jindex           = nlist->jindex;
 114     jjnr             = nlist->jjnr;
 115     shiftidx         = nlist->shift;
 116     gid              = nlist->gid;
 117     shiftvec         = fr->shift_vec[0];
 118     fshift           = fr->fshift[0];
 119     facel            = _mm_set1_ps(fr->ic->epsfac);
 120     charge           = mdatoms->chargeA;
 121     nvdwtype         = fr->ntype;
 122     vdwparam         = fr->nbfp;
 123     vdwtype          = mdatoms->typeA;
 124
 125     /* Setup water-specific parameters */
 126     inr              = nlist->iinr[0];
 127     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
 128     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
 129     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
 130     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 131
 132     /* Avoid stupid compiler warnings */
 133     jnrA = jnrB = jnrC = jnrD = 0;
 134     j_coord_offsetA = 0;
 135     j_coord_offsetB = 0;
 136     j_coord_offsetC = 0;
 137     j_coord_offsetD = 0;
 138
 139     outeriter        = 0;
 140     inneriter        = 0;
 141
 142     for(iidx=0;iidx<4*DIM;iidx++)
 143     {
 144         scratch[iidx] = 0.0;
 145     }
 146
 147     /* Start outer loop over neighborlists */
 148     for(iidx=0; iidx<nri; iidx++)
 149     {
 150         /* Load shift vector for this list */
 151         i_shift_offset   = DIM*shiftidx[iidx];
 152
 153         /* Load limits for loop over neighbors */
 154         j_index_start    = jindex[iidx];
 155         j_index_end      = jindex[iidx+1];
 156
 157         /* Get outer coordinate index */
 158         inr              = iinr[iidx];
 159         i_coord_offset   = DIM*inr;
 160
 161         /* Load i particle coords and add shift vector */
 162         gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 163                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 164
 165         fix0             = _mm_setzero_ps();
 166         fiy0             = _mm_setzero_ps();
 167         fiz0             = _mm_setzero_ps();
 168         fix1             = _mm_setzero_ps();
 169         fiy1             = _mm_setzero_ps();
 170         fiz1             = _mm_setzero_ps();
 171         fix2             = _mm_setzero_ps();
 172         fiy2             = _mm_setzero_ps();
 173         fiz2             = _mm_setzero_ps();
 174         fix3             = _mm_setzero_ps();
 175         fiy3             = _mm_setzero_ps();
 176         fiz3             = _mm_setzero_ps();
 177
 178         /* Reset potential sums */
 179         velecsum         = _mm_setzero_ps();
 180         vvdwsum          = _mm_setzero_ps();
 181
 182         /* Start inner kernel loop */
 183         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 184         {
 185
 186             /* Get j neighbor index, and coordinate index */
 187             jnrA             = jjnr[jidx];
 188             jnrB             = jjnr[jidx+1];
 189             jnrC             = jjnr[jidx+2];
 190             jnrD             = jjnr[jidx+3];
 191             j_coord_offsetA  = DIM*jnrA;
 192             j_coord_offsetB  = DIM*jnrB;
 193             j_coord_offsetC  = DIM*jnrC;
 194             j_coord_offsetD  = DIM*jnrD;
 195
 196             /* load j atom coordinates */
 197             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 198                                               x+j_coord_offsetC,x+j_coord_offsetD,
 199                                               &jx0,&jy0,&jz0);
 200
 201             /* Calculate displacement vector */
 202             dx00             = _mm_sub_ps(ix0,jx0);
 203             dy00             = _mm_sub_ps(iy0,jy0);
 204             dz00             = _mm_sub_ps(iz0,jz0);
 205             dx10             = _mm_sub_ps(ix1,jx0);
 206             dy10             = _mm_sub_ps(iy1,jy0);
 207             dz10             = _mm_sub_ps(iz1,jz0);
 208             dx20             = _mm_sub_ps(ix2,jx0);
 209             dy20             = _mm_sub_ps(iy2,jy0);
 210             dz20             = _mm_sub_ps(iz2,jz0);
 211             dx30             = _mm_sub_ps(ix3,jx0);
 212             dy30             = _mm_sub_ps(iy3,jy0);
 213             dz30             = _mm_sub_ps(iz3,jz0);
 214
 215             /* Calculate squared distance and things based on it */
 216             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 217             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 218             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 219             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 220
 221             rinv10           = avx128fma_invsqrt_f(rsq10);
 222             rinv20           = avx128fma_invsqrt_f(rsq20);
 223             rinv30           = avx128fma_invsqrt_f(rsq30);
 224
 225             rinvsq00         = avx128fma_inv_f(rsq00);
 226             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 227             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 228             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 229
 230             /* Load parameters for j particles */
 231             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 232                                                               charge+jnrC+0,charge+jnrD+0);
 233             vdwjidx0A        = 2*vdwtype[jnrA+0];
 234             vdwjidx0B        = 2*vdwtype[jnrB+0];
 235             vdwjidx0C        = 2*vdwtype[jnrC+0];
 236             vdwjidx0D        = 2*vdwtype[jnrD+0];
 237
 238             fjx0             = _mm_setzero_ps();
 239             fjy0             = _mm_setzero_ps();
 240             fjz0             = _mm_setzero_ps();
 241
 242             /**************************
 243              * CALCULATE INTERACTIONS *
 244              **************************/
 245
 246             /* Compute parameters for interactions between i and j atoms */
 247             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 248                                          vdwparam+vdwioffset0+vdwjidx0B,
 249                                          vdwparam+vdwioffset0+vdwjidx0C,
 250                                          vdwparam+vdwioffset0+vdwjidx0D,
 251                                          &c6_00,&c12_00);
 252
 253             /* LENNARD-JONES DISPERSION/REPULSION */
 254
 255             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 256             vvdw6            = _mm_mul_ps(c6_00,rinvsix);
 257             vvdw12           = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix));
 258             vvdw             = _mm_msub_ps(vvdw12,one_twelfth,_mm_mul_ps(vvdw6,one_sixth));
 259             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00);
 260
 261             /* Update potential sum for this i atom from the interaction with this j atom. */
 262             vvdwsum          = _mm_add_ps(vvdwsum,vvdw);
 263
 264             fscal            = fvdw;
 265
 266              /* Update vectorial force */
 267             fix0             = _mm_macc_ps(dx00,fscal,fix0);
 268             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
 269             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
 270
 271             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
 272             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
 273             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
 274
 275             /**************************
 276              * CALCULATE INTERACTIONS *
 277              **************************/
 278
 279             /* Compute parameters for interactions between i and j atoms */
 280             qq10             = _mm_mul_ps(iq1,jq0);
 281
 282             /* COULOMB ELECTROSTATICS */
 283             velec            = _mm_mul_ps(qq10,rinv10);
 284             felec            = _mm_mul_ps(velec,rinvsq10);
 285
 286             /* Update potential sum for this i atom from the interaction with this j atom. */
 287             velecsum         = _mm_add_ps(velecsum,velec);
 288
 289             fscal            = felec;
 290
 291              /* Update vectorial force */
 292             fix1             = _mm_macc_ps(dx10,fscal,fix1);
 293             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
 294             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
 295
 296             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
 297             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
 298             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
 299
 300             /**************************
 301              * CALCULATE INTERACTIONS *
 302              **************************/
 303
 304             /* Compute parameters for interactions between i and j atoms */
 305             qq20             = _mm_mul_ps(iq2,jq0);
 306
 307             /* COULOMB ELECTROSTATICS */
 308             velec            = _mm_mul_ps(qq20,rinv20);
 309             felec            = _mm_mul_ps(velec,rinvsq20);
 310
 311             /* Update potential sum for this i atom from the interaction with this j atom. */
 312             velecsum         = _mm_add_ps(velecsum,velec);
 313
 314             fscal            = felec;
 315
 316              /* Update vectorial force */
 317             fix2             = _mm_macc_ps(dx20,fscal,fix2);
 318             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
 319             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
 320
 321             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
 322             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
 323             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
 324
 325             /**************************
 326              * CALCULATE INTERACTIONS *
 327              **************************/
 328
 329             /* Compute parameters for interactions between i and j atoms */
 330             qq30             = _mm_mul_ps(iq3,jq0);
 331
 332             /* COULOMB ELECTROSTATICS */
 333             velec            = _mm_mul_ps(qq30,rinv30);
 334             felec            = _mm_mul_ps(velec,rinvsq30);
 335
 336             /* Update potential sum for this i atom from the interaction with this j atom. */
 337             velecsum         = _mm_add_ps(velecsum,velec);
 338
 339             fscal            = felec;
 340
 341              /* Update vectorial force */
 342             fix3             = _mm_macc_ps(dx30,fscal,fix3);
 343             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
 344             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
 345
 346             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
 347             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
 348             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
 349
 350             fjptrA             = f+j_coord_offsetA;
 351             fjptrB             = f+j_coord_offsetB;
 352             fjptrC             = f+j_coord_offsetC;
 353             fjptrD             = f+j_coord_offsetD;
 354
 355             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 356
 357             /* Inner loop uses 128 flops */
 358         }
 359
 360         if(jidx<j_index_end)
 361         {
 362
 363             /* Get j neighbor index, and coordinate index */
 364             jnrlistA         = jjnr[jidx];
 365             jnrlistB         = jjnr[jidx+1];
 366             jnrlistC         = jjnr[jidx+2];
 367             jnrlistD         = jjnr[jidx+3];
 368             /* Sign of each element will be negative for non-real atoms.
 369              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 370              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 371              */
 372             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 373             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 374             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 375             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 376             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 377             j_coord_offsetA  = DIM*jnrA;
 378             j_coord_offsetB  = DIM*jnrB;
 379             j_coord_offsetC  = DIM*jnrC;
 380             j_coord_offsetD  = DIM*jnrD;
 381
 382             /* load j atom coordinates */
 383             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 384                                               x+j_coord_offsetC,x+j_coord_offsetD,
 385                                               &jx0,&jy0,&jz0);
 386
 387             /* Calculate displacement vector */
 388             dx00             = _mm_sub_ps(ix0,jx0);
 389             dy00             = _mm_sub_ps(iy0,jy0);
 390             dz00             = _mm_sub_ps(iz0,jz0);
 391             dx10             = _mm_sub_ps(ix1,jx0);
 392             dy10             = _mm_sub_ps(iy1,jy0);
 393             dz10             = _mm_sub_ps(iz1,jz0);
 394             dx20             = _mm_sub_ps(ix2,jx0);
 395             dy20             = _mm_sub_ps(iy2,jy0);
 396             dz20             = _mm_sub_ps(iz2,jz0);
 397             dx30             = _mm_sub_ps(ix3,jx0);
 398             dy30             = _mm_sub_ps(iy3,jy0);
 399             dz30             = _mm_sub_ps(iz3,jz0);
 400
 401             /* Calculate squared distance and things based on it */
 402             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 403             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 404             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 405             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 406
 407             rinv10           = avx128fma_invsqrt_f(rsq10);
 408             rinv20           = avx128fma_invsqrt_f(rsq20);
 409             rinv30           = avx128fma_invsqrt_f(rsq30);
 410
 411             rinvsq00         = avx128fma_inv_f(rsq00);
 412             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 413             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 414             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 415
 416             /* Load parameters for j particles */
 417             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 418                                                               charge+jnrC+0,charge+jnrD+0);
 419             vdwjidx0A        = 2*vdwtype[jnrA+0];
 420             vdwjidx0B        = 2*vdwtype[jnrB+0];
 421             vdwjidx0C        = 2*vdwtype[jnrC+0];
 422             vdwjidx0D        = 2*vdwtype[jnrD+0];
 423
 424             fjx0             = _mm_setzero_ps();
 425             fjy0             = _mm_setzero_ps();
 426             fjz0             = _mm_setzero_ps();
 427
 428             /**************************
 429              * CALCULATE INTERACTIONS *
 430              **************************/
 431
 432             /* Compute parameters for interactions between i and j atoms */
 433             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 434                                          vdwparam+vdwioffset0+vdwjidx0B,
 435                                          vdwparam+vdwioffset0+vdwjidx0C,
 436                                          vdwparam+vdwioffset0+vdwjidx0D,
 437                                          &c6_00,&c12_00);
 438
 439             /* LENNARD-JONES DISPERSION/REPULSION */
 440
 441             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 442             vvdw6            = _mm_mul_ps(c6_00,rinvsix);
 443             vvdw12           = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix));
 444             vvdw             = _mm_msub_ps(vvdw12,one_twelfth,_mm_mul_ps(vvdw6,one_sixth));
 445             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00);
 446
 447             /* Update potential sum for this i atom from the interaction with this j atom. */
 448             vvdw             = _mm_andnot_ps(dummy_mask,vvdw);
 449             vvdwsum          = _mm_add_ps(vvdwsum,vvdw);
 450
 451             fscal            = fvdw;
 452
 453             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 454
 455              /* Update vectorial force */
 456             fix0             = _mm_macc_ps(dx00,fscal,fix0);
 457             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
 458             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
 459
 460             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
 461             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
 462             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
 463
 464             /**************************
 465              * CALCULATE INTERACTIONS *
 466              **************************/
 467
 468             /* Compute parameters for interactions between i and j atoms */
 469             qq10             = _mm_mul_ps(iq1,jq0);
 470
 471             /* COULOMB ELECTROSTATICS */
 472             velec            = _mm_mul_ps(qq10,rinv10);
 473             felec            = _mm_mul_ps(velec,rinvsq10);
 474
 475             /* Update potential sum for this i atom from the interaction with this j atom. */
 476             velec            = _mm_andnot_ps(dummy_mask,velec);
 477             velecsum         = _mm_add_ps(velecsum,velec);
 478
 479             fscal            = felec;
 480
 481             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 482
 483              /* Update vectorial force */
 484             fix1             = _mm_macc_ps(dx10,fscal,fix1);
 485             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
 486             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
 487
 488             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
 489             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
 490             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
 491
 492             /**************************
 493              * CALCULATE INTERACTIONS *
 494              **************************/
 495
 496             /* Compute parameters for interactions between i and j atoms */
 497             qq20             = _mm_mul_ps(iq2,jq0);
 498
 499             /* COULOMB ELECTROSTATICS */
 500             velec            = _mm_mul_ps(qq20,rinv20);
 501             felec            = _mm_mul_ps(velec,rinvsq20);
 502
 503             /* Update potential sum for this i atom from the interaction with this j atom. */
 504             velec            = _mm_andnot_ps(dummy_mask,velec);
 505             velecsum         = _mm_add_ps(velecsum,velec);
 506
 507             fscal            = felec;
 508
 509             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 510
 511              /* Update vectorial force */
 512             fix2             = _mm_macc_ps(dx20,fscal,fix2);
 513             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
 514             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
 515
 516             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
 517             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
 518             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
 519
 520             /**************************
 521              * CALCULATE INTERACTIONS *
 522              **************************/
 523
 524             /* Compute parameters for interactions between i and j atoms */
 525             qq30             = _mm_mul_ps(iq3,jq0);
 526
 527             /* COULOMB ELECTROSTATICS */
 528             velec            = _mm_mul_ps(qq30,rinv30);
 529             felec            = _mm_mul_ps(velec,rinvsq30);
 530
 531             /* Update potential sum for this i atom from the interaction with this j atom. */
 532             velec            = _mm_andnot_ps(dummy_mask,velec);
 533             velecsum         = _mm_add_ps(velecsum,velec);
 534
 535             fscal            = felec;
 536
 537             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 538
 539              /* Update vectorial force */
 540             fix3             = _mm_macc_ps(dx30,fscal,fix3);
 541             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
 542             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
 543
 544             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
 545             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
 546             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
 547
 548             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 549             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 550             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 551             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 552
 553             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 554
 555             /* Inner loop uses 128 flops */
 556         }
 557
 558         /* End of innermost loop */
 559
 560         gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 561                                               f+i_coord_offset,fshift+i_shift_offset);
 562
 563         ggid                        = gid[iidx];
 564         /* Update potential energies */
 565         gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 566         gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 567
 568         /* Increment number of inner iterations */
 569         inneriter                  += j_index_end - j_index_start;
 570
 571         /* Outer loop uses 26 flops */
 572     }
 573
 574     /* Increment number of outer iterations */
 575     outeriter        += nri;
 576
 577     /* Update outer/inner flops */
 578
 579     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*128);
 580 }
 581 /*
 582  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwLJ_GeomW4P1_F_avx_128_fma_single
 583  * Electrostatics interaction: Coulomb
 584  * VdW interaction:            LennardJones
 585  * Geometry:                   Water4-Particle
 586  * Calculate force/pot:        Force
 587  */
 588 void
 589 nb_kernel_ElecCoul_VdwLJ_GeomW4P1_F_avx_128_fma_single
 590                     (t_nblist                    * gmx_restrict       nlist,
 591                      rvec                        * gmx_restrict          xx,
 592                      rvec                        * gmx_restrict          ff,
 593                      struct t_forcerec           * gmx_restrict          fr,
 594                      t_mdatoms                   * gmx_restrict     mdatoms,
 595                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 596                      t_nrnb                      * gmx_restrict        nrnb)
 597 {
 598     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 599      * just 0 for non-waters.
 600      * Suffixes A,B,C,D refer to j loop unrolling done with AVX_128, e.g. for the four different
 601      * jnr indices corresponding to data put in the four positions in the SIMD register.
 602      */
 603     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 604     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 605     int              jnrA,jnrB,jnrC,jnrD;
 606     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 607     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 608     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 609     real             rcutoff_scalar;
 610     real             *shiftvec,*fshift,*x,*f;
 611     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 612     real             scratch[4*DIM];
 613     __m128           fscal,rcutoff,rcutoff2,jidxall;
 614     int              vdwioffset0;
 615     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 616     int              vdwioffset1;
 617     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 618     int              vdwioffset2;
 619     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 620     int              vdwioffset3;
 621     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 622     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 623     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 624     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 625     __m128           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 626     __m128           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 627     __m128           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 628     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
 629     real             *charge;
 630     int              nvdwtype;
 631     __m128           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 632     int              *vdwtype;
 633     real             *vdwparam;
 634     __m128           one_sixth   = _mm_set1_ps(1.0/6.0);
 635     __m128           one_twelfth = _mm_set1_ps(1.0/12.0);
 636     __m128           dummy_mask,cutoff_mask;
 637     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 638     __m128           one     = _mm_set1_ps(1.0);
 639     __m128           two     = _mm_set1_ps(2.0);
 640     x                = xx[0];
 641     f                = ff[0];
 642
 643     nri              = nlist->nri;
 644     iinr             = nlist->iinr;
 645     jindex           = nlist->jindex;
 646     jjnr             = nlist->jjnr;
 647     shiftidx         = nlist->shift;
 648     gid              = nlist->gid;
 649     shiftvec         = fr->shift_vec[0];
 650     fshift           = fr->fshift[0];
 651     facel            = _mm_set1_ps(fr->ic->epsfac);
 652     charge           = mdatoms->chargeA;
 653     nvdwtype         = fr->ntype;
 654     vdwparam         = fr->nbfp;
 655     vdwtype          = mdatoms->typeA;
 656
 657     /* Setup water-specific parameters */
 658     inr              = nlist->iinr[0];
 659     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
 660     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
 661     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
 662     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 663
 664     /* Avoid stupid compiler warnings */
 665     jnrA = jnrB = jnrC = jnrD = 0;
 666     j_coord_offsetA = 0;
 667     j_coord_offsetB = 0;
 668     j_coord_offsetC = 0;
 669     j_coord_offsetD = 0;
 670
 671     outeriter        = 0;
 672     inneriter        = 0;
 673
 674     for(iidx=0;iidx<4*DIM;iidx++)
 675     {
 676         scratch[iidx] = 0.0;
 677     }
 678
 679     /* Start outer loop over neighborlists */
 680     for(iidx=0; iidx<nri; iidx++)
 681     {
 682         /* Load shift vector for this list */
 683         i_shift_offset   = DIM*shiftidx[iidx];
 684
 685         /* Load limits for loop over neighbors */
 686         j_index_start    = jindex[iidx];
 687         j_index_end      = jindex[iidx+1];
 688
 689         /* Get outer coordinate index */
 690         inr              = iinr[iidx];
 691         i_coord_offset   = DIM*inr;
 692
 693         /* Load i particle coords and add shift vector */
 694         gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 695                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 696
 697         fix0             = _mm_setzero_ps();
 698         fiy0             = _mm_setzero_ps();
 699         fiz0             = _mm_setzero_ps();
 700         fix1             = _mm_setzero_ps();
 701         fiy1             = _mm_setzero_ps();
 702         fiz1             = _mm_setzero_ps();
 703         fix2             = _mm_setzero_ps();
 704         fiy2             = _mm_setzero_ps();
 705         fiz2             = _mm_setzero_ps();
 706         fix3             = _mm_setzero_ps();
 707         fiy3             = _mm_setzero_ps();
 708         fiz3             = _mm_setzero_ps();
 709
 710         /* Start inner kernel loop */
 711         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 712         {
 713
 714             /* Get j neighbor index, and coordinate index */
 715             jnrA             = jjnr[jidx];
 716             jnrB             = jjnr[jidx+1];
 717             jnrC             = jjnr[jidx+2];
 718             jnrD             = jjnr[jidx+3];
 719             j_coord_offsetA  = DIM*jnrA;
 720             j_coord_offsetB  = DIM*jnrB;
 721             j_coord_offsetC  = DIM*jnrC;
 722             j_coord_offsetD  = DIM*jnrD;
 723
 724             /* load j atom coordinates */
 725             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 726                                               x+j_coord_offsetC,x+j_coord_offsetD,
 727                                               &jx0,&jy0,&jz0);
 728
 729             /* Calculate displacement vector */
 730             dx00             = _mm_sub_ps(ix0,jx0);
 731             dy00             = _mm_sub_ps(iy0,jy0);
 732             dz00             = _mm_sub_ps(iz0,jz0);
 733             dx10             = _mm_sub_ps(ix1,jx0);
 734             dy10             = _mm_sub_ps(iy1,jy0);
 735             dz10             = _mm_sub_ps(iz1,jz0);
 736             dx20             = _mm_sub_ps(ix2,jx0);
 737             dy20             = _mm_sub_ps(iy2,jy0);
 738             dz20             = _mm_sub_ps(iz2,jz0);
 739             dx30             = _mm_sub_ps(ix3,jx0);
 740             dy30             = _mm_sub_ps(iy3,jy0);
 741             dz30             = _mm_sub_ps(iz3,jz0);
 742
 743             /* Calculate squared distance and things based on it */
 744             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 745             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 746             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 747             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 748
 749             rinv10           = avx128fma_invsqrt_f(rsq10);
 750             rinv20           = avx128fma_invsqrt_f(rsq20);
 751             rinv30           = avx128fma_invsqrt_f(rsq30);
 752
 753             rinvsq00         = avx128fma_inv_f(rsq00);
 754             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 755             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 756             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 757
 758             /* Load parameters for j particles */
 759             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 760                                                               charge+jnrC+0,charge+jnrD+0);
 761             vdwjidx0A        = 2*vdwtype[jnrA+0];
 762             vdwjidx0B        = 2*vdwtype[jnrB+0];
 763             vdwjidx0C        = 2*vdwtype[jnrC+0];
 764             vdwjidx0D        = 2*vdwtype[jnrD+0];
 765
 766             fjx0             = _mm_setzero_ps();
 767             fjy0             = _mm_setzero_ps();
 768             fjz0             = _mm_setzero_ps();
 769
 770             /**************************
 771              * CALCULATE INTERACTIONS *
 772              **************************/
 773
 774             /* Compute parameters for interactions between i and j atoms */
 775             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 776                                          vdwparam+vdwioffset0+vdwjidx0B,
 777                                          vdwparam+vdwioffset0+vdwjidx0C,
 778                                          vdwparam+vdwioffset0+vdwjidx0D,
 779                                          &c6_00,&c12_00);
 780
 781             /* LENNARD-JONES DISPERSION/REPULSION */
 782
 783             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 784             fvdw             = _mm_mul_ps(_mm_msub_ps(c12_00,rinvsix,c6_00),_mm_mul_ps(rinvsix,rinvsq00));
 785
 786             fscal            = fvdw;
 787
 788              /* Update vectorial force */
 789             fix0             = _mm_macc_ps(dx00,fscal,fix0);
 790             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
 791             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
 792
 793             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
 794             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
 795             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
 796
 797             /**************************
 798              * CALCULATE INTERACTIONS *
 799              **************************/
 800
 801             /* Compute parameters for interactions between i and j atoms */
 802             qq10             = _mm_mul_ps(iq1,jq0);
 803
 804             /* COULOMB ELECTROSTATICS */
 805             velec            = _mm_mul_ps(qq10,rinv10);
 806             felec            = _mm_mul_ps(velec,rinvsq10);
 807
 808             fscal            = felec;
 809
 810              /* Update vectorial force */
 811             fix1             = _mm_macc_ps(dx10,fscal,fix1);
 812             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
 813             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
 814
 815             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
 816             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
 817             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
 818
 819             /**************************
 820              * CALCULATE INTERACTIONS *
 821              **************************/
 822
 823             /* Compute parameters for interactions between i and j atoms */
 824             qq20             = _mm_mul_ps(iq2,jq0);
 825
 826             /* COULOMB ELECTROSTATICS */
 827             velec            = _mm_mul_ps(qq20,rinv20);
 828             felec            = _mm_mul_ps(velec,rinvsq20);
 829
 830             fscal            = felec;
 831
 832              /* Update vectorial force */
 833             fix2             = _mm_macc_ps(dx20,fscal,fix2);
 834             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
 835             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
 836
 837             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
 838             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
 839             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
 840
 841             /**************************
 842              * CALCULATE INTERACTIONS *
 843              **************************/
 844
 845             /* Compute parameters for interactions between i and j atoms */
 846             qq30             = _mm_mul_ps(iq3,jq0);
 847
 848             /* COULOMB ELECTROSTATICS */
 849             velec            = _mm_mul_ps(qq30,rinv30);
 850             felec            = _mm_mul_ps(velec,rinvsq30);
 851
 852             fscal            = felec;
 853
 854              /* Update vectorial force */
 855             fix3             = _mm_macc_ps(dx30,fscal,fix3);
 856             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
 857             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
 858
 859             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
 860             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
 861             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
 862
 863             fjptrA             = f+j_coord_offsetA;
 864             fjptrB             = f+j_coord_offsetB;
 865             fjptrC             = f+j_coord_offsetC;
 866             fjptrD             = f+j_coord_offsetD;
 867
 868             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 869
 870             /* Inner loop uses 120 flops */
 871         }
 872
 873         if(jidx<j_index_end)
 874         {
 875
 876             /* Get j neighbor index, and coordinate index */
 877             jnrlistA         = jjnr[jidx];
 878             jnrlistB         = jjnr[jidx+1];
 879             jnrlistC         = jjnr[jidx+2];
 880             jnrlistD         = jjnr[jidx+3];
 881             /* Sign of each element will be negative for non-real atoms.
 882              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 883              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 884              */
 885             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 886             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 887             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 888             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 889             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 890             j_coord_offsetA  = DIM*jnrA;
 891             j_coord_offsetB  = DIM*jnrB;
 892             j_coord_offsetC  = DIM*jnrC;
 893             j_coord_offsetD  = DIM*jnrD;
 894
 895             /* load j atom coordinates */
 896             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 897                                               x+j_coord_offsetC,x+j_coord_offsetD,
 898                                               &jx0,&jy0,&jz0);
 899
 900             /* Calculate displacement vector */
 901             dx00             = _mm_sub_ps(ix0,jx0);
 902             dy00             = _mm_sub_ps(iy0,jy0);
 903             dz00             = _mm_sub_ps(iz0,jz0);
 904             dx10             = _mm_sub_ps(ix1,jx0);
 905             dy10             = _mm_sub_ps(iy1,jy0);
 906             dz10             = _mm_sub_ps(iz1,jz0);
 907             dx20             = _mm_sub_ps(ix2,jx0);
 908             dy20             = _mm_sub_ps(iy2,jy0);
 909             dz20             = _mm_sub_ps(iz2,jz0);
 910             dx30             = _mm_sub_ps(ix3,jx0);
 911             dy30             = _mm_sub_ps(iy3,jy0);
 912             dz30             = _mm_sub_ps(iz3,jz0);
 913
 914             /* Calculate squared distance and things based on it */
 915             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 916             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 917             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 918             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 919
 920             rinv10           = avx128fma_invsqrt_f(rsq10);
 921             rinv20           = avx128fma_invsqrt_f(rsq20);
 922             rinv30           = avx128fma_invsqrt_f(rsq30);
 923
 924             rinvsq00         = avx128fma_inv_f(rsq00);
 925             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 926             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 927             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 928
 929             /* Load parameters for j particles */
 930             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 931                                                               charge+jnrC+0,charge+jnrD+0);
 932             vdwjidx0A        = 2*vdwtype[jnrA+0];
 933             vdwjidx0B        = 2*vdwtype[jnrB+0];
 934             vdwjidx0C        = 2*vdwtype[jnrC+0];
 935             vdwjidx0D        = 2*vdwtype[jnrD+0];
 936
 937             fjx0             = _mm_setzero_ps();
 938             fjy0             = _mm_setzero_ps();
 939             fjz0             = _mm_setzero_ps();
 940
 941             /**************************
 942              * CALCULATE INTERACTIONS *
 943              **************************/
 944
 945             /* Compute parameters for interactions between i and j atoms */
 946             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 947                                          vdwparam+vdwioffset0+vdwjidx0B,
 948                                          vdwparam+vdwioffset0+vdwjidx0C,
 949                                          vdwparam+vdwioffset0+vdwjidx0D,
 950                                          &c6_00,&c12_00);
 951
 952             /* LENNARD-JONES DISPERSION/REPULSION */
 953
 954             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 955             fvdw             = _mm_mul_ps(_mm_msub_ps(c12_00,rinvsix,c6_00),_mm_mul_ps(rinvsix,rinvsq00));
 956
 957             fscal            = fvdw;
 958
 959             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 960
 961              /* Update vectorial force */
 962             fix0             = _mm_macc_ps(dx00,fscal,fix0);
 963             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
 964             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
 965
 966             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
 967             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
 968             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
 969
 970             /**************************
 971              * CALCULATE INTERACTIONS *
 972              **************************/
 973
 974             /* Compute parameters for interactions between i and j atoms */
 975             qq10             = _mm_mul_ps(iq1,jq0);
 976
 977             /* COULOMB ELECTROSTATICS */
 978             velec            = _mm_mul_ps(qq10,rinv10);
 979             felec            = _mm_mul_ps(velec,rinvsq10);
 980
 981             fscal            = felec;
 982
 983             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 984
 985              /* Update vectorial force */
 986             fix1             = _mm_macc_ps(dx10,fscal,fix1);
 987             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
 988             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
 989
 990             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
 991             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
 992             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
 993
 994             /**************************
 995              * CALCULATE INTERACTIONS *
 996              **************************/
 997
 998             /* Compute parameters for interactions between i and j atoms */
 999             qq20             = _mm_mul_ps(iq2,jq0);
1000
1001             /* COULOMB ELECTROSTATICS */
1002             velec            = _mm_mul_ps(qq20,rinv20);
1003             felec            = _mm_mul_ps(velec,rinvsq20);
1004
1005             fscal            = felec;
1006
1007             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1008
1009              /* Update vectorial force */
1010             fix2             = _mm_macc_ps(dx20,fscal,fix2);
1011             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
1012             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
1013
1014             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
1015             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
1016             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
1017
1018             /**************************
1019              * CALCULATE INTERACTIONS *
1020              **************************/
1021
1022             /* Compute parameters for interactions between i and j atoms */
1023             qq30             = _mm_mul_ps(iq3,jq0);
1024
1025             /* COULOMB ELECTROSTATICS */
1026             velec            = _mm_mul_ps(qq30,rinv30);
1027             felec            = _mm_mul_ps(velec,rinvsq30);
1028
1029             fscal            = felec;
1030
1031             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1032
1033              /* Update vectorial force */
1034             fix3             = _mm_macc_ps(dx30,fscal,fix3);
1035             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
1036             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
1037
1038             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
1039             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
1040             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
1041
1042             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1043             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1044             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1045             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1046
1047             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1048
1049             /* Inner loop uses 120 flops */
1050         }
1051
1052         /* End of innermost loop */
1053
1054         gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1055                                               f+i_coord_offset,fshift+i_shift_offset);
1056
1057         /* Increment number of inner iterations */
1058         inneriter                  += j_index_end - j_index_start;
1059
1060         /* Outer loop uses 24 flops */
1061     }
1062
1063     /* Increment number of outer iterations */
1064     outeriter        += nri;
1065
1066     /* Update outer/inner flops */
1067
1068     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*120);
1069 }