src/gromacs/gmxlib/nonbonded/nb_kernel_sse4_1_single/nb_kernel_ElecEwSh_VdwLJEwSh_GeomW4P1_sse4_1_single.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, 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 sse4_1_single kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_sse4_1_single.h"
  48 #include "kernelutil_x86_sse4_1_single.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJEwSh_GeomW4P1_VF_sse4_1_single
  52  * Electrostatics interaction: Ewald
  53  * VdW interaction:            LJEwald
  54  * Geometry:                   Water4-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecEwSh_VdwLJEwSh_GeomW4P1_VF_sse4_1_single
  59                     (t_nblist                    * gmx_restrict       nlist,
  60                      rvec                        * gmx_restrict          xx,
  61                      rvec                        * gmx_restrict          ff,
  62                      t_forcerec                  * gmx_restrict          fr,
  63                      t_mdatoms                   * gmx_restrict     mdatoms,
  64                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  65                      t_nrnb                      * gmx_restrict        nrnb)
  66 {
  67     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  68      * just 0 for non-waters.
  69      * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
  70      * jnr indices corresponding to data put in the four positions in the SIMD register.
  71      */
  72     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  73     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  74     int              jnrA,jnrB,jnrC,jnrD;
  75     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  76     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  77     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  78     real             rcutoff_scalar;
  79     real             *shiftvec,*fshift,*x,*f;
  80     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  81     real             scratch[4*DIM];
  82     __m128           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  83     int              vdwioffset0;
  84     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  85     int              vdwioffset1;
  86     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  87     int              vdwioffset2;
  88     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  89     int              vdwioffset3;
  90     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  91     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  92     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  93     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  94     __m128           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  95     __m128           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  96     __m128           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  97     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
  98     real             *charge;
  99     int              nvdwtype;
 100     __m128           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 101     int              *vdwtype;
 102     real             *vdwparam;
 103     __m128           one_sixth   = _mm_set1_ps(1.0/6.0);
 104     __m128           one_twelfth = _mm_set1_ps(1.0/12.0);
 105     __m128           c6grid_00;
 106     __m128           c6grid_10;
 107     __m128           c6grid_20;
 108     __m128           c6grid_30;
 109     __m128           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 110     real             *vdwgridparam;
 111     __m128           one_half  = _mm_set1_ps(0.5);
 112     __m128           minus_one = _mm_set1_ps(-1.0);
 113     __m128i          ewitab;
 114     __m128           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 115     real             *ewtab;
 116     __m128           dummy_mask,cutoff_mask;
 117     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 118     __m128           one     = _mm_set1_ps(1.0);
 119     __m128           two     = _mm_set1_ps(2.0);
 120     x                = xx[0];
 121     f                = ff[0];
 122
 123     nri              = nlist->nri;
 124     iinr             = nlist->iinr;
 125     jindex           = nlist->jindex;
 126     jjnr             = nlist->jjnr;
 127     shiftidx         = nlist->shift;
 128     gid              = nlist->gid;
 129     shiftvec         = fr->shift_vec[0];
 130     fshift           = fr->fshift[0];
 131     facel            = _mm_set1_ps(fr->epsfac);
 132     charge           = mdatoms->chargeA;
 133     nvdwtype         = fr->ntype;
 134     vdwparam         = fr->nbfp;
 135     vdwtype          = mdatoms->typeA;
 136     vdwgridparam     = fr->ljpme_c6grid;
 137     sh_lj_ewald      = _mm_set1_ps(fr->ic->sh_lj_ewald);
 138     ewclj            = _mm_set1_ps(fr->ewaldcoeff_lj);
 139     ewclj2           = _mm_mul_ps(minus_one,_mm_mul_ps(ewclj,ewclj));
 140
 141     sh_ewald         = _mm_set1_ps(fr->ic->sh_ewald);
 142     ewtab            = fr->ic->tabq_coul_FDV0;
 143     ewtabscale       = _mm_set1_ps(fr->ic->tabq_scale);
 144     ewtabhalfspace   = _mm_set1_ps(0.5/fr->ic->tabq_scale);
 145
 146     /* Setup water-specific parameters */
 147     inr              = nlist->iinr[0];
 148     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
 149     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
 150     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
 151     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 152
 153     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 154     rcutoff_scalar   = fr->rcoulomb;
 155     rcutoff          = _mm_set1_ps(rcutoff_scalar);
 156     rcutoff2         = _mm_mul_ps(rcutoff,rcutoff);
 157
 158     sh_vdw_invrcut6  = _mm_set1_ps(fr->ic->sh_invrc6);
 159     rvdw             = _mm_set1_ps(fr->rvdw);
 160
 161     /* Avoid stupid compiler warnings */
 162     jnrA = jnrB = jnrC = jnrD = 0;
 163     j_coord_offsetA = 0;
 164     j_coord_offsetB = 0;
 165     j_coord_offsetC = 0;
 166     j_coord_offsetD = 0;
 167
 168     outeriter        = 0;
 169     inneriter        = 0;
 170
 171     for(iidx=0;iidx<4*DIM;iidx++)
 172     {
 173         scratch[iidx] = 0.0;
 174     }
 175
 176     /* Start outer loop over neighborlists */
 177     for(iidx=0; iidx<nri; iidx++)
 178     {
 179         /* Load shift vector for this list */
 180         i_shift_offset   = DIM*shiftidx[iidx];
 181
 182         /* Load limits for loop over neighbors */
 183         j_index_start    = jindex[iidx];
 184         j_index_end      = jindex[iidx+1];
 185
 186         /* Get outer coordinate index */
 187         inr              = iinr[iidx];
 188         i_coord_offset   = DIM*inr;
 189
 190         /* Load i particle coords and add shift vector */
 191         gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 192                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 193
 194         fix0             = _mm_setzero_ps();
 195         fiy0             = _mm_setzero_ps();
 196         fiz0             = _mm_setzero_ps();
 197         fix1             = _mm_setzero_ps();
 198         fiy1             = _mm_setzero_ps();
 199         fiz1             = _mm_setzero_ps();
 200         fix2             = _mm_setzero_ps();
 201         fiy2             = _mm_setzero_ps();
 202         fiz2             = _mm_setzero_ps();
 203         fix3             = _mm_setzero_ps();
 204         fiy3             = _mm_setzero_ps();
 205         fiz3             = _mm_setzero_ps();
 206
 207         /* Reset potential sums */
 208         velecsum         = _mm_setzero_ps();
 209         vvdwsum          = _mm_setzero_ps();
 210
 211         /* Start inner kernel loop */
 212         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 213         {
 214
 215             /* Get j neighbor index, and coordinate index */
 216             jnrA             = jjnr[jidx];
 217             jnrB             = jjnr[jidx+1];
 218             jnrC             = jjnr[jidx+2];
 219             jnrD             = jjnr[jidx+3];
 220             j_coord_offsetA  = DIM*jnrA;
 221             j_coord_offsetB  = DIM*jnrB;
 222             j_coord_offsetC  = DIM*jnrC;
 223             j_coord_offsetD  = DIM*jnrD;
 224
 225             /* load j atom coordinates */
 226             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 227                                               x+j_coord_offsetC,x+j_coord_offsetD,
 228                                               &jx0,&jy0,&jz0);
 229
 230             /* Calculate displacement vector */
 231             dx00             = _mm_sub_ps(ix0,jx0);
 232             dy00             = _mm_sub_ps(iy0,jy0);
 233             dz00             = _mm_sub_ps(iz0,jz0);
 234             dx10             = _mm_sub_ps(ix1,jx0);
 235             dy10             = _mm_sub_ps(iy1,jy0);
 236             dz10             = _mm_sub_ps(iz1,jz0);
 237             dx20             = _mm_sub_ps(ix2,jx0);
 238             dy20             = _mm_sub_ps(iy2,jy0);
 239             dz20             = _mm_sub_ps(iz2,jz0);
 240             dx30             = _mm_sub_ps(ix3,jx0);
 241             dy30             = _mm_sub_ps(iy3,jy0);
 242             dz30             = _mm_sub_ps(iz3,jz0);
 243
 244             /* Calculate squared distance and things based on it */
 245             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 246             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 247             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 248             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 249
 250             rinv00           = gmx_mm_invsqrt_ps(rsq00);
 251             rinv10           = gmx_mm_invsqrt_ps(rsq10);
 252             rinv20           = gmx_mm_invsqrt_ps(rsq20);
 253             rinv30           = gmx_mm_invsqrt_ps(rsq30);
 254
 255             rinvsq00         = _mm_mul_ps(rinv00,rinv00);
 256             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 257             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 258             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 259
 260             /* Load parameters for j particles */
 261             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 262                                                               charge+jnrC+0,charge+jnrD+0);
 263             vdwjidx0A        = 2*vdwtype[jnrA+0];
 264             vdwjidx0B        = 2*vdwtype[jnrB+0];
 265             vdwjidx0C        = 2*vdwtype[jnrC+0];
 266             vdwjidx0D        = 2*vdwtype[jnrD+0];
 267
 268             fjx0             = _mm_setzero_ps();
 269             fjy0             = _mm_setzero_ps();
 270             fjz0             = _mm_setzero_ps();
 271
 272             /**************************
 273              * CALCULATE INTERACTIONS *
 274              **************************/
 275
 276             if (gmx_mm_any_lt(rsq00,rcutoff2))
 277             {
 278
 279             r00              = _mm_mul_ps(rsq00,rinv00);
 280
 281             /* Compute parameters for interactions between i and j atoms */
 282             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 283                                          vdwparam+vdwioffset0+vdwjidx0B,
 284                                          vdwparam+vdwioffset0+vdwjidx0C,
 285                                          vdwparam+vdwioffset0+vdwjidx0D,
 286                                          &c6_00,&c12_00);
 287
 288             c6grid_00       = gmx_mm_load_4real_swizzle_ps(vdwgridparam+vdwioffset0+vdwjidx0A,
 289                                                                vdwgridparam+vdwioffset0+vdwjidx0B,
 290                                                                vdwgridparam+vdwioffset0+vdwjidx0C,
 291                                                                vdwgridparam+vdwioffset0+vdwjidx0D);
 292
 293             /* Analytical LJ-PME */
 294             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 295             ewcljrsq         = _mm_mul_ps(ewclj2,rsq00);
 296             ewclj6           = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2));
 297             exponent         = gmx_simd_exp_r(ewcljrsq);
 298             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 299             poly             = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 300             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 301             vvdw6            = _mm_mul_ps(_mm_sub_ps(c6_00,_mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly))),rinvsix);
 302             vvdw12           = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix));
 303             vvdw             = _mm_sub_ps(_mm_mul_ps( _mm_sub_ps(vvdw12 , _mm_mul_ps(c12_00,_mm_mul_ps(sh_vdw_invrcut6,sh_vdw_invrcut6))),one_twelfth),
 304                                _mm_mul_ps( _mm_sub_ps(vvdw6,_mm_add_ps(_mm_mul_ps(c6_00,sh_vdw_invrcut6),_mm_mul_ps(c6grid_00,sh_lj_ewald))),one_sixth));
 305             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 306             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,_mm_sub_ps(vvdw6,_mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6)))),rinvsq00);
 307
 308             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
 309
 310             /* Update potential sum for this i atom from the interaction with this j atom. */
 311             vvdw             = _mm_and_ps(vvdw,cutoff_mask);
 312             vvdwsum          = _mm_add_ps(vvdwsum,vvdw);
 313
 314             fscal            = fvdw;
 315
 316             fscal            = _mm_and_ps(fscal,cutoff_mask);
 317
 318             /* Calculate temporary vectorial force */
 319             tx               = _mm_mul_ps(fscal,dx00);
 320             ty               = _mm_mul_ps(fscal,dy00);
 321             tz               = _mm_mul_ps(fscal,dz00);
 322
 323             /* Update vectorial force */
 324             fix0             = _mm_add_ps(fix0,tx);
 325             fiy0             = _mm_add_ps(fiy0,ty);
 326             fiz0             = _mm_add_ps(fiz0,tz);
 327
 328             fjx0             = _mm_add_ps(fjx0,tx);
 329             fjy0             = _mm_add_ps(fjy0,ty);
 330             fjz0             = _mm_add_ps(fjz0,tz);
 331
 332             }
 333
 334             /**************************
 335              * CALCULATE INTERACTIONS *
 336              **************************/
 337
 338             if (gmx_mm_any_lt(rsq10,rcutoff2))
 339             {
 340
 341             r10              = _mm_mul_ps(rsq10,rinv10);
 342
 343             /* Compute parameters for interactions between i and j atoms */
 344             qq10             = _mm_mul_ps(iq1,jq0);
 345
 346             /* EWALD ELECTROSTATICS */
 347
 348             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 349             ewrt             = _mm_mul_ps(r10,ewtabscale);
 350             ewitab           = _mm_cvttps_epi32(ewrt);
 351             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
 352             ewitab           = _mm_slli_epi32(ewitab,2);
 353             ewtabF           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
 354             ewtabD           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
 355             ewtabV           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
 356             ewtabFn          = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
 357             _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
 358             felec            = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
 359             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 360             velec            = _mm_mul_ps(qq10,_mm_sub_ps(_mm_sub_ps(rinv10,sh_ewald),velec));
 361             felec            = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec));
 362
 363             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
 364
 365             /* Update potential sum for this i atom from the interaction with this j atom. */
 366             velec            = _mm_and_ps(velec,cutoff_mask);
 367             velecsum         = _mm_add_ps(velecsum,velec);
 368
 369             fscal            = felec;
 370
 371             fscal            = _mm_and_ps(fscal,cutoff_mask);
 372
 373             /* Calculate temporary vectorial force */
 374             tx               = _mm_mul_ps(fscal,dx10);
 375             ty               = _mm_mul_ps(fscal,dy10);
 376             tz               = _mm_mul_ps(fscal,dz10);
 377
 378             /* Update vectorial force */
 379             fix1             = _mm_add_ps(fix1,tx);
 380             fiy1             = _mm_add_ps(fiy1,ty);
 381             fiz1             = _mm_add_ps(fiz1,tz);
 382
 383             fjx0             = _mm_add_ps(fjx0,tx);
 384             fjy0             = _mm_add_ps(fjy0,ty);
 385             fjz0             = _mm_add_ps(fjz0,tz);
 386
 387             }
 388
 389             /**************************
 390              * CALCULATE INTERACTIONS *
 391              **************************/
 392
 393             if (gmx_mm_any_lt(rsq20,rcutoff2))
 394             {
 395
 396             r20              = _mm_mul_ps(rsq20,rinv20);
 397
 398             /* Compute parameters for interactions between i and j atoms */
 399             qq20             = _mm_mul_ps(iq2,jq0);
 400
 401             /* EWALD ELECTROSTATICS */
 402
 403             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 404             ewrt             = _mm_mul_ps(r20,ewtabscale);
 405             ewitab           = _mm_cvttps_epi32(ewrt);
 406             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
 407             ewitab           = _mm_slli_epi32(ewitab,2);
 408             ewtabF           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
 409             ewtabD           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
 410             ewtabV           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
 411             ewtabFn          = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
 412             _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
 413             felec            = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
 414             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 415             velec            = _mm_mul_ps(qq20,_mm_sub_ps(_mm_sub_ps(rinv20,sh_ewald),velec));
 416             felec            = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec));
 417
 418             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
 419
 420             /* Update potential sum for this i atom from the interaction with this j atom. */
 421             velec            = _mm_and_ps(velec,cutoff_mask);
 422             velecsum         = _mm_add_ps(velecsum,velec);
 423
 424             fscal            = felec;
 425
 426             fscal            = _mm_and_ps(fscal,cutoff_mask);
 427
 428             /* Calculate temporary vectorial force */
 429             tx               = _mm_mul_ps(fscal,dx20);
 430             ty               = _mm_mul_ps(fscal,dy20);
 431             tz               = _mm_mul_ps(fscal,dz20);
 432
 433             /* Update vectorial force */
 434             fix2             = _mm_add_ps(fix2,tx);
 435             fiy2             = _mm_add_ps(fiy2,ty);
 436             fiz2             = _mm_add_ps(fiz2,tz);
 437
 438             fjx0             = _mm_add_ps(fjx0,tx);
 439             fjy0             = _mm_add_ps(fjy0,ty);
 440             fjz0             = _mm_add_ps(fjz0,tz);
 441
 442             }
 443
 444             /**************************
 445              * CALCULATE INTERACTIONS *
 446              **************************/
 447
 448             if (gmx_mm_any_lt(rsq30,rcutoff2))
 449             {
 450
 451             r30              = _mm_mul_ps(rsq30,rinv30);
 452
 453             /* Compute parameters for interactions between i and j atoms */
 454             qq30             = _mm_mul_ps(iq3,jq0);
 455
 456             /* EWALD ELECTROSTATICS */
 457
 458             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 459             ewrt             = _mm_mul_ps(r30,ewtabscale);
 460             ewitab           = _mm_cvttps_epi32(ewrt);
 461             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
 462             ewitab           = _mm_slli_epi32(ewitab,2);
 463             ewtabF           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
 464             ewtabD           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
 465             ewtabV           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
 466             ewtabFn          = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
 467             _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
 468             felec            = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
 469             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 470             velec            = _mm_mul_ps(qq30,_mm_sub_ps(_mm_sub_ps(rinv30,sh_ewald),velec));
 471             felec            = _mm_mul_ps(_mm_mul_ps(qq30,rinv30),_mm_sub_ps(rinvsq30,felec));
 472
 473             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
 474
 475             /* Update potential sum for this i atom from the interaction with this j atom. */
 476             velec            = _mm_and_ps(velec,cutoff_mask);
 477             velecsum         = _mm_add_ps(velecsum,velec);
 478
 479             fscal            = felec;
 480
 481             fscal            = _mm_and_ps(fscal,cutoff_mask);
 482
 483             /* Calculate temporary vectorial force */
 484             tx               = _mm_mul_ps(fscal,dx30);
 485             ty               = _mm_mul_ps(fscal,dy30);
 486             tz               = _mm_mul_ps(fscal,dz30);
 487
 488             /* Update vectorial force */
 489             fix3             = _mm_add_ps(fix3,tx);
 490             fiy3             = _mm_add_ps(fiy3,ty);
 491             fiz3             = _mm_add_ps(fiz3,tz);
 492
 493             fjx0             = _mm_add_ps(fjx0,tx);
 494             fjy0             = _mm_add_ps(fjy0,ty);
 495             fjz0             = _mm_add_ps(fjz0,tz);
 496
 497             }
 498
 499             fjptrA             = f+j_coord_offsetA;
 500             fjptrB             = f+j_coord_offsetB;
 501             fjptrC             = f+j_coord_offsetC;
 502             fjptrD             = f+j_coord_offsetD;
 503
 504             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 505
 506             /* Inner loop uses 200 flops */
 507         }
 508
 509         if(jidx<j_index_end)
 510         {
 511
 512             /* Get j neighbor index, and coordinate index */
 513             jnrlistA         = jjnr[jidx];
 514             jnrlistB         = jjnr[jidx+1];
 515             jnrlistC         = jjnr[jidx+2];
 516             jnrlistD         = jjnr[jidx+3];
 517             /* Sign of each element will be negative for non-real atoms.
 518              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 519              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 520              */
 521             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 522             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 523             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 524             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 525             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 526             j_coord_offsetA  = DIM*jnrA;
 527             j_coord_offsetB  = DIM*jnrB;
 528             j_coord_offsetC  = DIM*jnrC;
 529             j_coord_offsetD  = DIM*jnrD;
 530
 531             /* load j atom coordinates */
 532             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 533                                               x+j_coord_offsetC,x+j_coord_offsetD,
 534                                               &jx0,&jy0,&jz0);
 535
 536             /* Calculate displacement vector */
 537             dx00             = _mm_sub_ps(ix0,jx0);
 538             dy00             = _mm_sub_ps(iy0,jy0);
 539             dz00             = _mm_sub_ps(iz0,jz0);
 540             dx10             = _mm_sub_ps(ix1,jx0);
 541             dy10             = _mm_sub_ps(iy1,jy0);
 542             dz10             = _mm_sub_ps(iz1,jz0);
 543             dx20             = _mm_sub_ps(ix2,jx0);
 544             dy20             = _mm_sub_ps(iy2,jy0);
 545             dz20             = _mm_sub_ps(iz2,jz0);
 546             dx30             = _mm_sub_ps(ix3,jx0);
 547             dy30             = _mm_sub_ps(iy3,jy0);
 548             dz30             = _mm_sub_ps(iz3,jz0);
 549
 550             /* Calculate squared distance and things based on it */
 551             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 552             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 553             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 554             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 555
 556             rinv00           = gmx_mm_invsqrt_ps(rsq00);
 557             rinv10           = gmx_mm_invsqrt_ps(rsq10);
 558             rinv20           = gmx_mm_invsqrt_ps(rsq20);
 559             rinv30           = gmx_mm_invsqrt_ps(rsq30);
 560
 561             rinvsq00         = _mm_mul_ps(rinv00,rinv00);
 562             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 563             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 564             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 565
 566             /* Load parameters for j particles */
 567             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 568                                                               charge+jnrC+0,charge+jnrD+0);
 569             vdwjidx0A        = 2*vdwtype[jnrA+0];
 570             vdwjidx0B        = 2*vdwtype[jnrB+0];
 571             vdwjidx0C        = 2*vdwtype[jnrC+0];
 572             vdwjidx0D        = 2*vdwtype[jnrD+0];
 573
 574             fjx0             = _mm_setzero_ps();
 575             fjy0             = _mm_setzero_ps();
 576             fjz0             = _mm_setzero_ps();
 577
 578             /**************************
 579              * CALCULATE INTERACTIONS *
 580              **************************/
 581
 582             if (gmx_mm_any_lt(rsq00,rcutoff2))
 583             {
 584
 585             r00              = _mm_mul_ps(rsq00,rinv00);
 586             r00              = _mm_andnot_ps(dummy_mask,r00);
 587
 588             /* Compute parameters for interactions between i and j atoms */
 589             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 590                                          vdwparam+vdwioffset0+vdwjidx0B,
 591                                          vdwparam+vdwioffset0+vdwjidx0C,
 592                                          vdwparam+vdwioffset0+vdwjidx0D,
 593                                          &c6_00,&c12_00);
 594
 595             c6grid_00       = gmx_mm_load_4real_swizzle_ps(vdwgridparam+vdwioffset0+vdwjidx0A,
 596                                                                vdwgridparam+vdwioffset0+vdwjidx0B,
 597                                                                vdwgridparam+vdwioffset0+vdwjidx0C,
 598                                                                vdwgridparam+vdwioffset0+vdwjidx0D);
 599
 600             /* Analytical LJ-PME */
 601             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 602             ewcljrsq         = _mm_mul_ps(ewclj2,rsq00);
 603             ewclj6           = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2));
 604             exponent         = gmx_simd_exp_r(ewcljrsq);
 605             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 606             poly             = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 607             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 608             vvdw6            = _mm_mul_ps(_mm_sub_ps(c6_00,_mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly))),rinvsix);
 609             vvdw12           = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix));
 610             vvdw             = _mm_sub_ps(_mm_mul_ps( _mm_sub_ps(vvdw12 , _mm_mul_ps(c12_00,_mm_mul_ps(sh_vdw_invrcut6,sh_vdw_invrcut6))),one_twelfth),
 611                                _mm_mul_ps( _mm_sub_ps(vvdw6,_mm_add_ps(_mm_mul_ps(c6_00,sh_vdw_invrcut6),_mm_mul_ps(c6grid_00,sh_lj_ewald))),one_sixth));
 612             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 613             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,_mm_sub_ps(vvdw6,_mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6)))),rinvsq00);
 614
 615             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
 616
 617             /* Update potential sum for this i atom from the interaction with this j atom. */
 618             vvdw             = _mm_and_ps(vvdw,cutoff_mask);
 619             vvdw             = _mm_andnot_ps(dummy_mask,vvdw);
 620             vvdwsum          = _mm_add_ps(vvdwsum,vvdw);
 621
 622             fscal            = fvdw;
 623
 624             fscal            = _mm_and_ps(fscal,cutoff_mask);
 625
 626             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 627
 628             /* Calculate temporary vectorial force */
 629             tx               = _mm_mul_ps(fscal,dx00);
 630             ty               = _mm_mul_ps(fscal,dy00);
 631             tz               = _mm_mul_ps(fscal,dz00);
 632
 633             /* Update vectorial force */
 634             fix0             = _mm_add_ps(fix0,tx);
 635             fiy0             = _mm_add_ps(fiy0,ty);
 636             fiz0             = _mm_add_ps(fiz0,tz);
 637
 638             fjx0             = _mm_add_ps(fjx0,tx);
 639             fjy0             = _mm_add_ps(fjy0,ty);
 640             fjz0             = _mm_add_ps(fjz0,tz);
 641
 642             }
 643
 644             /**************************
 645              * CALCULATE INTERACTIONS *
 646              **************************/
 647
 648             if (gmx_mm_any_lt(rsq10,rcutoff2))
 649             {
 650
 651             r10              = _mm_mul_ps(rsq10,rinv10);
 652             r10              = _mm_andnot_ps(dummy_mask,r10);
 653
 654             /* Compute parameters for interactions between i and j atoms */
 655             qq10             = _mm_mul_ps(iq1,jq0);
 656
 657             /* EWALD ELECTROSTATICS */
 658
 659             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 660             ewrt             = _mm_mul_ps(r10,ewtabscale);
 661             ewitab           = _mm_cvttps_epi32(ewrt);
 662             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
 663             ewitab           = _mm_slli_epi32(ewitab,2);
 664             ewtabF           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
 665             ewtabD           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
 666             ewtabV           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
 667             ewtabFn          = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
 668             _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
 669             felec            = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
 670             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 671             velec            = _mm_mul_ps(qq10,_mm_sub_ps(_mm_sub_ps(rinv10,sh_ewald),velec));
 672             felec            = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec));
 673
 674             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
 675
 676             /* Update potential sum for this i atom from the interaction with this j atom. */
 677             velec            = _mm_and_ps(velec,cutoff_mask);
 678             velec            = _mm_andnot_ps(dummy_mask,velec);
 679             velecsum         = _mm_add_ps(velecsum,velec);
 680
 681             fscal            = felec;
 682
 683             fscal            = _mm_and_ps(fscal,cutoff_mask);
 684
 685             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 686
 687             /* Calculate temporary vectorial force */
 688             tx               = _mm_mul_ps(fscal,dx10);
 689             ty               = _mm_mul_ps(fscal,dy10);
 690             tz               = _mm_mul_ps(fscal,dz10);
 691
 692             /* Update vectorial force */
 693             fix1             = _mm_add_ps(fix1,tx);
 694             fiy1             = _mm_add_ps(fiy1,ty);
 695             fiz1             = _mm_add_ps(fiz1,tz);
 696
 697             fjx0             = _mm_add_ps(fjx0,tx);
 698             fjy0             = _mm_add_ps(fjy0,ty);
 699             fjz0             = _mm_add_ps(fjz0,tz);
 700
 701             }
 702
 703             /**************************
 704              * CALCULATE INTERACTIONS *
 705              **************************/
 706
 707             if (gmx_mm_any_lt(rsq20,rcutoff2))
 708             {
 709
 710             r20              = _mm_mul_ps(rsq20,rinv20);
 711             r20              = _mm_andnot_ps(dummy_mask,r20);
 712
 713             /* Compute parameters for interactions between i and j atoms */
 714             qq20             = _mm_mul_ps(iq2,jq0);
 715
 716             /* EWALD ELECTROSTATICS */
 717
 718             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 719             ewrt             = _mm_mul_ps(r20,ewtabscale);
 720             ewitab           = _mm_cvttps_epi32(ewrt);
 721             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
 722             ewitab           = _mm_slli_epi32(ewitab,2);
 723             ewtabF           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
 724             ewtabD           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
 725             ewtabV           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
 726             ewtabFn          = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
 727             _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
 728             felec            = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
 729             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 730             velec            = _mm_mul_ps(qq20,_mm_sub_ps(_mm_sub_ps(rinv20,sh_ewald),velec));
 731             felec            = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec));
 732
 733             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
 734
 735             /* Update potential sum for this i atom from the interaction with this j atom. */
 736             velec            = _mm_and_ps(velec,cutoff_mask);
 737             velec            = _mm_andnot_ps(dummy_mask,velec);
 738             velecsum         = _mm_add_ps(velecsum,velec);
 739
 740             fscal            = felec;
 741
 742             fscal            = _mm_and_ps(fscal,cutoff_mask);
 743
 744             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 745
 746             /* Calculate temporary vectorial force */
 747             tx               = _mm_mul_ps(fscal,dx20);
 748             ty               = _mm_mul_ps(fscal,dy20);
 749             tz               = _mm_mul_ps(fscal,dz20);
 750
 751             /* Update vectorial force */
 752             fix2             = _mm_add_ps(fix2,tx);
 753             fiy2             = _mm_add_ps(fiy2,ty);
 754             fiz2             = _mm_add_ps(fiz2,tz);
 755
 756             fjx0             = _mm_add_ps(fjx0,tx);
 757             fjy0             = _mm_add_ps(fjy0,ty);
 758             fjz0             = _mm_add_ps(fjz0,tz);
 759
 760             }
 761
 762             /**************************
 763              * CALCULATE INTERACTIONS *
 764              **************************/
 765
 766             if (gmx_mm_any_lt(rsq30,rcutoff2))
 767             {
 768
 769             r30              = _mm_mul_ps(rsq30,rinv30);
 770             r30              = _mm_andnot_ps(dummy_mask,r30);
 771
 772             /* Compute parameters for interactions between i and j atoms */
 773             qq30             = _mm_mul_ps(iq3,jq0);
 774
 775             /* EWALD ELECTROSTATICS */
 776
 777             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 778             ewrt             = _mm_mul_ps(r30,ewtabscale);
 779             ewitab           = _mm_cvttps_epi32(ewrt);
 780             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
 781             ewitab           = _mm_slli_epi32(ewitab,2);
 782             ewtabF           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
 783             ewtabD           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
 784             ewtabV           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
 785             ewtabFn          = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
 786             _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
 787             felec            = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
 788             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 789             velec            = _mm_mul_ps(qq30,_mm_sub_ps(_mm_sub_ps(rinv30,sh_ewald),velec));
 790             felec            = _mm_mul_ps(_mm_mul_ps(qq30,rinv30),_mm_sub_ps(rinvsq30,felec));
 791
 792             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
 793
 794             /* Update potential sum for this i atom from the interaction with this j atom. */
 795             velec            = _mm_and_ps(velec,cutoff_mask);
 796             velec            = _mm_andnot_ps(dummy_mask,velec);
 797             velecsum         = _mm_add_ps(velecsum,velec);
 798
 799             fscal            = felec;
 800
 801             fscal            = _mm_and_ps(fscal,cutoff_mask);
 802
 803             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 804
 805             /* Calculate temporary vectorial force */
 806             tx               = _mm_mul_ps(fscal,dx30);
 807             ty               = _mm_mul_ps(fscal,dy30);
 808             tz               = _mm_mul_ps(fscal,dz30);
 809
 810             /* Update vectorial force */
 811             fix3             = _mm_add_ps(fix3,tx);
 812             fiy3             = _mm_add_ps(fiy3,ty);
 813             fiz3             = _mm_add_ps(fiz3,tz);
 814
 815             fjx0             = _mm_add_ps(fjx0,tx);
 816             fjy0             = _mm_add_ps(fjy0,ty);
 817             fjz0             = _mm_add_ps(fjz0,tz);
 818
 819             }
 820
 821             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 822             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 823             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 824             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 825
 826             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 827
 828             /* Inner loop uses 204 flops */
 829         }
 830
 831         /* End of innermost loop */
 832
 833         gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 834                                               f+i_coord_offset,fshift+i_shift_offset);
 835
 836         ggid                        = gid[iidx];
 837         /* Update potential energies */
 838         gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 839         gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 840
 841         /* Increment number of inner iterations */
 842         inneriter                  += j_index_end - j_index_start;
 843
 844         /* Outer loop uses 26 flops */
 845     }
 846
 847     /* Increment number of outer iterations */
 848     outeriter        += nri;
 849
 850     /* Update outer/inner flops */
 851
 852     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*204);
 853 }
 854 /*
 855  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJEwSh_GeomW4P1_F_sse4_1_single
 856  * Electrostatics interaction: Ewald
 857  * VdW interaction:            LJEwald
 858  * Geometry:                   Water4-Particle
 859  * Calculate force/pot:        Force
 860  */
 861 void
 862 nb_kernel_ElecEwSh_VdwLJEwSh_GeomW4P1_F_sse4_1_single
 863                     (t_nblist                    * gmx_restrict       nlist,
 864                      rvec                        * gmx_restrict          xx,
 865                      rvec                        * gmx_restrict          ff,
 866                      t_forcerec                  * gmx_restrict          fr,
 867                      t_mdatoms                   * gmx_restrict     mdatoms,
 868                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 869                      t_nrnb                      * gmx_restrict        nrnb)
 870 {
 871     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 872      * just 0 for non-waters.
 873      * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
 874      * jnr indices corresponding to data put in the four positions in the SIMD register.
 875      */
 876     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 877     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 878     int              jnrA,jnrB,jnrC,jnrD;
 879     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 880     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 881     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 882     real             rcutoff_scalar;
 883     real             *shiftvec,*fshift,*x,*f;
 884     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 885     real             scratch[4*DIM];
 886     __m128           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 887     int              vdwioffset0;
 888     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 889     int              vdwioffset1;
 890     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 891     int              vdwioffset2;
 892     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 893     int              vdwioffset3;
 894     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 895     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 896     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 897     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 898     __m128           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 899     __m128           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 900     __m128           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 901     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
 902     real             *charge;
 903     int              nvdwtype;
 904     __m128           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 905     int              *vdwtype;
 906     real             *vdwparam;
 907     __m128           one_sixth   = _mm_set1_ps(1.0/6.0);
 908     __m128           one_twelfth = _mm_set1_ps(1.0/12.0);
 909     __m128           c6grid_00;
 910     __m128           c6grid_10;
 911     __m128           c6grid_20;
 912     __m128           c6grid_30;
 913     __m128           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 914     real             *vdwgridparam;
 915     __m128           one_half  = _mm_set1_ps(0.5);
 916     __m128           minus_one = _mm_set1_ps(-1.0);
 917     __m128i          ewitab;
 918     __m128           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 919     real             *ewtab;
 920     __m128           dummy_mask,cutoff_mask;
 921     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 922     __m128           one     = _mm_set1_ps(1.0);
 923     __m128           two     = _mm_set1_ps(2.0);
 924     x                = xx[0];
 925     f                = ff[0];
 926
 927     nri              = nlist->nri;
 928     iinr             = nlist->iinr;
 929     jindex           = nlist->jindex;
 930     jjnr             = nlist->jjnr;
 931     shiftidx         = nlist->shift;
 932     gid              = nlist->gid;
 933     shiftvec         = fr->shift_vec[0];
 934     fshift           = fr->fshift[0];
 935     facel            = _mm_set1_ps(fr->epsfac);
 936     charge           = mdatoms->chargeA;
 937     nvdwtype         = fr->ntype;
 938     vdwparam         = fr->nbfp;
 939     vdwtype          = mdatoms->typeA;
 940     vdwgridparam     = fr->ljpme_c6grid;
 941     sh_lj_ewald      = _mm_set1_ps(fr->ic->sh_lj_ewald);
 942     ewclj            = _mm_set1_ps(fr->ewaldcoeff_lj);
 943     ewclj2           = _mm_mul_ps(minus_one,_mm_mul_ps(ewclj,ewclj));
 944
 945     sh_ewald         = _mm_set1_ps(fr->ic->sh_ewald);
 946     ewtab            = fr->ic->tabq_coul_F;
 947     ewtabscale       = _mm_set1_ps(fr->ic->tabq_scale);
 948     ewtabhalfspace   = _mm_set1_ps(0.5/fr->ic->tabq_scale);
 949
 950     /* Setup water-specific parameters */
 951     inr              = nlist->iinr[0];
 952     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
 953     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
 954     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
 955     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 956
 957     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 958     rcutoff_scalar   = fr->rcoulomb;
 959     rcutoff          = _mm_set1_ps(rcutoff_scalar);
 960     rcutoff2         = _mm_mul_ps(rcutoff,rcutoff);
 961
 962     sh_vdw_invrcut6  = _mm_set1_ps(fr->ic->sh_invrc6);
 963     rvdw             = _mm_set1_ps(fr->rvdw);
 964
 965     /* Avoid stupid compiler warnings */
 966     jnrA = jnrB = jnrC = jnrD = 0;
 967     j_coord_offsetA = 0;
 968     j_coord_offsetB = 0;
 969     j_coord_offsetC = 0;
 970     j_coord_offsetD = 0;
 971
 972     outeriter        = 0;
 973     inneriter        = 0;
 974
 975     for(iidx=0;iidx<4*DIM;iidx++)
 976     {
 977         scratch[iidx] = 0.0;
 978     }
 979
 980     /* Start outer loop over neighborlists */
 981     for(iidx=0; iidx<nri; iidx++)
 982     {
 983         /* Load shift vector for this list */
 984         i_shift_offset   = DIM*shiftidx[iidx];
 985
 986         /* Load limits for loop over neighbors */
 987         j_index_start    = jindex[iidx];
 988         j_index_end      = jindex[iidx+1];
 989
 990         /* Get outer coordinate index */
 991         inr              = iinr[iidx];
 992         i_coord_offset   = DIM*inr;
 993
 994         /* Load i particle coords and add shift vector */
 995         gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 996                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 997
 998         fix0             = _mm_setzero_ps();
 999         fiy0             = _mm_setzero_ps();
1000         fiz0             = _mm_setzero_ps();
1001         fix1             = _mm_setzero_ps();
1002         fiy1             = _mm_setzero_ps();
1003         fiz1             = _mm_setzero_ps();
1004         fix2             = _mm_setzero_ps();
1005         fiy2             = _mm_setzero_ps();
1006         fiz2             = _mm_setzero_ps();
1007         fix3             = _mm_setzero_ps();
1008         fiy3             = _mm_setzero_ps();
1009         fiz3             = _mm_setzero_ps();
1010
1011         /* Start inner kernel loop */
1012         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1013         {
1014
1015             /* Get j neighbor index, and coordinate index */
1016             jnrA             = jjnr[jidx];
1017             jnrB             = jjnr[jidx+1];
1018             jnrC             = jjnr[jidx+2];
1019             jnrD             = jjnr[jidx+3];
1020             j_coord_offsetA  = DIM*jnrA;
1021             j_coord_offsetB  = DIM*jnrB;
1022             j_coord_offsetC  = DIM*jnrC;
1023             j_coord_offsetD  = DIM*jnrD;
1024
1025             /* load j atom coordinates */
1026             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1027                                               x+j_coord_offsetC,x+j_coord_offsetD,
1028                                               &jx0,&jy0,&jz0);
1029
1030             /* Calculate displacement vector */
1031             dx00             = _mm_sub_ps(ix0,jx0);
1032             dy00             = _mm_sub_ps(iy0,jy0);
1033             dz00             = _mm_sub_ps(iz0,jz0);
1034             dx10             = _mm_sub_ps(ix1,jx0);
1035             dy10             = _mm_sub_ps(iy1,jy0);
1036             dz10             = _mm_sub_ps(iz1,jz0);
1037             dx20             = _mm_sub_ps(ix2,jx0);
1038             dy20             = _mm_sub_ps(iy2,jy0);
1039             dz20             = _mm_sub_ps(iz2,jz0);
1040             dx30             = _mm_sub_ps(ix3,jx0);
1041             dy30             = _mm_sub_ps(iy3,jy0);
1042             dz30             = _mm_sub_ps(iz3,jz0);
1043
1044             /* Calculate squared distance and things based on it */
1045             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
1046             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
1047             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
1048             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
1049
1050             rinv00           = gmx_mm_invsqrt_ps(rsq00);
1051             rinv10           = gmx_mm_invsqrt_ps(rsq10);
1052             rinv20           = gmx_mm_invsqrt_ps(rsq20);
1053             rinv30           = gmx_mm_invsqrt_ps(rsq30);
1054
1055             rinvsq00         = _mm_mul_ps(rinv00,rinv00);
1056             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
1057             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
1058             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
1059
1060             /* Load parameters for j particles */
1061             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1062                                                               charge+jnrC+0,charge+jnrD+0);
1063             vdwjidx0A        = 2*vdwtype[jnrA+0];
1064             vdwjidx0B        = 2*vdwtype[jnrB+0];
1065             vdwjidx0C        = 2*vdwtype[jnrC+0];
1066             vdwjidx0D        = 2*vdwtype[jnrD+0];
1067
1068             fjx0             = _mm_setzero_ps();
1069             fjy0             = _mm_setzero_ps();
1070             fjz0             = _mm_setzero_ps();
1071
1072             /**************************
1073              * CALCULATE INTERACTIONS *
1074              **************************/
1075
1076             if (gmx_mm_any_lt(rsq00,rcutoff2))
1077             {
1078
1079             r00              = _mm_mul_ps(rsq00,rinv00);
1080
1081             /* Compute parameters for interactions between i and j atoms */
1082             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
1083                                          vdwparam+vdwioffset0+vdwjidx0B,
1084                                          vdwparam+vdwioffset0+vdwjidx0C,
1085                                          vdwparam+vdwioffset0+vdwjidx0D,
1086                                          &c6_00,&c12_00);
1087
1088             c6grid_00       = gmx_mm_load_4real_swizzle_ps(vdwgridparam+vdwioffset0+vdwjidx0A,
1089                                                                vdwgridparam+vdwioffset0+vdwjidx0B,
1090                                                                vdwgridparam+vdwioffset0+vdwjidx0C,
1091                                                                vdwgridparam+vdwioffset0+vdwjidx0D);
1092
1093             /* Analytical LJ-PME */
1094             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1095             ewcljrsq         = _mm_mul_ps(ewclj2,rsq00);
1096             ewclj6           = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2));
1097             exponent         = gmx_simd_exp_r(ewcljrsq);
1098             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1099             poly             = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half)));
1100             /* f6A = 6 * C6grid * (1 - poly) */
1101             f6A              = _mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly));
1102             /* f6B = C6grid * exponent * beta^6 */
1103             f6B              = _mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6));
1104             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
1105             fvdw              = _mm_mul_ps(_mm_add_ps(_mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_00,rinvsix),_mm_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
1106
1107             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
1108
1109             fscal            = fvdw;
1110
1111             fscal            = _mm_and_ps(fscal,cutoff_mask);
1112
1113             /* Calculate temporary vectorial force */
1114             tx               = _mm_mul_ps(fscal,dx00);
1115             ty               = _mm_mul_ps(fscal,dy00);
1116             tz               = _mm_mul_ps(fscal,dz00);
1117
1118             /* Update vectorial force */
1119             fix0             = _mm_add_ps(fix0,tx);
1120             fiy0             = _mm_add_ps(fiy0,ty);
1121             fiz0             = _mm_add_ps(fiz0,tz);
1122
1123             fjx0             = _mm_add_ps(fjx0,tx);
1124             fjy0             = _mm_add_ps(fjy0,ty);
1125             fjz0             = _mm_add_ps(fjz0,tz);
1126
1127             }
1128
1129             /**************************
1130              * CALCULATE INTERACTIONS *
1131              **************************/
1132
1133             if (gmx_mm_any_lt(rsq10,rcutoff2))
1134             {
1135
1136             r10              = _mm_mul_ps(rsq10,rinv10);
1137
1138             /* Compute parameters for interactions between i and j atoms */
1139             qq10             = _mm_mul_ps(iq1,jq0);
1140
1141             /* EWALD ELECTROSTATICS */
1142
1143             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1144             ewrt             = _mm_mul_ps(r10,ewtabscale);
1145             ewitab           = _mm_cvttps_epi32(ewrt);
1146             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
1147             gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0),ewtab + gmx_mm_extract_epi32(ewitab,1),
1148                                          ewtab + gmx_mm_extract_epi32(ewitab,2),ewtab + gmx_mm_extract_epi32(ewitab,3),
1149                                          &ewtabF,&ewtabFn);
1150             felec            = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1151             felec            = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec));
1152
1153             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
1154
1155             fscal            = felec;
1156
1157             fscal            = _mm_and_ps(fscal,cutoff_mask);
1158
1159             /* Calculate temporary vectorial force */
1160             tx               = _mm_mul_ps(fscal,dx10);
1161             ty               = _mm_mul_ps(fscal,dy10);
1162             tz               = _mm_mul_ps(fscal,dz10);
1163
1164             /* Update vectorial force */
1165             fix1             = _mm_add_ps(fix1,tx);
1166             fiy1             = _mm_add_ps(fiy1,ty);
1167             fiz1             = _mm_add_ps(fiz1,tz);
1168
1169             fjx0             = _mm_add_ps(fjx0,tx);
1170             fjy0             = _mm_add_ps(fjy0,ty);
1171             fjz0             = _mm_add_ps(fjz0,tz);
1172
1173             }
1174
1175             /**************************
1176              * CALCULATE INTERACTIONS *
1177              **************************/
1178
1179             if (gmx_mm_any_lt(rsq20,rcutoff2))
1180             {
1181
1182             r20              = _mm_mul_ps(rsq20,rinv20);
1183
1184             /* Compute parameters for interactions between i and j atoms */
1185             qq20             = _mm_mul_ps(iq2,jq0);
1186
1187             /* EWALD ELECTROSTATICS */
1188
1189             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1190             ewrt             = _mm_mul_ps(r20,ewtabscale);
1191             ewitab           = _mm_cvttps_epi32(ewrt);
1192             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
1193             gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0),ewtab + gmx_mm_extract_epi32(ewitab,1),
1194                                          ewtab + gmx_mm_extract_epi32(ewitab,2),ewtab + gmx_mm_extract_epi32(ewitab,3),
1195                                          &ewtabF,&ewtabFn);
1196             felec            = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1197             felec            = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec));
1198
1199             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
1200
1201             fscal            = felec;
1202
1203             fscal            = _mm_and_ps(fscal,cutoff_mask);
1204
1205             /* Calculate temporary vectorial force */
1206             tx               = _mm_mul_ps(fscal,dx20);
1207             ty               = _mm_mul_ps(fscal,dy20);
1208             tz               = _mm_mul_ps(fscal,dz20);
1209
1210             /* Update vectorial force */
1211             fix2             = _mm_add_ps(fix2,tx);
1212             fiy2             = _mm_add_ps(fiy2,ty);
1213             fiz2             = _mm_add_ps(fiz2,tz);
1214
1215             fjx0             = _mm_add_ps(fjx0,tx);
1216             fjy0             = _mm_add_ps(fjy0,ty);
1217             fjz0             = _mm_add_ps(fjz0,tz);
1218
1219             }
1220
1221             /**************************
1222              * CALCULATE INTERACTIONS *
1223              **************************/
1224
1225             if (gmx_mm_any_lt(rsq30,rcutoff2))
1226             {
1227
1228             r30              = _mm_mul_ps(rsq30,rinv30);
1229
1230             /* Compute parameters for interactions between i and j atoms */
1231             qq30             = _mm_mul_ps(iq3,jq0);
1232
1233             /* EWALD ELECTROSTATICS */
1234
1235             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1236             ewrt             = _mm_mul_ps(r30,ewtabscale);
1237             ewitab           = _mm_cvttps_epi32(ewrt);
1238             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
1239             gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0),ewtab + gmx_mm_extract_epi32(ewitab,1),
1240                                          ewtab + gmx_mm_extract_epi32(ewitab,2),ewtab + gmx_mm_extract_epi32(ewitab,3),
1241                                          &ewtabF,&ewtabFn);
1242             felec            = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1243             felec            = _mm_mul_ps(_mm_mul_ps(qq30,rinv30),_mm_sub_ps(rinvsq30,felec));
1244
1245             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
1246
1247             fscal            = felec;
1248
1249             fscal            = _mm_and_ps(fscal,cutoff_mask);
1250
1251             /* Calculate temporary vectorial force */
1252             tx               = _mm_mul_ps(fscal,dx30);
1253             ty               = _mm_mul_ps(fscal,dy30);
1254             tz               = _mm_mul_ps(fscal,dz30);
1255
1256             /* Update vectorial force */
1257             fix3             = _mm_add_ps(fix3,tx);
1258             fiy3             = _mm_add_ps(fiy3,ty);
1259             fiz3             = _mm_add_ps(fiz3,tz);
1260
1261             fjx0             = _mm_add_ps(fjx0,tx);
1262             fjy0             = _mm_add_ps(fjy0,ty);
1263             fjz0             = _mm_add_ps(fjz0,tz);
1264
1265             }
1266
1267             fjptrA             = f+j_coord_offsetA;
1268             fjptrB             = f+j_coord_offsetB;
1269             fjptrC             = f+j_coord_offsetC;
1270             fjptrD             = f+j_coord_offsetD;
1271
1272             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1273
1274             /* Inner loop uses 166 flops */
1275         }
1276
1277         if(jidx<j_index_end)
1278         {
1279
1280             /* Get j neighbor index, and coordinate index */
1281             jnrlistA         = jjnr[jidx];
1282             jnrlistB         = jjnr[jidx+1];
1283             jnrlistC         = jjnr[jidx+2];
1284             jnrlistD         = jjnr[jidx+3];
1285             /* Sign of each element will be negative for non-real atoms.
1286              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1287              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1288              */
1289             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1290             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1291             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1292             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1293             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1294             j_coord_offsetA  = DIM*jnrA;
1295             j_coord_offsetB  = DIM*jnrB;
1296             j_coord_offsetC  = DIM*jnrC;
1297             j_coord_offsetD  = DIM*jnrD;
1298
1299             /* load j atom coordinates */
1300             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1301                                               x+j_coord_offsetC,x+j_coord_offsetD,
1302                                               &jx0,&jy0,&jz0);
1303
1304             /* Calculate displacement vector */
1305             dx00             = _mm_sub_ps(ix0,jx0);
1306             dy00             = _mm_sub_ps(iy0,jy0);
1307             dz00             = _mm_sub_ps(iz0,jz0);
1308             dx10             = _mm_sub_ps(ix1,jx0);
1309             dy10             = _mm_sub_ps(iy1,jy0);
1310             dz10             = _mm_sub_ps(iz1,jz0);
1311             dx20             = _mm_sub_ps(ix2,jx0);
1312             dy20             = _mm_sub_ps(iy2,jy0);
1313             dz20             = _mm_sub_ps(iz2,jz0);
1314             dx30             = _mm_sub_ps(ix3,jx0);
1315             dy30             = _mm_sub_ps(iy3,jy0);
1316             dz30             = _mm_sub_ps(iz3,jz0);
1317
1318             /* Calculate squared distance and things based on it */
1319             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
1320             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
1321             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
1322             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
1323
1324             rinv00           = gmx_mm_invsqrt_ps(rsq00);
1325             rinv10           = gmx_mm_invsqrt_ps(rsq10);
1326             rinv20           = gmx_mm_invsqrt_ps(rsq20);
1327             rinv30           = gmx_mm_invsqrt_ps(rsq30);
1328
1329             rinvsq00         = _mm_mul_ps(rinv00,rinv00);
1330             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
1331             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
1332             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
1333
1334             /* Load parameters for j particles */
1335             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1336                                                               charge+jnrC+0,charge+jnrD+0);
1337             vdwjidx0A        = 2*vdwtype[jnrA+0];
1338             vdwjidx0B        = 2*vdwtype[jnrB+0];
1339             vdwjidx0C        = 2*vdwtype[jnrC+0];
1340             vdwjidx0D        = 2*vdwtype[jnrD+0];
1341
1342             fjx0             = _mm_setzero_ps();
1343             fjy0             = _mm_setzero_ps();
1344             fjz0             = _mm_setzero_ps();
1345
1346             /**************************
1347              * CALCULATE INTERACTIONS *
1348              **************************/
1349
1350             if (gmx_mm_any_lt(rsq00,rcutoff2))
1351             {
1352
1353             r00              = _mm_mul_ps(rsq00,rinv00);
1354             r00              = _mm_andnot_ps(dummy_mask,r00);
1355
1356             /* Compute parameters for interactions between i and j atoms */
1357             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
1358                                          vdwparam+vdwioffset0+vdwjidx0B,
1359                                          vdwparam+vdwioffset0+vdwjidx0C,
1360                                          vdwparam+vdwioffset0+vdwjidx0D,
1361                                          &c6_00,&c12_00);
1362
1363             c6grid_00       = gmx_mm_load_4real_swizzle_ps(vdwgridparam+vdwioffset0+vdwjidx0A,
1364                                                                vdwgridparam+vdwioffset0+vdwjidx0B,
1365                                                                vdwgridparam+vdwioffset0+vdwjidx0C,
1366                                                                vdwgridparam+vdwioffset0+vdwjidx0D);
1367
1368             /* Analytical LJ-PME */
1369             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1370             ewcljrsq         = _mm_mul_ps(ewclj2,rsq00);
1371             ewclj6           = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2));
1372             exponent         = gmx_simd_exp_r(ewcljrsq);
1373             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1374             poly             = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half)));
1375             /* f6A = 6 * C6grid * (1 - poly) */
1376             f6A              = _mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly));
1377             /* f6B = C6grid * exponent * beta^6 */
1378             f6B              = _mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6));
1379             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
1380             fvdw              = _mm_mul_ps(_mm_add_ps(_mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_00,rinvsix),_mm_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
1381
1382             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
1383
1384             fscal            = fvdw;
1385
1386             fscal            = _mm_and_ps(fscal,cutoff_mask);
1387
1388             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1389
1390             /* Calculate temporary vectorial force */
1391             tx               = _mm_mul_ps(fscal,dx00);
1392             ty               = _mm_mul_ps(fscal,dy00);
1393             tz               = _mm_mul_ps(fscal,dz00);
1394
1395             /* Update vectorial force */
1396             fix0             = _mm_add_ps(fix0,tx);
1397             fiy0             = _mm_add_ps(fiy0,ty);
1398             fiz0             = _mm_add_ps(fiz0,tz);
1399
1400             fjx0             = _mm_add_ps(fjx0,tx);
1401             fjy0             = _mm_add_ps(fjy0,ty);
1402             fjz0             = _mm_add_ps(fjz0,tz);
1403
1404             }
1405
1406             /**************************
1407              * CALCULATE INTERACTIONS *
1408              **************************/
1409
1410             if (gmx_mm_any_lt(rsq10,rcutoff2))
1411             {
1412
1413             r10              = _mm_mul_ps(rsq10,rinv10);
1414             r10              = _mm_andnot_ps(dummy_mask,r10);
1415
1416             /* Compute parameters for interactions between i and j atoms */
1417             qq10             = _mm_mul_ps(iq1,jq0);
1418
1419             /* EWALD ELECTROSTATICS */
1420
1421             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1422             ewrt             = _mm_mul_ps(r10,ewtabscale);
1423             ewitab           = _mm_cvttps_epi32(ewrt);
1424             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
1425             gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0),ewtab + gmx_mm_extract_epi32(ewitab,1),
1426                                          ewtab + gmx_mm_extract_epi32(ewitab,2),ewtab + gmx_mm_extract_epi32(ewitab,3),
1427                                          &ewtabF,&ewtabFn);
1428             felec            = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1429             felec            = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec));
1430
1431             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
1432
1433             fscal            = felec;
1434
1435             fscal            = _mm_and_ps(fscal,cutoff_mask);
1436
1437             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1438
1439             /* Calculate temporary vectorial force */
1440             tx               = _mm_mul_ps(fscal,dx10);
1441             ty               = _mm_mul_ps(fscal,dy10);
1442             tz               = _mm_mul_ps(fscal,dz10);
1443
1444             /* Update vectorial force */
1445             fix1             = _mm_add_ps(fix1,tx);
1446             fiy1             = _mm_add_ps(fiy1,ty);
1447             fiz1             = _mm_add_ps(fiz1,tz);
1448
1449             fjx0             = _mm_add_ps(fjx0,tx);
1450             fjy0             = _mm_add_ps(fjy0,ty);
1451             fjz0             = _mm_add_ps(fjz0,tz);
1452
1453             }
1454
1455             /**************************
1456              * CALCULATE INTERACTIONS *
1457              **************************/
1458
1459             if (gmx_mm_any_lt(rsq20,rcutoff2))
1460             {
1461
1462             r20              = _mm_mul_ps(rsq20,rinv20);
1463             r20              = _mm_andnot_ps(dummy_mask,r20);
1464
1465             /* Compute parameters for interactions between i and j atoms */
1466             qq20             = _mm_mul_ps(iq2,jq0);
1467
1468             /* EWALD ELECTROSTATICS */
1469
1470             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1471             ewrt             = _mm_mul_ps(r20,ewtabscale);
1472             ewitab           = _mm_cvttps_epi32(ewrt);
1473             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
1474             gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0),ewtab + gmx_mm_extract_epi32(ewitab,1),
1475                                          ewtab + gmx_mm_extract_epi32(ewitab,2),ewtab + gmx_mm_extract_epi32(ewitab,3),
1476                                          &ewtabF,&ewtabFn);
1477             felec            = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1478             felec            = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec));
1479
1480             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
1481
1482             fscal            = felec;
1483
1484             fscal            = _mm_and_ps(fscal,cutoff_mask);
1485
1486             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1487
1488             /* Calculate temporary vectorial force */
1489             tx               = _mm_mul_ps(fscal,dx20);
1490             ty               = _mm_mul_ps(fscal,dy20);
1491             tz               = _mm_mul_ps(fscal,dz20);
1492
1493             /* Update vectorial force */
1494             fix2             = _mm_add_ps(fix2,tx);
1495             fiy2             = _mm_add_ps(fiy2,ty);
1496             fiz2             = _mm_add_ps(fiz2,tz);
1497
1498             fjx0             = _mm_add_ps(fjx0,tx);
1499             fjy0             = _mm_add_ps(fjy0,ty);
1500             fjz0             = _mm_add_ps(fjz0,tz);
1501
1502             }
1503
1504             /**************************
1505              * CALCULATE INTERACTIONS *
1506              **************************/
1507
1508             if (gmx_mm_any_lt(rsq30,rcutoff2))
1509             {
1510
1511             r30              = _mm_mul_ps(rsq30,rinv30);
1512             r30              = _mm_andnot_ps(dummy_mask,r30);
1513
1514             /* Compute parameters for interactions between i and j atoms */
1515             qq30             = _mm_mul_ps(iq3,jq0);
1516
1517             /* EWALD ELECTROSTATICS */
1518
1519             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1520             ewrt             = _mm_mul_ps(r30,ewtabscale);
1521             ewitab           = _mm_cvttps_epi32(ewrt);
1522             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
1523             gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0),ewtab + gmx_mm_extract_epi32(ewitab,1),
1524                                          ewtab + gmx_mm_extract_epi32(ewitab,2),ewtab + gmx_mm_extract_epi32(ewitab,3),
1525                                          &ewtabF,&ewtabFn);
1526             felec            = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1527             felec            = _mm_mul_ps(_mm_mul_ps(qq30,rinv30),_mm_sub_ps(rinvsq30,felec));
1528
1529             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
1530
1531             fscal            = felec;
1532
1533             fscal            = _mm_and_ps(fscal,cutoff_mask);
1534
1535             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1536
1537             /* Calculate temporary vectorial force */
1538             tx               = _mm_mul_ps(fscal,dx30);
1539             ty               = _mm_mul_ps(fscal,dy30);
1540             tz               = _mm_mul_ps(fscal,dz30);
1541
1542             /* Update vectorial force */
1543             fix3             = _mm_add_ps(fix3,tx);
1544             fiy3             = _mm_add_ps(fiy3,ty);
1545             fiz3             = _mm_add_ps(fiz3,tz);
1546
1547             fjx0             = _mm_add_ps(fjx0,tx);
1548             fjy0             = _mm_add_ps(fjy0,ty);
1549             fjz0             = _mm_add_ps(fjz0,tz);
1550
1551             }
1552
1553             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1554             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1555             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1556             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1557
1558             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1559
1560             /* Inner loop uses 170 flops */
1561         }
1562
1563         /* End of innermost loop */
1564
1565         gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1566                                               f+i_coord_offset,fshift+i_shift_offset);
1567
1568         /* Increment number of inner iterations */
1569         inneriter                  += j_index_end - j_index_start;
1570
1571         /* Outer loop uses 24 flops */
1572     }
1573
1574     /* Increment number of outer iterations */
1575     outeriter        += nri;
1576
1577     /* Update outer/inner flops */
1578
1579     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*170);
1580 }