src/gromacs/gmxlib/nonbonded/nb_kernel_avx_128_fma_single/nb_kernel_ElecEwSh_VdwLJSh_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, 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/legacyheaders/types/simple.h"
  46 #include "gromacs/math/vec.h"
  47 #include "gromacs/legacyheaders/nrnb.h"
  48
  49 #include "gromacs/simd/math_x86_avx_128_fma_single.h"
  50 #include "kernelutil_x86_avx_128_fma_single.h"
  51
  52 /*
  53  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_VF_avx_128_fma_single
  54  * Electrostatics interaction: Ewald
  55  * VdW interaction:            LennardJones
  56  * Geometry:                   Water4-Particle
  57  * Calculate force/pot:        PotentialAndForce
  58  */
  59 void
  60 nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_VF_avx_128_fma_single
  61                     (t_nblist                    * gmx_restrict       nlist,
  62                      rvec                        * gmx_restrict          xx,
  63                      rvec                        * gmx_restrict          ff,
  64                      t_forcerec                  * gmx_restrict          fr,
  65                      t_mdatoms                   * gmx_restrict     mdatoms,
  66                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  67                      t_nrnb                      * gmx_restrict        nrnb)
  68 {
  69     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  70      * just 0 for non-waters.
  71      * Suffixes A,B,C,D refer to j loop unrolling done with AVX_128, e.g. for the four different
  72      * jnr indices corresponding to data put in the four positions in the SIMD register.
  73      */
  74     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  75     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  76     int              jnrA,jnrB,jnrC,jnrD;
  77     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  78     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  79     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  80     real             rcutoff_scalar;
  81     real             *shiftvec,*fshift,*x,*f;
  82     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  83     real             scratch[4*DIM];
  84     __m128           fscal,rcutoff,rcutoff2,jidxall;
  85     int              vdwioffset0;
  86     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  87     int              vdwioffset1;
  88     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  89     int              vdwioffset2;
  90     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  91     int              vdwioffset3;
  92     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  93     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  94     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  95     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  96     __m128           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  97     __m128           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  98     __m128           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  99     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
 100     real             *charge;
 101     int              nvdwtype;
 102     __m128           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 103     int              *vdwtype;
 104     real             *vdwparam;
 105     __m128           one_sixth   = _mm_set1_ps(1.0/6.0);
 106     __m128           one_twelfth = _mm_set1_ps(1.0/12.0);
 107     __m128i          ewitab;
 108     __m128           ewtabscale,eweps,twoeweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 109     __m128           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 110     real             *ewtab;
 111     __m128           dummy_mask,cutoff_mask;
 112     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 113     __m128           one     = _mm_set1_ps(1.0);
 114     __m128           two     = _mm_set1_ps(2.0);
 115     x                = xx[0];
 116     f                = ff[0];
 117
 118     nri              = nlist->nri;
 119     iinr             = nlist->iinr;
 120     jindex           = nlist->jindex;
 121     jjnr             = nlist->jjnr;
 122     shiftidx         = nlist->shift;
 123     gid              = nlist->gid;
 124     shiftvec         = fr->shift_vec[0];
 125     fshift           = fr->fshift[0];
 126     facel            = _mm_set1_ps(fr->epsfac);
 127     charge           = mdatoms->chargeA;
 128     nvdwtype         = fr->ntype;
 129     vdwparam         = fr->nbfp;
 130     vdwtype          = mdatoms->typeA;
 131
 132     sh_ewald         = _mm_set1_ps(fr->ic->sh_ewald);
 133     beta             = _mm_set1_ps(fr->ic->ewaldcoeff_q);
 134     beta2            = _mm_mul_ps(beta,beta);
 135     beta3            = _mm_mul_ps(beta,beta2);
 136     ewtab            = fr->ic->tabq_coul_FDV0;
 137     ewtabscale       = _mm_set1_ps(fr->ic->tabq_scale);
 138     ewtabhalfspace   = _mm_set1_ps(0.5/fr->ic->tabq_scale);
 139
 140     /* Setup water-specific parameters */
 141     inr              = nlist->iinr[0];
 142     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
 143     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
 144     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
 145     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 146
 147     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 148     rcutoff_scalar   = fr->rcoulomb;
 149     rcutoff          = _mm_set1_ps(rcutoff_scalar);
 150     rcutoff2         = _mm_mul_ps(rcutoff,rcutoff);
 151
 152     sh_vdw_invrcut6  = _mm_set1_ps(fr->ic->sh_invrc6);
 153     rvdw             = _mm_set1_ps(fr->rvdw);
 154
 155     /* Avoid stupid compiler warnings */
 156     jnrA = jnrB = jnrC = jnrD = 0;
 157     j_coord_offsetA = 0;
 158     j_coord_offsetB = 0;
 159     j_coord_offsetC = 0;
 160     j_coord_offsetD = 0;
 161
 162     outeriter        = 0;
 163     inneriter        = 0;
 164
 165     for(iidx=0;iidx<4*DIM;iidx++)
 166     {
 167         scratch[iidx] = 0.0;
 168     }
 169
 170     /* Start outer loop over neighborlists */
 171     for(iidx=0; iidx<nri; iidx++)
 172     {
 173         /* Load shift vector for this list */
 174         i_shift_offset   = DIM*shiftidx[iidx];
 175
 176         /* Load limits for loop over neighbors */
 177         j_index_start    = jindex[iidx];
 178         j_index_end      = jindex[iidx+1];
 179
 180         /* Get outer coordinate index */
 181         inr              = iinr[iidx];
 182         i_coord_offset   = DIM*inr;
 183
 184         /* Load i particle coords and add shift vector */
 185         gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 186                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 187
 188         fix0             = _mm_setzero_ps();
 189         fiy0             = _mm_setzero_ps();
 190         fiz0             = _mm_setzero_ps();
 191         fix1             = _mm_setzero_ps();
 192         fiy1             = _mm_setzero_ps();
 193         fiz1             = _mm_setzero_ps();
 194         fix2             = _mm_setzero_ps();
 195         fiy2             = _mm_setzero_ps();
 196         fiz2             = _mm_setzero_ps();
 197         fix3             = _mm_setzero_ps();
 198         fiy3             = _mm_setzero_ps();
 199         fiz3             = _mm_setzero_ps();
 200
 201         /* Reset potential sums */
 202         velecsum         = _mm_setzero_ps();
 203         vvdwsum          = _mm_setzero_ps();
 204
 205         /* Start inner kernel loop */
 206         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 207         {
 208
 209             /* Get j neighbor index, and coordinate index */
 210             jnrA             = jjnr[jidx];
 211             jnrB             = jjnr[jidx+1];
 212             jnrC             = jjnr[jidx+2];
 213             jnrD             = jjnr[jidx+3];
 214             j_coord_offsetA  = DIM*jnrA;
 215             j_coord_offsetB  = DIM*jnrB;
 216             j_coord_offsetC  = DIM*jnrC;
 217             j_coord_offsetD  = DIM*jnrD;
 218
 219             /* load j atom coordinates */
 220             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 221                                               x+j_coord_offsetC,x+j_coord_offsetD,
 222                                               &jx0,&jy0,&jz0);
 223
 224             /* Calculate displacement vector */
 225             dx00             = _mm_sub_ps(ix0,jx0);
 226             dy00             = _mm_sub_ps(iy0,jy0);
 227             dz00             = _mm_sub_ps(iz0,jz0);
 228             dx10             = _mm_sub_ps(ix1,jx0);
 229             dy10             = _mm_sub_ps(iy1,jy0);
 230             dz10             = _mm_sub_ps(iz1,jz0);
 231             dx20             = _mm_sub_ps(ix2,jx0);
 232             dy20             = _mm_sub_ps(iy2,jy0);
 233             dz20             = _mm_sub_ps(iz2,jz0);
 234             dx30             = _mm_sub_ps(ix3,jx0);
 235             dy30             = _mm_sub_ps(iy3,jy0);
 236             dz30             = _mm_sub_ps(iz3,jz0);
 237
 238             /* Calculate squared distance and things based on it */
 239             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 240             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 241             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 242             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 243
 244             rinv10           = gmx_mm_invsqrt_ps(rsq10);
 245             rinv20           = gmx_mm_invsqrt_ps(rsq20);
 246             rinv30           = gmx_mm_invsqrt_ps(rsq30);
 247
 248             rinvsq00         = gmx_mm_inv_ps(rsq00);
 249             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 250             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 251             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 252
 253             /* Load parameters for j particles */
 254             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 255                                                               charge+jnrC+0,charge+jnrD+0);
 256             vdwjidx0A        = 2*vdwtype[jnrA+0];
 257             vdwjidx0B        = 2*vdwtype[jnrB+0];
 258             vdwjidx0C        = 2*vdwtype[jnrC+0];
 259             vdwjidx0D        = 2*vdwtype[jnrD+0];
 260
 261             fjx0             = _mm_setzero_ps();
 262             fjy0             = _mm_setzero_ps();
 263             fjz0             = _mm_setzero_ps();
 264
 265             /**************************
 266              * CALCULATE INTERACTIONS *
 267              **************************/
 268
 269             if (gmx_mm_any_lt(rsq00,rcutoff2))
 270             {
 271
 272             /* Compute parameters for interactions between i and j atoms */
 273             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 274                                          vdwparam+vdwioffset0+vdwjidx0B,
 275                                          vdwparam+vdwioffset0+vdwjidx0C,
 276                                          vdwparam+vdwioffset0+vdwjidx0D,
 277                                          &c6_00,&c12_00);
 278
 279             /* LENNARD-JONES DISPERSION/REPULSION */
 280
 281             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 282             vvdw6            = _mm_mul_ps(c6_00,rinvsix);
 283             vvdw12           = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix));
 284             vvdw             = _mm_msub_ps(_mm_nmacc_ps(c12_00,_mm_mul_ps(sh_vdw_invrcut6,sh_vdw_invrcut6),vvdw12),one_twelfth,
 285                                           _mm_mul_ps( _mm_nmacc_ps(c6_00,sh_vdw_invrcut6,vvdw6),one_sixth));
 286             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00);
 287
 288             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
 289
 290             /* Update potential sum for this i atom from the interaction with this j atom. */
 291             vvdw             = _mm_and_ps(vvdw,cutoff_mask);
 292             vvdwsum          = _mm_add_ps(vvdwsum,vvdw);
 293
 294             fscal            = fvdw;
 295
 296             fscal            = _mm_and_ps(fscal,cutoff_mask);
 297
 298              /* Update vectorial force */
 299             fix0             = _mm_macc_ps(dx00,fscal,fix0);
 300             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
 301             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
 302
 303             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
 304             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
 305             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
 306
 307             }
 308
 309             /**************************
 310              * CALCULATE INTERACTIONS *
 311              **************************/
 312
 313             if (gmx_mm_any_lt(rsq10,rcutoff2))
 314             {
 315
 316             r10              = _mm_mul_ps(rsq10,rinv10);
 317
 318             /* Compute parameters for interactions between i and j atoms */
 319             qq10             = _mm_mul_ps(iq1,jq0);
 320
 321             /* EWALD ELECTROSTATICS */
 322
 323             /* Analytical PME correction */
 324             zeta2            = _mm_mul_ps(beta2,rsq10);
 325             rinv3            = _mm_mul_ps(rinvsq10,rinv10);
 326             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
 327             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
 328             felec            = _mm_mul_ps(qq10,felec);
 329             pmecorrV         = gmx_mm_pmecorrV_ps(zeta2);
 330             velec            = _mm_nmacc_ps(pmecorrV,beta,_mm_sub_ps(rinv10,sh_ewald));
 331             velec            = _mm_mul_ps(qq10,velec);
 332
 333             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
 334
 335             /* Update potential sum for this i atom from the interaction with this j atom. */
 336             velec            = _mm_and_ps(velec,cutoff_mask);
 337             velecsum         = _mm_add_ps(velecsum,velec);
 338
 339             fscal            = felec;
 340
 341             fscal            = _mm_and_ps(fscal,cutoff_mask);
 342
 343              /* Update vectorial force */
 344             fix1             = _mm_macc_ps(dx10,fscal,fix1);
 345             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
 346             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
 347
 348             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
 349             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
 350             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
 351
 352             }
 353
 354             /**************************
 355              * CALCULATE INTERACTIONS *
 356              **************************/
 357
 358             if (gmx_mm_any_lt(rsq20,rcutoff2))
 359             {
 360
 361             r20              = _mm_mul_ps(rsq20,rinv20);
 362
 363             /* Compute parameters for interactions between i and j atoms */
 364             qq20             = _mm_mul_ps(iq2,jq0);
 365
 366             /* EWALD ELECTROSTATICS */
 367
 368             /* Analytical PME correction */
 369             zeta2            = _mm_mul_ps(beta2,rsq20);
 370             rinv3            = _mm_mul_ps(rinvsq20,rinv20);
 371             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
 372             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
 373             felec            = _mm_mul_ps(qq20,felec);
 374             pmecorrV         = gmx_mm_pmecorrV_ps(zeta2);
 375             velec            = _mm_nmacc_ps(pmecorrV,beta,_mm_sub_ps(rinv20,sh_ewald));
 376             velec            = _mm_mul_ps(qq20,velec);
 377
 378             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
 379
 380             /* Update potential sum for this i atom from the interaction with this j atom. */
 381             velec            = _mm_and_ps(velec,cutoff_mask);
 382             velecsum         = _mm_add_ps(velecsum,velec);
 383
 384             fscal            = felec;
 385
 386             fscal            = _mm_and_ps(fscal,cutoff_mask);
 387
 388              /* Update vectorial force */
 389             fix2             = _mm_macc_ps(dx20,fscal,fix2);
 390             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
 391             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
 392
 393             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
 394             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
 395             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
 396
 397             }
 398
 399             /**************************
 400              * CALCULATE INTERACTIONS *
 401              **************************/
 402
 403             if (gmx_mm_any_lt(rsq30,rcutoff2))
 404             {
 405
 406             r30              = _mm_mul_ps(rsq30,rinv30);
 407
 408             /* Compute parameters for interactions between i and j atoms */
 409             qq30             = _mm_mul_ps(iq3,jq0);
 410
 411             /* EWALD ELECTROSTATICS */
 412
 413             /* Analytical PME correction */
 414             zeta2            = _mm_mul_ps(beta2,rsq30);
 415             rinv3            = _mm_mul_ps(rinvsq30,rinv30);
 416             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
 417             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
 418             felec            = _mm_mul_ps(qq30,felec);
 419             pmecorrV         = gmx_mm_pmecorrV_ps(zeta2);
 420             velec            = _mm_nmacc_ps(pmecorrV,beta,_mm_sub_ps(rinv30,sh_ewald));
 421             velec            = _mm_mul_ps(qq30,velec);
 422
 423             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
 424
 425             /* Update potential sum for this i atom from the interaction with this j atom. */
 426             velec            = _mm_and_ps(velec,cutoff_mask);
 427             velecsum         = _mm_add_ps(velecsum,velec);
 428
 429             fscal            = felec;
 430
 431             fscal            = _mm_and_ps(fscal,cutoff_mask);
 432
 433              /* Update vectorial force */
 434             fix3             = _mm_macc_ps(dx30,fscal,fix3);
 435             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
 436             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
 437
 438             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
 439             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
 440             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
 441
 442             }
 443
 444             fjptrA             = f+j_coord_offsetA;
 445             fjptrB             = f+j_coord_offsetB;
 446             fjptrC             = f+j_coord_offsetC;
 447             fjptrD             = f+j_coord_offsetD;
 448
 449             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 450
 451             /* Inner loop uses 143 flops */
 452         }
 453
 454         if(jidx<j_index_end)
 455         {
 456
 457             /* Get j neighbor index, and coordinate index */
 458             jnrlistA         = jjnr[jidx];
 459             jnrlistB         = jjnr[jidx+1];
 460             jnrlistC         = jjnr[jidx+2];
 461             jnrlistD         = jjnr[jidx+3];
 462             /* Sign of each element will be negative for non-real atoms.
 463              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 464              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 465              */
 466             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 467             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 468             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 469             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 470             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 471             j_coord_offsetA  = DIM*jnrA;
 472             j_coord_offsetB  = DIM*jnrB;
 473             j_coord_offsetC  = DIM*jnrC;
 474             j_coord_offsetD  = DIM*jnrD;
 475
 476             /* load j atom coordinates */
 477             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 478                                               x+j_coord_offsetC,x+j_coord_offsetD,
 479                                               &jx0,&jy0,&jz0);
 480
 481             /* Calculate displacement vector */
 482             dx00             = _mm_sub_ps(ix0,jx0);
 483             dy00             = _mm_sub_ps(iy0,jy0);
 484             dz00             = _mm_sub_ps(iz0,jz0);
 485             dx10             = _mm_sub_ps(ix1,jx0);
 486             dy10             = _mm_sub_ps(iy1,jy0);
 487             dz10             = _mm_sub_ps(iz1,jz0);
 488             dx20             = _mm_sub_ps(ix2,jx0);
 489             dy20             = _mm_sub_ps(iy2,jy0);
 490             dz20             = _mm_sub_ps(iz2,jz0);
 491             dx30             = _mm_sub_ps(ix3,jx0);
 492             dy30             = _mm_sub_ps(iy3,jy0);
 493             dz30             = _mm_sub_ps(iz3,jz0);
 494
 495             /* Calculate squared distance and things based on it */
 496             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 497             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 498             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 499             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 500
 501             rinv10           = gmx_mm_invsqrt_ps(rsq10);
 502             rinv20           = gmx_mm_invsqrt_ps(rsq20);
 503             rinv30           = gmx_mm_invsqrt_ps(rsq30);
 504
 505             rinvsq00         = gmx_mm_inv_ps(rsq00);
 506             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 507             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 508             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 509
 510             /* Load parameters for j particles */
 511             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 512                                                               charge+jnrC+0,charge+jnrD+0);
 513             vdwjidx0A        = 2*vdwtype[jnrA+0];
 514             vdwjidx0B        = 2*vdwtype[jnrB+0];
 515             vdwjidx0C        = 2*vdwtype[jnrC+0];
 516             vdwjidx0D        = 2*vdwtype[jnrD+0];
 517
 518             fjx0             = _mm_setzero_ps();
 519             fjy0             = _mm_setzero_ps();
 520             fjz0             = _mm_setzero_ps();
 521
 522             /**************************
 523              * CALCULATE INTERACTIONS *
 524              **************************/
 525
 526             if (gmx_mm_any_lt(rsq00,rcutoff2))
 527             {
 528
 529             /* Compute parameters for interactions between i and j atoms */
 530             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 531                                          vdwparam+vdwioffset0+vdwjidx0B,
 532                                          vdwparam+vdwioffset0+vdwjidx0C,
 533                                          vdwparam+vdwioffset0+vdwjidx0D,
 534                                          &c6_00,&c12_00);
 535
 536             /* LENNARD-JONES DISPERSION/REPULSION */
 537
 538             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 539             vvdw6            = _mm_mul_ps(c6_00,rinvsix);
 540             vvdw12           = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix));
 541             vvdw             = _mm_msub_ps(_mm_nmacc_ps(c12_00,_mm_mul_ps(sh_vdw_invrcut6,sh_vdw_invrcut6),vvdw12),one_twelfth,
 542                                           _mm_mul_ps( _mm_nmacc_ps(c6_00,sh_vdw_invrcut6,vvdw6),one_sixth));
 543             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00);
 544
 545             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
 546
 547             /* Update potential sum for this i atom from the interaction with this j atom. */
 548             vvdw             = _mm_and_ps(vvdw,cutoff_mask);
 549             vvdw             = _mm_andnot_ps(dummy_mask,vvdw);
 550             vvdwsum          = _mm_add_ps(vvdwsum,vvdw);
 551
 552             fscal            = fvdw;
 553
 554             fscal            = _mm_and_ps(fscal,cutoff_mask);
 555
 556             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 557
 558              /* Update vectorial force */
 559             fix0             = _mm_macc_ps(dx00,fscal,fix0);
 560             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
 561             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
 562
 563             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
 564             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
 565             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
 566
 567             }
 568
 569             /**************************
 570              * CALCULATE INTERACTIONS *
 571              **************************/
 572
 573             if (gmx_mm_any_lt(rsq10,rcutoff2))
 574             {
 575
 576             r10              = _mm_mul_ps(rsq10,rinv10);
 577             r10              = _mm_andnot_ps(dummy_mask,r10);
 578
 579             /* Compute parameters for interactions between i and j atoms */
 580             qq10             = _mm_mul_ps(iq1,jq0);
 581
 582             /* EWALD ELECTROSTATICS */
 583
 584             /* Analytical PME correction */
 585             zeta2            = _mm_mul_ps(beta2,rsq10);
 586             rinv3            = _mm_mul_ps(rinvsq10,rinv10);
 587             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
 588             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
 589             felec            = _mm_mul_ps(qq10,felec);
 590             pmecorrV         = gmx_mm_pmecorrV_ps(zeta2);
 591             velec            = _mm_nmacc_ps(pmecorrV,beta,_mm_sub_ps(rinv10,sh_ewald));
 592             velec            = _mm_mul_ps(qq10,velec);
 593
 594             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
 595
 596             /* Update potential sum for this i atom from the interaction with this j atom. */
 597             velec            = _mm_and_ps(velec,cutoff_mask);
 598             velec            = _mm_andnot_ps(dummy_mask,velec);
 599             velecsum         = _mm_add_ps(velecsum,velec);
 600
 601             fscal            = felec;
 602
 603             fscal            = _mm_and_ps(fscal,cutoff_mask);
 604
 605             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 606
 607              /* Update vectorial force */
 608             fix1             = _mm_macc_ps(dx10,fscal,fix1);
 609             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
 610             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
 611
 612             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
 613             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
 614             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
 615
 616             }
 617
 618             /**************************
 619              * CALCULATE INTERACTIONS *
 620              **************************/
 621
 622             if (gmx_mm_any_lt(rsq20,rcutoff2))
 623             {
 624
 625             r20              = _mm_mul_ps(rsq20,rinv20);
 626             r20              = _mm_andnot_ps(dummy_mask,r20);
 627
 628             /* Compute parameters for interactions between i and j atoms */
 629             qq20             = _mm_mul_ps(iq2,jq0);
 630
 631             /* EWALD ELECTROSTATICS */
 632
 633             /* Analytical PME correction */
 634             zeta2            = _mm_mul_ps(beta2,rsq20);
 635             rinv3            = _mm_mul_ps(rinvsq20,rinv20);
 636             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
 637             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
 638             felec            = _mm_mul_ps(qq20,felec);
 639             pmecorrV         = gmx_mm_pmecorrV_ps(zeta2);
 640             velec            = _mm_nmacc_ps(pmecorrV,beta,_mm_sub_ps(rinv20,sh_ewald));
 641             velec            = _mm_mul_ps(qq20,velec);
 642
 643             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
 644
 645             /* Update potential sum for this i atom from the interaction with this j atom. */
 646             velec            = _mm_and_ps(velec,cutoff_mask);
 647             velec            = _mm_andnot_ps(dummy_mask,velec);
 648             velecsum         = _mm_add_ps(velecsum,velec);
 649
 650             fscal            = felec;
 651
 652             fscal            = _mm_and_ps(fscal,cutoff_mask);
 653
 654             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 655
 656              /* Update vectorial force */
 657             fix2             = _mm_macc_ps(dx20,fscal,fix2);
 658             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
 659             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
 660
 661             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
 662             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
 663             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
 664
 665             }
 666
 667             /**************************
 668              * CALCULATE INTERACTIONS *
 669              **************************/
 670
 671             if (gmx_mm_any_lt(rsq30,rcutoff2))
 672             {
 673
 674             r30              = _mm_mul_ps(rsq30,rinv30);
 675             r30              = _mm_andnot_ps(dummy_mask,r30);
 676
 677             /* Compute parameters for interactions between i and j atoms */
 678             qq30             = _mm_mul_ps(iq3,jq0);
 679
 680             /* EWALD ELECTROSTATICS */
 681
 682             /* Analytical PME correction */
 683             zeta2            = _mm_mul_ps(beta2,rsq30);
 684             rinv3            = _mm_mul_ps(rinvsq30,rinv30);
 685             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
 686             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
 687             felec            = _mm_mul_ps(qq30,felec);
 688             pmecorrV         = gmx_mm_pmecorrV_ps(zeta2);
 689             velec            = _mm_nmacc_ps(pmecorrV,beta,_mm_sub_ps(rinv30,sh_ewald));
 690             velec            = _mm_mul_ps(qq30,velec);
 691
 692             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
 693
 694             /* Update potential sum for this i atom from the interaction with this j atom. */
 695             velec            = _mm_and_ps(velec,cutoff_mask);
 696             velec            = _mm_andnot_ps(dummy_mask,velec);
 697             velecsum         = _mm_add_ps(velecsum,velec);
 698
 699             fscal            = felec;
 700
 701             fscal            = _mm_and_ps(fscal,cutoff_mask);
 702
 703             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 704
 705              /* Update vectorial force */
 706             fix3             = _mm_macc_ps(dx30,fscal,fix3);
 707             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
 708             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
 709
 710             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
 711             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
 712             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
 713
 714             }
 715
 716             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 717             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 718             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 719             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 720
 721             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 722
 723             /* Inner loop uses 146 flops */
 724         }
 725
 726         /* End of innermost loop */
 727
 728         gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 729                                               f+i_coord_offset,fshift+i_shift_offset);
 730
 731         ggid                        = gid[iidx];
 732         /* Update potential energies */
 733         gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 734         gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 735
 736         /* Increment number of inner iterations */
 737         inneriter                  += j_index_end - j_index_start;
 738
 739         /* Outer loop uses 26 flops */
 740     }
 741
 742     /* Increment number of outer iterations */
 743     outeriter        += nri;
 744
 745     /* Update outer/inner flops */
 746
 747     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*146);
 748 }
 749 /*
 750  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_F_avx_128_fma_single
 751  * Electrostatics interaction: Ewald
 752  * VdW interaction:            LennardJones
 753  * Geometry:                   Water4-Particle
 754  * Calculate force/pot:        Force
 755  */
 756 void
 757 nb_kernel_ElecEwSh_VdwLJSh_GeomW4P1_F_avx_128_fma_single
 758                     (t_nblist                    * gmx_restrict       nlist,
 759                      rvec                        * gmx_restrict          xx,
 760                      rvec                        * gmx_restrict          ff,
 761                      t_forcerec                  * gmx_restrict          fr,
 762                      t_mdatoms                   * gmx_restrict     mdatoms,
 763                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 764                      t_nrnb                      * gmx_restrict        nrnb)
 765 {
 766     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 767      * just 0 for non-waters.
 768      * Suffixes A,B,C,D refer to j loop unrolling done with AVX_128, e.g. for the four different
 769      * jnr indices corresponding to data put in the four positions in the SIMD register.
 770      */
 771     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 772     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 773     int              jnrA,jnrB,jnrC,jnrD;
 774     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 775     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 776     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 777     real             rcutoff_scalar;
 778     real             *shiftvec,*fshift,*x,*f;
 779     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 780     real             scratch[4*DIM];
 781     __m128           fscal,rcutoff,rcutoff2,jidxall;
 782     int              vdwioffset0;
 783     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 784     int              vdwioffset1;
 785     __m128           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 786     int              vdwioffset2;
 787     __m128           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 788     int              vdwioffset3;
 789     __m128           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 790     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 791     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 792     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 793     __m128           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 794     __m128           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 795     __m128           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 796     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
 797     real             *charge;
 798     int              nvdwtype;
 799     __m128           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 800     int              *vdwtype;
 801     real             *vdwparam;
 802     __m128           one_sixth   = _mm_set1_ps(1.0/6.0);
 803     __m128           one_twelfth = _mm_set1_ps(1.0/12.0);
 804     __m128i          ewitab;
 805     __m128           ewtabscale,eweps,twoeweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 806     __m128           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 807     real             *ewtab;
 808     __m128           dummy_mask,cutoff_mask;
 809     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 810     __m128           one     = _mm_set1_ps(1.0);
 811     __m128           two     = _mm_set1_ps(2.0);
 812     x                = xx[0];
 813     f                = ff[0];
 814
 815     nri              = nlist->nri;
 816     iinr             = nlist->iinr;
 817     jindex           = nlist->jindex;
 818     jjnr             = nlist->jjnr;
 819     shiftidx         = nlist->shift;
 820     gid              = nlist->gid;
 821     shiftvec         = fr->shift_vec[0];
 822     fshift           = fr->fshift[0];
 823     facel            = _mm_set1_ps(fr->epsfac);
 824     charge           = mdatoms->chargeA;
 825     nvdwtype         = fr->ntype;
 826     vdwparam         = fr->nbfp;
 827     vdwtype          = mdatoms->typeA;
 828
 829     sh_ewald         = _mm_set1_ps(fr->ic->sh_ewald);
 830     beta             = _mm_set1_ps(fr->ic->ewaldcoeff_q);
 831     beta2            = _mm_mul_ps(beta,beta);
 832     beta3            = _mm_mul_ps(beta,beta2);
 833     ewtab            = fr->ic->tabq_coul_F;
 834     ewtabscale       = _mm_set1_ps(fr->ic->tabq_scale);
 835     ewtabhalfspace   = _mm_set1_ps(0.5/fr->ic->tabq_scale);
 836
 837     /* Setup water-specific parameters */
 838     inr              = nlist->iinr[0];
 839     iq1              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
 840     iq2              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
 841     iq3              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
 842     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 843
 844     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 845     rcutoff_scalar   = fr->rcoulomb;
 846     rcutoff          = _mm_set1_ps(rcutoff_scalar);
 847     rcutoff2         = _mm_mul_ps(rcutoff,rcutoff);
 848
 849     sh_vdw_invrcut6  = _mm_set1_ps(fr->ic->sh_invrc6);
 850     rvdw             = _mm_set1_ps(fr->rvdw);
 851
 852     /* Avoid stupid compiler warnings */
 853     jnrA = jnrB = jnrC = jnrD = 0;
 854     j_coord_offsetA = 0;
 855     j_coord_offsetB = 0;
 856     j_coord_offsetC = 0;
 857     j_coord_offsetD = 0;
 858
 859     outeriter        = 0;
 860     inneriter        = 0;
 861
 862     for(iidx=0;iidx<4*DIM;iidx++)
 863     {
 864         scratch[iidx] = 0.0;
 865     }
 866
 867     /* Start outer loop over neighborlists */
 868     for(iidx=0; iidx<nri; iidx++)
 869     {
 870         /* Load shift vector for this list */
 871         i_shift_offset   = DIM*shiftidx[iidx];
 872
 873         /* Load limits for loop over neighbors */
 874         j_index_start    = jindex[iidx];
 875         j_index_end      = jindex[iidx+1];
 876
 877         /* Get outer coordinate index */
 878         inr              = iinr[iidx];
 879         i_coord_offset   = DIM*inr;
 880
 881         /* Load i particle coords and add shift vector */
 882         gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 883                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 884
 885         fix0             = _mm_setzero_ps();
 886         fiy0             = _mm_setzero_ps();
 887         fiz0             = _mm_setzero_ps();
 888         fix1             = _mm_setzero_ps();
 889         fiy1             = _mm_setzero_ps();
 890         fiz1             = _mm_setzero_ps();
 891         fix2             = _mm_setzero_ps();
 892         fiy2             = _mm_setzero_ps();
 893         fiz2             = _mm_setzero_ps();
 894         fix3             = _mm_setzero_ps();
 895         fiy3             = _mm_setzero_ps();
 896         fiz3             = _mm_setzero_ps();
 897
 898         /* Start inner kernel loop */
 899         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 900         {
 901
 902             /* Get j neighbor index, and coordinate index */
 903             jnrA             = jjnr[jidx];
 904             jnrB             = jjnr[jidx+1];
 905             jnrC             = jjnr[jidx+2];
 906             jnrD             = jjnr[jidx+3];
 907             j_coord_offsetA  = DIM*jnrA;
 908             j_coord_offsetB  = DIM*jnrB;
 909             j_coord_offsetC  = DIM*jnrC;
 910             j_coord_offsetD  = DIM*jnrD;
 911
 912             /* load j atom coordinates */
 913             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 914                                               x+j_coord_offsetC,x+j_coord_offsetD,
 915                                               &jx0,&jy0,&jz0);
 916
 917             /* Calculate displacement vector */
 918             dx00             = _mm_sub_ps(ix0,jx0);
 919             dy00             = _mm_sub_ps(iy0,jy0);
 920             dz00             = _mm_sub_ps(iz0,jz0);
 921             dx10             = _mm_sub_ps(ix1,jx0);
 922             dy10             = _mm_sub_ps(iy1,jy0);
 923             dz10             = _mm_sub_ps(iz1,jz0);
 924             dx20             = _mm_sub_ps(ix2,jx0);
 925             dy20             = _mm_sub_ps(iy2,jy0);
 926             dz20             = _mm_sub_ps(iz2,jz0);
 927             dx30             = _mm_sub_ps(ix3,jx0);
 928             dy30             = _mm_sub_ps(iy3,jy0);
 929             dz30             = _mm_sub_ps(iz3,jz0);
 930
 931             /* Calculate squared distance and things based on it */
 932             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 933             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
 934             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
 935             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
 936
 937             rinv10           = gmx_mm_invsqrt_ps(rsq10);
 938             rinv20           = gmx_mm_invsqrt_ps(rsq20);
 939             rinv30           = gmx_mm_invsqrt_ps(rsq30);
 940
 941             rinvsq00         = gmx_mm_inv_ps(rsq00);
 942             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
 943             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
 944             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
 945
 946             /* Load parameters for j particles */
 947             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 948                                                               charge+jnrC+0,charge+jnrD+0);
 949             vdwjidx0A        = 2*vdwtype[jnrA+0];
 950             vdwjidx0B        = 2*vdwtype[jnrB+0];
 951             vdwjidx0C        = 2*vdwtype[jnrC+0];
 952             vdwjidx0D        = 2*vdwtype[jnrD+0];
 953
 954             fjx0             = _mm_setzero_ps();
 955             fjy0             = _mm_setzero_ps();
 956             fjz0             = _mm_setzero_ps();
 957
 958             /**************************
 959              * CALCULATE INTERACTIONS *
 960              **************************/
 961
 962             if (gmx_mm_any_lt(rsq00,rcutoff2))
 963             {
 964
 965             /* Compute parameters for interactions between i and j atoms */
 966             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
 967                                          vdwparam+vdwioffset0+vdwjidx0B,
 968                                          vdwparam+vdwioffset0+vdwjidx0C,
 969                                          vdwparam+vdwioffset0+vdwjidx0D,
 970                                          &c6_00,&c12_00);
 971
 972             /* LENNARD-JONES DISPERSION/REPULSION */
 973
 974             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 975             fvdw             = _mm_mul_ps(_mm_msub_ps(c12_00,rinvsix,c6_00),_mm_mul_ps(rinvsix,rinvsq00));
 976
 977             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
 978
 979             fscal            = fvdw;
 980
 981             fscal            = _mm_and_ps(fscal,cutoff_mask);
 982
 983              /* Update vectorial force */
 984             fix0             = _mm_macc_ps(dx00,fscal,fix0);
 985             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
 986             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
 987
 988             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
 989             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
 990             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
 991
 992             }
 993
 994             /**************************
 995              * CALCULATE INTERACTIONS *
 996              **************************/
 997
 998             if (gmx_mm_any_lt(rsq10,rcutoff2))
 999             {
1000
1001             r10              = _mm_mul_ps(rsq10,rinv10);
1002
1003             /* Compute parameters for interactions between i and j atoms */
1004             qq10             = _mm_mul_ps(iq1,jq0);
1005
1006             /* EWALD ELECTROSTATICS */
1007
1008             /* Analytical PME correction */
1009             zeta2            = _mm_mul_ps(beta2,rsq10);
1010             rinv3            = _mm_mul_ps(rinvsq10,rinv10);
1011             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
1012             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
1013             felec            = _mm_mul_ps(qq10,felec);
1014
1015             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
1016
1017             fscal            = felec;
1018
1019             fscal            = _mm_and_ps(fscal,cutoff_mask);
1020
1021              /* Update vectorial force */
1022             fix1             = _mm_macc_ps(dx10,fscal,fix1);
1023             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
1024             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
1025
1026             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
1027             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
1028             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
1029
1030             }
1031
1032             /**************************
1033              * CALCULATE INTERACTIONS *
1034              **************************/
1035
1036             if (gmx_mm_any_lt(rsq20,rcutoff2))
1037             {
1038
1039             r20              = _mm_mul_ps(rsq20,rinv20);
1040
1041             /* Compute parameters for interactions between i and j atoms */
1042             qq20             = _mm_mul_ps(iq2,jq0);
1043
1044             /* EWALD ELECTROSTATICS */
1045
1046             /* Analytical PME correction */
1047             zeta2            = _mm_mul_ps(beta2,rsq20);
1048             rinv3            = _mm_mul_ps(rinvsq20,rinv20);
1049             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
1050             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
1051             felec            = _mm_mul_ps(qq20,felec);
1052
1053             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
1054
1055             fscal            = felec;
1056
1057             fscal            = _mm_and_ps(fscal,cutoff_mask);
1058
1059              /* Update vectorial force */
1060             fix2             = _mm_macc_ps(dx20,fscal,fix2);
1061             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
1062             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
1063
1064             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
1065             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
1066             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
1067
1068             }
1069
1070             /**************************
1071              * CALCULATE INTERACTIONS *
1072              **************************/
1073
1074             if (gmx_mm_any_lt(rsq30,rcutoff2))
1075             {
1076
1077             r30              = _mm_mul_ps(rsq30,rinv30);
1078
1079             /* Compute parameters for interactions between i and j atoms */
1080             qq30             = _mm_mul_ps(iq3,jq0);
1081
1082             /* EWALD ELECTROSTATICS */
1083
1084             /* Analytical PME correction */
1085             zeta2            = _mm_mul_ps(beta2,rsq30);
1086             rinv3            = _mm_mul_ps(rinvsq30,rinv30);
1087             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
1088             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
1089             felec            = _mm_mul_ps(qq30,felec);
1090
1091             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
1092
1093             fscal            = felec;
1094
1095             fscal            = _mm_and_ps(fscal,cutoff_mask);
1096
1097              /* Update vectorial force */
1098             fix3             = _mm_macc_ps(dx30,fscal,fix3);
1099             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
1100             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
1101
1102             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
1103             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
1104             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
1105
1106             }
1107
1108             fjptrA             = f+j_coord_offsetA;
1109             fjptrB             = f+j_coord_offsetB;
1110             fjptrC             = f+j_coord_offsetC;
1111             fjptrD             = f+j_coord_offsetD;
1112
1113             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1114
1115             /* Inner loop uses 126 flops */
1116         }
1117
1118         if(jidx<j_index_end)
1119         {
1120
1121             /* Get j neighbor index, and coordinate index */
1122             jnrlistA         = jjnr[jidx];
1123             jnrlistB         = jjnr[jidx+1];
1124             jnrlistC         = jjnr[jidx+2];
1125             jnrlistD         = jjnr[jidx+3];
1126             /* Sign of each element will be negative for non-real atoms.
1127              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1128              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1129              */
1130             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1131             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1132             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1133             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1134             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1135             j_coord_offsetA  = DIM*jnrA;
1136             j_coord_offsetB  = DIM*jnrB;
1137             j_coord_offsetC  = DIM*jnrC;
1138             j_coord_offsetD  = DIM*jnrD;
1139
1140             /* load j atom coordinates */
1141             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1142                                               x+j_coord_offsetC,x+j_coord_offsetD,
1143                                               &jx0,&jy0,&jz0);
1144
1145             /* Calculate displacement vector */
1146             dx00             = _mm_sub_ps(ix0,jx0);
1147             dy00             = _mm_sub_ps(iy0,jy0);
1148             dz00             = _mm_sub_ps(iz0,jz0);
1149             dx10             = _mm_sub_ps(ix1,jx0);
1150             dy10             = _mm_sub_ps(iy1,jy0);
1151             dz10             = _mm_sub_ps(iz1,jz0);
1152             dx20             = _mm_sub_ps(ix2,jx0);
1153             dy20             = _mm_sub_ps(iy2,jy0);
1154             dz20             = _mm_sub_ps(iz2,jz0);
1155             dx30             = _mm_sub_ps(ix3,jx0);
1156             dy30             = _mm_sub_ps(iy3,jy0);
1157             dz30             = _mm_sub_ps(iz3,jz0);
1158
1159             /* Calculate squared distance and things based on it */
1160             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
1161             rsq10            = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
1162             rsq20            = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
1163             rsq30            = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
1164
1165             rinv10           = gmx_mm_invsqrt_ps(rsq10);
1166             rinv20           = gmx_mm_invsqrt_ps(rsq20);
1167             rinv30           = gmx_mm_invsqrt_ps(rsq30);
1168
1169             rinvsq00         = gmx_mm_inv_ps(rsq00);
1170             rinvsq10         = _mm_mul_ps(rinv10,rinv10);
1171             rinvsq20         = _mm_mul_ps(rinv20,rinv20);
1172             rinvsq30         = _mm_mul_ps(rinv30,rinv30);
1173
1174             /* Load parameters for j particles */
1175             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1176                                                               charge+jnrC+0,charge+jnrD+0);
1177             vdwjidx0A        = 2*vdwtype[jnrA+0];
1178             vdwjidx0B        = 2*vdwtype[jnrB+0];
1179             vdwjidx0C        = 2*vdwtype[jnrC+0];
1180             vdwjidx0D        = 2*vdwtype[jnrD+0];
1181
1182             fjx0             = _mm_setzero_ps();
1183             fjy0             = _mm_setzero_ps();
1184             fjz0             = _mm_setzero_ps();
1185
1186             /**************************
1187              * CALCULATE INTERACTIONS *
1188              **************************/
1189
1190             if (gmx_mm_any_lt(rsq00,rcutoff2))
1191             {
1192
1193             /* Compute parameters for interactions between i and j atoms */
1194             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
1195                                          vdwparam+vdwioffset0+vdwjidx0B,
1196                                          vdwparam+vdwioffset0+vdwjidx0C,
1197                                          vdwparam+vdwioffset0+vdwjidx0D,
1198                                          &c6_00,&c12_00);
1199
1200             /* LENNARD-JONES DISPERSION/REPULSION */
1201
1202             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1203             fvdw             = _mm_mul_ps(_mm_msub_ps(c12_00,rinvsix,c6_00),_mm_mul_ps(rinvsix,rinvsq00));
1204
1205             cutoff_mask      = _mm_cmplt_ps(rsq00,rcutoff2);
1206
1207             fscal            = fvdw;
1208
1209             fscal            = _mm_and_ps(fscal,cutoff_mask);
1210
1211             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1212
1213              /* Update vectorial force */
1214             fix0             = _mm_macc_ps(dx00,fscal,fix0);
1215             fiy0             = _mm_macc_ps(dy00,fscal,fiy0);
1216             fiz0             = _mm_macc_ps(dz00,fscal,fiz0);
1217
1218             fjx0             = _mm_macc_ps(dx00,fscal,fjx0);
1219             fjy0             = _mm_macc_ps(dy00,fscal,fjy0);
1220             fjz0             = _mm_macc_ps(dz00,fscal,fjz0);
1221
1222             }
1223
1224             /**************************
1225              * CALCULATE INTERACTIONS *
1226              **************************/
1227
1228             if (gmx_mm_any_lt(rsq10,rcutoff2))
1229             {
1230
1231             r10              = _mm_mul_ps(rsq10,rinv10);
1232             r10              = _mm_andnot_ps(dummy_mask,r10);
1233
1234             /* Compute parameters for interactions between i and j atoms */
1235             qq10             = _mm_mul_ps(iq1,jq0);
1236
1237             /* EWALD ELECTROSTATICS */
1238
1239             /* Analytical PME correction */
1240             zeta2            = _mm_mul_ps(beta2,rsq10);
1241             rinv3            = _mm_mul_ps(rinvsq10,rinv10);
1242             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
1243             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
1244             felec            = _mm_mul_ps(qq10,felec);
1245
1246             cutoff_mask      = _mm_cmplt_ps(rsq10,rcutoff2);
1247
1248             fscal            = felec;
1249
1250             fscal            = _mm_and_ps(fscal,cutoff_mask);
1251
1252             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1253
1254              /* Update vectorial force */
1255             fix1             = _mm_macc_ps(dx10,fscal,fix1);
1256             fiy1             = _mm_macc_ps(dy10,fscal,fiy1);
1257             fiz1             = _mm_macc_ps(dz10,fscal,fiz1);
1258
1259             fjx0             = _mm_macc_ps(dx10,fscal,fjx0);
1260             fjy0             = _mm_macc_ps(dy10,fscal,fjy0);
1261             fjz0             = _mm_macc_ps(dz10,fscal,fjz0);
1262
1263             }
1264
1265             /**************************
1266              * CALCULATE INTERACTIONS *
1267              **************************/
1268
1269             if (gmx_mm_any_lt(rsq20,rcutoff2))
1270             {
1271
1272             r20              = _mm_mul_ps(rsq20,rinv20);
1273             r20              = _mm_andnot_ps(dummy_mask,r20);
1274
1275             /* Compute parameters for interactions between i and j atoms */
1276             qq20             = _mm_mul_ps(iq2,jq0);
1277
1278             /* EWALD ELECTROSTATICS */
1279
1280             /* Analytical PME correction */
1281             zeta2            = _mm_mul_ps(beta2,rsq20);
1282             rinv3            = _mm_mul_ps(rinvsq20,rinv20);
1283             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
1284             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
1285             felec            = _mm_mul_ps(qq20,felec);
1286
1287             cutoff_mask      = _mm_cmplt_ps(rsq20,rcutoff2);
1288
1289             fscal            = felec;
1290
1291             fscal            = _mm_and_ps(fscal,cutoff_mask);
1292
1293             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1294
1295              /* Update vectorial force */
1296             fix2             = _mm_macc_ps(dx20,fscal,fix2);
1297             fiy2             = _mm_macc_ps(dy20,fscal,fiy2);
1298             fiz2             = _mm_macc_ps(dz20,fscal,fiz2);
1299
1300             fjx0             = _mm_macc_ps(dx20,fscal,fjx0);
1301             fjy0             = _mm_macc_ps(dy20,fscal,fjy0);
1302             fjz0             = _mm_macc_ps(dz20,fscal,fjz0);
1303
1304             }
1305
1306             /**************************
1307              * CALCULATE INTERACTIONS *
1308              **************************/
1309
1310             if (gmx_mm_any_lt(rsq30,rcutoff2))
1311             {
1312
1313             r30              = _mm_mul_ps(rsq30,rinv30);
1314             r30              = _mm_andnot_ps(dummy_mask,r30);
1315
1316             /* Compute parameters for interactions between i and j atoms */
1317             qq30             = _mm_mul_ps(iq3,jq0);
1318
1319             /* EWALD ELECTROSTATICS */
1320
1321             /* Analytical PME correction */
1322             zeta2            = _mm_mul_ps(beta2,rsq30);
1323             rinv3            = _mm_mul_ps(rinvsq30,rinv30);
1324             pmecorrF         = gmx_mm_pmecorrF_ps(zeta2);
1325             felec            = _mm_macc_ps(pmecorrF,beta3,rinv3);
1326             felec            = _mm_mul_ps(qq30,felec);
1327
1328             cutoff_mask      = _mm_cmplt_ps(rsq30,rcutoff2);
1329
1330             fscal            = felec;
1331
1332             fscal            = _mm_and_ps(fscal,cutoff_mask);
1333
1334             fscal            = _mm_andnot_ps(dummy_mask,fscal);
1335
1336              /* Update vectorial force */
1337             fix3             = _mm_macc_ps(dx30,fscal,fix3);
1338             fiy3             = _mm_macc_ps(dy30,fscal,fiy3);
1339             fiz3             = _mm_macc_ps(dz30,fscal,fiz3);
1340
1341             fjx0             = _mm_macc_ps(dx30,fscal,fjx0);
1342             fjy0             = _mm_macc_ps(dy30,fscal,fjy0);
1343             fjz0             = _mm_macc_ps(dz30,fscal,fjz0);
1344
1345             }
1346
1347             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1348             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1349             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1350             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1351
1352             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1353
1354             /* Inner loop uses 129 flops */
1355         }
1356
1357         /* End of innermost loop */
1358
1359         gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1360                                               f+i_coord_offset,fshift+i_shift_offset);
1361
1362         /* Increment number of inner iterations */
1363         inneriter                  += j_index_end - j_index_start;
1364
1365         /* Outer loop uses 24 flops */
1366     }
1367
1368     /* Increment number of outer iterations */
1369     outeriter        += nri;
1370
1371     /* Update outer/inner flops */
1372
1373     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*129);
1374 }