src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecEw_VdwNone_GeomW3P1_avx_256_double.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_256_double kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_256_double.h"
  48 #include "kernelutil_x86_avx_256_double.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwNone_GeomW3P1_VF_avx_256_double
  52  * Electrostatics interaction: Ewald
  53  * VdW interaction:            None
  54  * Geometry:                   Water3-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecEw_VdwNone_GeomW3P1_VF_avx_256_double
  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 AVX, 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              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  77     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  78     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  79     real             rcutoff_scalar;
  80     real             *shiftvec,*fshift,*x,*f;
  81     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  82     real             scratch[4*DIM];
  83     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  84     real *           vdwioffsetptr0;
  85     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  86     real *           vdwioffsetptr1;
  87     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  88     real *           vdwioffsetptr2;
  89     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  90     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  91     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  92     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  93     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  94     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  95     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
  96     real             *charge;
  97     __m128i          ewitab;
  98     __m256d          ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
  99     __m256d          beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 100     real             *ewtab;
 101     __m256d          dummy_mask,cutoff_mask;
 102     __m128           tmpmask0,tmpmask1;
 103     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 104     __m256d          one     = _mm256_set1_pd(1.0);
 105     __m256d          two     = _mm256_set1_pd(2.0);
 106     x                = xx[0];
 107     f                = ff[0];
 108
 109     nri              = nlist->nri;
 110     iinr             = nlist->iinr;
 111     jindex           = nlist->jindex;
 112     jjnr             = nlist->jjnr;
 113     shiftidx         = nlist->shift;
 114     gid              = nlist->gid;
 115     shiftvec         = fr->shift_vec[0];
 116     fshift           = fr->fshift[0];
 117     facel            = _mm256_set1_pd(fr->epsfac);
 118     charge           = mdatoms->chargeA;
 119
 120     sh_ewald         = _mm256_set1_pd(fr->ic->sh_ewald);
 121     beta             = _mm256_set1_pd(fr->ic->ewaldcoeff_q);
 122     beta2            = _mm256_mul_pd(beta,beta);
 123     beta3            = _mm256_mul_pd(beta,beta2);
 124
 125     ewtab            = fr->ic->tabq_coul_FDV0;
 126     ewtabscale       = _mm256_set1_pd(fr->ic->tabq_scale);
 127     ewtabhalfspace   = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
 128
 129     /* Setup water-specific parameters */
 130     inr              = nlist->iinr[0];
 131     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 132     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 133     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 134
 135     /* Avoid stupid compiler warnings */
 136     jnrA = jnrB = jnrC = jnrD = 0;
 137     j_coord_offsetA = 0;
 138     j_coord_offsetB = 0;
 139     j_coord_offsetC = 0;
 140     j_coord_offsetD = 0;
 141
 142     outeriter        = 0;
 143     inneriter        = 0;
 144
 145     for(iidx=0;iidx<4*DIM;iidx++)
 146     {
 147         scratch[iidx] = 0.0;
 148     }
 149
 150     /* Start outer loop over neighborlists */
 151     for(iidx=0; iidx<nri; iidx++)
 152     {
 153         /* Load shift vector for this list */
 154         i_shift_offset   = DIM*shiftidx[iidx];
 155
 156         /* Load limits for loop over neighbors */
 157         j_index_start    = jindex[iidx];
 158         j_index_end      = jindex[iidx+1];
 159
 160         /* Get outer coordinate index */
 161         inr              = iinr[iidx];
 162         i_coord_offset   = DIM*inr;
 163
 164         /* Load i particle coords and add shift vector */
 165         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 166                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 167
 168         fix0             = _mm256_setzero_pd();
 169         fiy0             = _mm256_setzero_pd();
 170         fiz0             = _mm256_setzero_pd();
 171         fix1             = _mm256_setzero_pd();
 172         fiy1             = _mm256_setzero_pd();
 173         fiz1             = _mm256_setzero_pd();
 174         fix2             = _mm256_setzero_pd();
 175         fiy2             = _mm256_setzero_pd();
 176         fiz2             = _mm256_setzero_pd();
 177
 178         /* Reset potential sums */
 179         velecsum         = _mm256_setzero_pd();
 180
 181         /* Start inner kernel loop */
 182         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 183         {
 184
 185             /* Get j neighbor index, and coordinate index */
 186             jnrA             = jjnr[jidx];
 187             jnrB             = jjnr[jidx+1];
 188             jnrC             = jjnr[jidx+2];
 189             jnrD             = jjnr[jidx+3];
 190             j_coord_offsetA  = DIM*jnrA;
 191             j_coord_offsetB  = DIM*jnrB;
 192             j_coord_offsetC  = DIM*jnrC;
 193             j_coord_offsetD  = DIM*jnrD;
 194
 195             /* load j atom coordinates */
 196             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 197                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 198                                                  &jx0,&jy0,&jz0);
 199
 200             /* Calculate displacement vector */
 201             dx00             = _mm256_sub_pd(ix0,jx0);
 202             dy00             = _mm256_sub_pd(iy0,jy0);
 203             dz00             = _mm256_sub_pd(iz0,jz0);
 204             dx10             = _mm256_sub_pd(ix1,jx0);
 205             dy10             = _mm256_sub_pd(iy1,jy0);
 206             dz10             = _mm256_sub_pd(iz1,jz0);
 207             dx20             = _mm256_sub_pd(ix2,jx0);
 208             dy20             = _mm256_sub_pd(iy2,jy0);
 209             dz20             = _mm256_sub_pd(iz2,jz0);
 210
 211             /* Calculate squared distance and things based on it */
 212             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 213             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 214             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 215
 216             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 217             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 218             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 219
 220             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 221             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 222             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 223
 224             /* Load parameters for j particles */
 225             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 226                                                                  charge+jnrC+0,charge+jnrD+0);
 227
 228             fjx0             = _mm256_setzero_pd();
 229             fjy0             = _mm256_setzero_pd();
 230             fjz0             = _mm256_setzero_pd();
 231
 232             /**************************
 233              * CALCULATE INTERACTIONS *
 234              **************************/
 235
 236             r00              = _mm256_mul_pd(rsq00,rinv00);
 237
 238             /* Compute parameters for interactions between i and j atoms */
 239             qq00             = _mm256_mul_pd(iq0,jq0);
 240
 241             /* EWALD ELECTROSTATICS */
 242
 243             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 244             ewrt             = _mm256_mul_pd(r00,ewtabscale);
 245             ewitab           = _mm256_cvttpd_epi32(ewrt);
 246             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 247             ewitab           = _mm_slli_epi32(ewitab,2);
 248             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 249             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 250             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 251             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 252             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 253             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 254             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 255             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(rinv00,velec));
 256             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
 257
 258             /* Update potential sum for this i atom from the interaction with this j atom. */
 259             velecsum         = _mm256_add_pd(velecsum,velec);
 260
 261             fscal            = felec;
 262
 263             /* Calculate temporary vectorial force */
 264             tx               = _mm256_mul_pd(fscal,dx00);
 265             ty               = _mm256_mul_pd(fscal,dy00);
 266             tz               = _mm256_mul_pd(fscal,dz00);
 267
 268             /* Update vectorial force */
 269             fix0             = _mm256_add_pd(fix0,tx);
 270             fiy0             = _mm256_add_pd(fiy0,ty);
 271             fiz0             = _mm256_add_pd(fiz0,tz);
 272
 273             fjx0             = _mm256_add_pd(fjx0,tx);
 274             fjy0             = _mm256_add_pd(fjy0,ty);
 275             fjz0             = _mm256_add_pd(fjz0,tz);
 276
 277             /**************************
 278              * CALCULATE INTERACTIONS *
 279              **************************/
 280
 281             r10              = _mm256_mul_pd(rsq10,rinv10);
 282
 283             /* Compute parameters for interactions between i and j atoms */
 284             qq10             = _mm256_mul_pd(iq1,jq0);
 285
 286             /* EWALD ELECTROSTATICS */
 287
 288             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 289             ewrt             = _mm256_mul_pd(r10,ewtabscale);
 290             ewitab           = _mm256_cvttpd_epi32(ewrt);
 291             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 292             ewitab           = _mm_slli_epi32(ewitab,2);
 293             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 294             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 295             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 296             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 297             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 298             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 299             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 300             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(rinv10,velec));
 301             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
 302
 303             /* Update potential sum for this i atom from the interaction with this j atom. */
 304             velecsum         = _mm256_add_pd(velecsum,velec);
 305
 306             fscal            = felec;
 307
 308             /* Calculate temporary vectorial force */
 309             tx               = _mm256_mul_pd(fscal,dx10);
 310             ty               = _mm256_mul_pd(fscal,dy10);
 311             tz               = _mm256_mul_pd(fscal,dz10);
 312
 313             /* Update vectorial force */
 314             fix1             = _mm256_add_pd(fix1,tx);
 315             fiy1             = _mm256_add_pd(fiy1,ty);
 316             fiz1             = _mm256_add_pd(fiz1,tz);
 317
 318             fjx0             = _mm256_add_pd(fjx0,tx);
 319             fjy0             = _mm256_add_pd(fjy0,ty);
 320             fjz0             = _mm256_add_pd(fjz0,tz);
 321
 322             /**************************
 323              * CALCULATE INTERACTIONS *
 324              **************************/
 325
 326             r20              = _mm256_mul_pd(rsq20,rinv20);
 327
 328             /* Compute parameters for interactions between i and j atoms */
 329             qq20             = _mm256_mul_pd(iq2,jq0);
 330
 331             /* EWALD ELECTROSTATICS */
 332
 333             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 334             ewrt             = _mm256_mul_pd(r20,ewtabscale);
 335             ewitab           = _mm256_cvttpd_epi32(ewrt);
 336             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 337             ewitab           = _mm_slli_epi32(ewitab,2);
 338             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 339             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 340             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 341             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 342             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 343             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 344             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 345             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(rinv20,velec));
 346             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
 347
 348             /* Update potential sum for this i atom from the interaction with this j atom. */
 349             velecsum         = _mm256_add_pd(velecsum,velec);
 350
 351             fscal            = felec;
 352
 353             /* Calculate temporary vectorial force */
 354             tx               = _mm256_mul_pd(fscal,dx20);
 355             ty               = _mm256_mul_pd(fscal,dy20);
 356             tz               = _mm256_mul_pd(fscal,dz20);
 357
 358             /* Update vectorial force */
 359             fix2             = _mm256_add_pd(fix2,tx);
 360             fiy2             = _mm256_add_pd(fiy2,ty);
 361             fiz2             = _mm256_add_pd(fiz2,tz);
 362
 363             fjx0             = _mm256_add_pd(fjx0,tx);
 364             fjy0             = _mm256_add_pd(fjy0,ty);
 365             fjz0             = _mm256_add_pd(fjz0,tz);
 366
 367             fjptrA             = f+j_coord_offsetA;
 368             fjptrB             = f+j_coord_offsetB;
 369             fjptrC             = f+j_coord_offsetC;
 370             fjptrD             = f+j_coord_offsetD;
 371
 372             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 373
 374             /* Inner loop uses 126 flops */
 375         }
 376
 377         if(jidx<j_index_end)
 378         {
 379
 380             /* Get j neighbor index, and coordinate index */
 381             jnrlistA         = jjnr[jidx];
 382             jnrlistB         = jjnr[jidx+1];
 383             jnrlistC         = jjnr[jidx+2];
 384             jnrlistD         = jjnr[jidx+3];
 385             /* Sign of each element will be negative for non-real atoms.
 386              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 387              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 388              */
 389             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 390
 391             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 392             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 393             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 394
 395             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 396             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 397             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 398             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 399             j_coord_offsetA  = DIM*jnrA;
 400             j_coord_offsetB  = DIM*jnrB;
 401             j_coord_offsetC  = DIM*jnrC;
 402             j_coord_offsetD  = DIM*jnrD;
 403
 404             /* load j atom coordinates */
 405             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 406                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 407                                                  &jx0,&jy0,&jz0);
 408
 409             /* Calculate displacement vector */
 410             dx00             = _mm256_sub_pd(ix0,jx0);
 411             dy00             = _mm256_sub_pd(iy0,jy0);
 412             dz00             = _mm256_sub_pd(iz0,jz0);
 413             dx10             = _mm256_sub_pd(ix1,jx0);
 414             dy10             = _mm256_sub_pd(iy1,jy0);
 415             dz10             = _mm256_sub_pd(iz1,jz0);
 416             dx20             = _mm256_sub_pd(ix2,jx0);
 417             dy20             = _mm256_sub_pd(iy2,jy0);
 418             dz20             = _mm256_sub_pd(iz2,jz0);
 419
 420             /* Calculate squared distance and things based on it */
 421             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 422             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 423             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 424
 425             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 426             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 427             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 428
 429             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 430             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 431             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 432
 433             /* Load parameters for j particles */
 434             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 435                                                                  charge+jnrC+0,charge+jnrD+0);
 436
 437             fjx0             = _mm256_setzero_pd();
 438             fjy0             = _mm256_setzero_pd();
 439             fjz0             = _mm256_setzero_pd();
 440
 441             /**************************
 442              * CALCULATE INTERACTIONS *
 443              **************************/
 444
 445             r00              = _mm256_mul_pd(rsq00,rinv00);
 446             r00              = _mm256_andnot_pd(dummy_mask,r00);
 447
 448             /* Compute parameters for interactions between i and j atoms */
 449             qq00             = _mm256_mul_pd(iq0,jq0);
 450
 451             /* EWALD ELECTROSTATICS */
 452
 453             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 454             ewrt             = _mm256_mul_pd(r00,ewtabscale);
 455             ewitab           = _mm256_cvttpd_epi32(ewrt);
 456             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 457             ewitab           = _mm_slli_epi32(ewitab,2);
 458             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 459             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 460             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 461             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 462             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 463             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 464             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 465             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(rinv00,velec));
 466             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
 467
 468             /* Update potential sum for this i atom from the interaction with this j atom. */
 469             velec            = _mm256_andnot_pd(dummy_mask,velec);
 470             velecsum         = _mm256_add_pd(velecsum,velec);
 471
 472             fscal            = felec;
 473
 474             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 475
 476             /* Calculate temporary vectorial force */
 477             tx               = _mm256_mul_pd(fscal,dx00);
 478             ty               = _mm256_mul_pd(fscal,dy00);
 479             tz               = _mm256_mul_pd(fscal,dz00);
 480
 481             /* Update vectorial force */
 482             fix0             = _mm256_add_pd(fix0,tx);
 483             fiy0             = _mm256_add_pd(fiy0,ty);
 484             fiz0             = _mm256_add_pd(fiz0,tz);
 485
 486             fjx0             = _mm256_add_pd(fjx0,tx);
 487             fjy0             = _mm256_add_pd(fjy0,ty);
 488             fjz0             = _mm256_add_pd(fjz0,tz);
 489
 490             /**************************
 491              * CALCULATE INTERACTIONS *
 492              **************************/
 493
 494             r10              = _mm256_mul_pd(rsq10,rinv10);
 495             r10              = _mm256_andnot_pd(dummy_mask,r10);
 496
 497             /* Compute parameters for interactions between i and j atoms */
 498             qq10             = _mm256_mul_pd(iq1,jq0);
 499
 500             /* EWALD ELECTROSTATICS */
 501
 502             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 503             ewrt             = _mm256_mul_pd(r10,ewtabscale);
 504             ewitab           = _mm256_cvttpd_epi32(ewrt);
 505             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 506             ewitab           = _mm_slli_epi32(ewitab,2);
 507             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 508             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 509             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 510             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 511             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 512             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 513             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 514             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(rinv10,velec));
 515             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
 516
 517             /* Update potential sum for this i atom from the interaction with this j atom. */
 518             velec            = _mm256_andnot_pd(dummy_mask,velec);
 519             velecsum         = _mm256_add_pd(velecsum,velec);
 520
 521             fscal            = felec;
 522
 523             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 524
 525             /* Calculate temporary vectorial force */
 526             tx               = _mm256_mul_pd(fscal,dx10);
 527             ty               = _mm256_mul_pd(fscal,dy10);
 528             tz               = _mm256_mul_pd(fscal,dz10);
 529
 530             /* Update vectorial force */
 531             fix1             = _mm256_add_pd(fix1,tx);
 532             fiy1             = _mm256_add_pd(fiy1,ty);
 533             fiz1             = _mm256_add_pd(fiz1,tz);
 534
 535             fjx0             = _mm256_add_pd(fjx0,tx);
 536             fjy0             = _mm256_add_pd(fjy0,ty);
 537             fjz0             = _mm256_add_pd(fjz0,tz);
 538
 539             /**************************
 540              * CALCULATE INTERACTIONS *
 541              **************************/
 542
 543             r20              = _mm256_mul_pd(rsq20,rinv20);
 544             r20              = _mm256_andnot_pd(dummy_mask,r20);
 545
 546             /* Compute parameters for interactions between i and j atoms */
 547             qq20             = _mm256_mul_pd(iq2,jq0);
 548
 549             /* EWALD ELECTROSTATICS */
 550
 551             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 552             ewrt             = _mm256_mul_pd(r20,ewtabscale);
 553             ewitab           = _mm256_cvttpd_epi32(ewrt);
 554             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 555             ewitab           = _mm_slli_epi32(ewitab,2);
 556             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 557             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 558             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 559             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 560             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 561             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 562             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 563             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(rinv20,velec));
 564             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
 565
 566             /* Update potential sum for this i atom from the interaction with this j atom. */
 567             velec            = _mm256_andnot_pd(dummy_mask,velec);
 568             velecsum         = _mm256_add_pd(velecsum,velec);
 569
 570             fscal            = felec;
 571
 572             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 573
 574             /* Calculate temporary vectorial force */
 575             tx               = _mm256_mul_pd(fscal,dx20);
 576             ty               = _mm256_mul_pd(fscal,dy20);
 577             tz               = _mm256_mul_pd(fscal,dz20);
 578
 579             /* Update vectorial force */
 580             fix2             = _mm256_add_pd(fix2,tx);
 581             fiy2             = _mm256_add_pd(fiy2,ty);
 582             fiz2             = _mm256_add_pd(fiz2,tz);
 583
 584             fjx0             = _mm256_add_pd(fjx0,tx);
 585             fjy0             = _mm256_add_pd(fjy0,ty);
 586             fjz0             = _mm256_add_pd(fjz0,tz);
 587
 588             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 589             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 590             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 591             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 592
 593             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 594
 595             /* Inner loop uses 129 flops */
 596         }
 597
 598         /* End of innermost loop */
 599
 600         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
 601                                                  f+i_coord_offset,fshift+i_shift_offset);
 602
 603         ggid                        = gid[iidx];
 604         /* Update potential energies */
 605         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 606
 607         /* Increment number of inner iterations */
 608         inneriter                  += j_index_end - j_index_start;
 609
 610         /* Outer loop uses 19 flops */
 611     }
 612
 613     /* Increment number of outer iterations */
 614     outeriter        += nri;
 615
 616     /* Update outer/inner flops */
 617
 618     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3_VF,outeriter*19 + inneriter*129);
 619 }
 620 /*
 621  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwNone_GeomW3P1_F_avx_256_double
 622  * Electrostatics interaction: Ewald
 623  * VdW interaction:            None
 624  * Geometry:                   Water3-Particle
 625  * Calculate force/pot:        Force
 626  */
 627 void
 628 nb_kernel_ElecEw_VdwNone_GeomW3P1_F_avx_256_double
 629                     (t_nblist                    * gmx_restrict       nlist,
 630                      rvec                        * gmx_restrict          xx,
 631                      rvec                        * gmx_restrict          ff,
 632                      t_forcerec                  * gmx_restrict          fr,
 633                      t_mdatoms                   * gmx_restrict     mdatoms,
 634                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 635                      t_nrnb                      * gmx_restrict        nrnb)
 636 {
 637     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 638      * just 0 for non-waters.
 639      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
 640      * jnr indices corresponding to data put in the four positions in the SIMD register.
 641      */
 642     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 643     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 644     int              jnrA,jnrB,jnrC,jnrD;
 645     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 646     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 647     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 648     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 649     real             rcutoff_scalar;
 650     real             *shiftvec,*fshift,*x,*f;
 651     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 652     real             scratch[4*DIM];
 653     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 654     real *           vdwioffsetptr0;
 655     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 656     real *           vdwioffsetptr1;
 657     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 658     real *           vdwioffsetptr2;
 659     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 660     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 661     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 662     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 663     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 664     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 665     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 666     real             *charge;
 667     __m128i          ewitab;
 668     __m256d          ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 669     __m256d          beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 670     real             *ewtab;
 671     __m256d          dummy_mask,cutoff_mask;
 672     __m128           tmpmask0,tmpmask1;
 673     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 674     __m256d          one     = _mm256_set1_pd(1.0);
 675     __m256d          two     = _mm256_set1_pd(2.0);
 676     x                = xx[0];
 677     f                = ff[0];
 678
 679     nri              = nlist->nri;
 680     iinr             = nlist->iinr;
 681     jindex           = nlist->jindex;
 682     jjnr             = nlist->jjnr;
 683     shiftidx         = nlist->shift;
 684     gid              = nlist->gid;
 685     shiftvec         = fr->shift_vec[0];
 686     fshift           = fr->fshift[0];
 687     facel            = _mm256_set1_pd(fr->epsfac);
 688     charge           = mdatoms->chargeA;
 689
 690     sh_ewald         = _mm256_set1_pd(fr->ic->sh_ewald);
 691     beta             = _mm256_set1_pd(fr->ic->ewaldcoeff_q);
 692     beta2            = _mm256_mul_pd(beta,beta);
 693     beta3            = _mm256_mul_pd(beta,beta2);
 694
 695     ewtab            = fr->ic->tabq_coul_F;
 696     ewtabscale       = _mm256_set1_pd(fr->ic->tabq_scale);
 697     ewtabhalfspace   = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
 698
 699     /* Setup water-specific parameters */
 700     inr              = nlist->iinr[0];
 701     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 702     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 703     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 704
 705     /* Avoid stupid compiler warnings */
 706     jnrA = jnrB = jnrC = jnrD = 0;
 707     j_coord_offsetA = 0;
 708     j_coord_offsetB = 0;
 709     j_coord_offsetC = 0;
 710     j_coord_offsetD = 0;
 711
 712     outeriter        = 0;
 713     inneriter        = 0;
 714
 715     for(iidx=0;iidx<4*DIM;iidx++)
 716     {
 717         scratch[iidx] = 0.0;
 718     }
 719
 720     /* Start outer loop over neighborlists */
 721     for(iidx=0; iidx<nri; iidx++)
 722     {
 723         /* Load shift vector for this list */
 724         i_shift_offset   = DIM*shiftidx[iidx];
 725
 726         /* Load limits for loop over neighbors */
 727         j_index_start    = jindex[iidx];
 728         j_index_end      = jindex[iidx+1];
 729
 730         /* Get outer coordinate index */
 731         inr              = iinr[iidx];
 732         i_coord_offset   = DIM*inr;
 733
 734         /* Load i particle coords and add shift vector */
 735         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 736                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 737
 738         fix0             = _mm256_setzero_pd();
 739         fiy0             = _mm256_setzero_pd();
 740         fiz0             = _mm256_setzero_pd();
 741         fix1             = _mm256_setzero_pd();
 742         fiy1             = _mm256_setzero_pd();
 743         fiz1             = _mm256_setzero_pd();
 744         fix2             = _mm256_setzero_pd();
 745         fiy2             = _mm256_setzero_pd();
 746         fiz2             = _mm256_setzero_pd();
 747
 748         /* Start inner kernel loop */
 749         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 750         {
 751
 752             /* Get j neighbor index, and coordinate index */
 753             jnrA             = jjnr[jidx];
 754             jnrB             = jjnr[jidx+1];
 755             jnrC             = jjnr[jidx+2];
 756             jnrD             = jjnr[jidx+3];
 757             j_coord_offsetA  = DIM*jnrA;
 758             j_coord_offsetB  = DIM*jnrB;
 759             j_coord_offsetC  = DIM*jnrC;
 760             j_coord_offsetD  = DIM*jnrD;
 761
 762             /* load j atom coordinates */
 763             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 764                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 765                                                  &jx0,&jy0,&jz0);
 766
 767             /* Calculate displacement vector */
 768             dx00             = _mm256_sub_pd(ix0,jx0);
 769             dy00             = _mm256_sub_pd(iy0,jy0);
 770             dz00             = _mm256_sub_pd(iz0,jz0);
 771             dx10             = _mm256_sub_pd(ix1,jx0);
 772             dy10             = _mm256_sub_pd(iy1,jy0);
 773             dz10             = _mm256_sub_pd(iz1,jz0);
 774             dx20             = _mm256_sub_pd(ix2,jx0);
 775             dy20             = _mm256_sub_pd(iy2,jy0);
 776             dz20             = _mm256_sub_pd(iz2,jz0);
 777
 778             /* Calculate squared distance and things based on it */
 779             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 780             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 781             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 782
 783             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 784             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 785             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 786
 787             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 788             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 789             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 790
 791             /* Load parameters for j particles */
 792             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 793                                                                  charge+jnrC+0,charge+jnrD+0);
 794
 795             fjx0             = _mm256_setzero_pd();
 796             fjy0             = _mm256_setzero_pd();
 797             fjz0             = _mm256_setzero_pd();
 798
 799             /**************************
 800              * CALCULATE INTERACTIONS *
 801              **************************/
 802
 803             r00              = _mm256_mul_pd(rsq00,rinv00);
 804
 805             /* Compute parameters for interactions between i and j atoms */
 806             qq00             = _mm256_mul_pd(iq0,jq0);
 807
 808             /* EWALD ELECTROSTATICS */
 809
 810             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 811             ewrt             = _mm256_mul_pd(r00,ewtabscale);
 812             ewitab           = _mm256_cvttpd_epi32(ewrt);
 813             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 814             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
 815                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
 816                                             &ewtabF,&ewtabFn);
 817             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
 818             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
 819
 820             fscal            = felec;
 821
 822             /* Calculate temporary vectorial force */
 823             tx               = _mm256_mul_pd(fscal,dx00);
 824             ty               = _mm256_mul_pd(fscal,dy00);
 825             tz               = _mm256_mul_pd(fscal,dz00);
 826
 827             /* Update vectorial force */
 828             fix0             = _mm256_add_pd(fix0,tx);
 829             fiy0             = _mm256_add_pd(fiy0,ty);
 830             fiz0             = _mm256_add_pd(fiz0,tz);
 831
 832             fjx0             = _mm256_add_pd(fjx0,tx);
 833             fjy0             = _mm256_add_pd(fjy0,ty);
 834             fjz0             = _mm256_add_pd(fjz0,tz);
 835
 836             /**************************
 837              * CALCULATE INTERACTIONS *
 838              **************************/
 839
 840             r10              = _mm256_mul_pd(rsq10,rinv10);
 841
 842             /* Compute parameters for interactions between i and j atoms */
 843             qq10             = _mm256_mul_pd(iq1,jq0);
 844
 845             /* EWALD ELECTROSTATICS */
 846
 847             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 848             ewrt             = _mm256_mul_pd(r10,ewtabscale);
 849             ewitab           = _mm256_cvttpd_epi32(ewrt);
 850             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 851             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
 852                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
 853                                             &ewtabF,&ewtabFn);
 854             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
 855             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
 856
 857             fscal            = felec;
 858
 859             /* Calculate temporary vectorial force */
 860             tx               = _mm256_mul_pd(fscal,dx10);
 861             ty               = _mm256_mul_pd(fscal,dy10);
 862             tz               = _mm256_mul_pd(fscal,dz10);
 863
 864             /* Update vectorial force */
 865             fix1             = _mm256_add_pd(fix1,tx);
 866             fiy1             = _mm256_add_pd(fiy1,ty);
 867             fiz1             = _mm256_add_pd(fiz1,tz);
 868
 869             fjx0             = _mm256_add_pd(fjx0,tx);
 870             fjy0             = _mm256_add_pd(fjy0,ty);
 871             fjz0             = _mm256_add_pd(fjz0,tz);
 872
 873             /**************************
 874              * CALCULATE INTERACTIONS *
 875              **************************/
 876
 877             r20              = _mm256_mul_pd(rsq20,rinv20);
 878
 879             /* Compute parameters for interactions between i and j atoms */
 880             qq20             = _mm256_mul_pd(iq2,jq0);
 881
 882             /* EWALD ELECTROSTATICS */
 883
 884             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 885             ewrt             = _mm256_mul_pd(r20,ewtabscale);
 886             ewitab           = _mm256_cvttpd_epi32(ewrt);
 887             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 888             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
 889                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
 890                                             &ewtabF,&ewtabFn);
 891             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
 892             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
 893
 894             fscal            = felec;
 895
 896             /* Calculate temporary vectorial force */
 897             tx               = _mm256_mul_pd(fscal,dx20);
 898             ty               = _mm256_mul_pd(fscal,dy20);
 899             tz               = _mm256_mul_pd(fscal,dz20);
 900
 901             /* Update vectorial force */
 902             fix2             = _mm256_add_pd(fix2,tx);
 903             fiy2             = _mm256_add_pd(fiy2,ty);
 904             fiz2             = _mm256_add_pd(fiz2,tz);
 905
 906             fjx0             = _mm256_add_pd(fjx0,tx);
 907             fjy0             = _mm256_add_pd(fjy0,ty);
 908             fjz0             = _mm256_add_pd(fjz0,tz);
 909
 910             fjptrA             = f+j_coord_offsetA;
 911             fjptrB             = f+j_coord_offsetB;
 912             fjptrC             = f+j_coord_offsetC;
 913             fjptrD             = f+j_coord_offsetD;
 914
 915             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 916
 917             /* Inner loop uses 111 flops */
 918         }
 919
 920         if(jidx<j_index_end)
 921         {
 922
 923             /* Get j neighbor index, and coordinate index */
 924             jnrlistA         = jjnr[jidx];
 925             jnrlistB         = jjnr[jidx+1];
 926             jnrlistC         = jjnr[jidx+2];
 927             jnrlistD         = jjnr[jidx+3];
 928             /* Sign of each element will be negative for non-real atoms.
 929              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 930              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 931              */
 932             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 933
 934             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 935             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 936             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 937
 938             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 939             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 940             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 941             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 942             j_coord_offsetA  = DIM*jnrA;
 943             j_coord_offsetB  = DIM*jnrB;
 944             j_coord_offsetC  = DIM*jnrC;
 945             j_coord_offsetD  = DIM*jnrD;
 946
 947             /* load j atom coordinates */
 948             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 949                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 950                                                  &jx0,&jy0,&jz0);
 951
 952             /* Calculate displacement vector */
 953             dx00             = _mm256_sub_pd(ix0,jx0);
 954             dy00             = _mm256_sub_pd(iy0,jy0);
 955             dz00             = _mm256_sub_pd(iz0,jz0);
 956             dx10             = _mm256_sub_pd(ix1,jx0);
 957             dy10             = _mm256_sub_pd(iy1,jy0);
 958             dz10             = _mm256_sub_pd(iz1,jz0);
 959             dx20             = _mm256_sub_pd(ix2,jx0);
 960             dy20             = _mm256_sub_pd(iy2,jy0);
 961             dz20             = _mm256_sub_pd(iz2,jz0);
 962
 963             /* Calculate squared distance and things based on it */
 964             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 965             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 966             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 967
 968             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 969             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 970             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 971
 972             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 973             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 974             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 975
 976             /* Load parameters for j particles */
 977             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 978                                                                  charge+jnrC+0,charge+jnrD+0);
 979
 980             fjx0             = _mm256_setzero_pd();
 981             fjy0             = _mm256_setzero_pd();
 982             fjz0             = _mm256_setzero_pd();
 983
 984             /**************************
 985              * CALCULATE INTERACTIONS *
 986              **************************/
 987
 988             r00              = _mm256_mul_pd(rsq00,rinv00);
 989             r00              = _mm256_andnot_pd(dummy_mask,r00);
 990
 991             /* Compute parameters for interactions between i and j atoms */
 992             qq00             = _mm256_mul_pd(iq0,jq0);
 993
 994             /* EWALD ELECTROSTATICS */
 995
 996             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 997             ewrt             = _mm256_mul_pd(r00,ewtabscale);
 998             ewitab           = _mm256_cvttpd_epi32(ewrt);
 999             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1000             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1001                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1002                                             &ewtabF,&ewtabFn);
1003             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1004             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
1005
1006             fscal            = felec;
1007
1008             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1009
1010             /* Calculate temporary vectorial force */
1011             tx               = _mm256_mul_pd(fscal,dx00);
1012             ty               = _mm256_mul_pd(fscal,dy00);
1013             tz               = _mm256_mul_pd(fscal,dz00);
1014
1015             /* Update vectorial force */
1016             fix0             = _mm256_add_pd(fix0,tx);
1017             fiy0             = _mm256_add_pd(fiy0,ty);
1018             fiz0             = _mm256_add_pd(fiz0,tz);
1019
1020             fjx0             = _mm256_add_pd(fjx0,tx);
1021             fjy0             = _mm256_add_pd(fjy0,ty);
1022             fjz0             = _mm256_add_pd(fjz0,tz);
1023
1024             /**************************
1025              * CALCULATE INTERACTIONS *
1026              **************************/
1027
1028             r10              = _mm256_mul_pd(rsq10,rinv10);
1029             r10              = _mm256_andnot_pd(dummy_mask,r10);
1030
1031             /* Compute parameters for interactions between i and j atoms */
1032             qq10             = _mm256_mul_pd(iq1,jq0);
1033
1034             /* EWALD ELECTROSTATICS */
1035
1036             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1037             ewrt             = _mm256_mul_pd(r10,ewtabscale);
1038             ewitab           = _mm256_cvttpd_epi32(ewrt);
1039             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1040             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1041                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1042                                             &ewtabF,&ewtabFn);
1043             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1044             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
1045
1046             fscal            = felec;
1047
1048             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1049
1050             /* Calculate temporary vectorial force */
1051             tx               = _mm256_mul_pd(fscal,dx10);
1052             ty               = _mm256_mul_pd(fscal,dy10);
1053             tz               = _mm256_mul_pd(fscal,dz10);
1054
1055             /* Update vectorial force */
1056             fix1             = _mm256_add_pd(fix1,tx);
1057             fiy1             = _mm256_add_pd(fiy1,ty);
1058             fiz1             = _mm256_add_pd(fiz1,tz);
1059
1060             fjx0             = _mm256_add_pd(fjx0,tx);
1061             fjy0             = _mm256_add_pd(fjy0,ty);
1062             fjz0             = _mm256_add_pd(fjz0,tz);
1063
1064             /**************************
1065              * CALCULATE INTERACTIONS *
1066              **************************/
1067
1068             r20              = _mm256_mul_pd(rsq20,rinv20);
1069             r20              = _mm256_andnot_pd(dummy_mask,r20);
1070
1071             /* Compute parameters for interactions between i and j atoms */
1072             qq20             = _mm256_mul_pd(iq2,jq0);
1073
1074             /* EWALD ELECTROSTATICS */
1075
1076             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1077             ewrt             = _mm256_mul_pd(r20,ewtabscale);
1078             ewitab           = _mm256_cvttpd_epi32(ewrt);
1079             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1080             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1081                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1082                                             &ewtabF,&ewtabFn);
1083             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1084             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
1085
1086             fscal            = felec;
1087
1088             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1089
1090             /* Calculate temporary vectorial force */
1091             tx               = _mm256_mul_pd(fscal,dx20);
1092             ty               = _mm256_mul_pd(fscal,dy20);
1093             tz               = _mm256_mul_pd(fscal,dz20);
1094
1095             /* Update vectorial force */
1096             fix2             = _mm256_add_pd(fix2,tx);
1097             fiy2             = _mm256_add_pd(fiy2,ty);
1098             fiz2             = _mm256_add_pd(fiz2,tz);
1099
1100             fjx0             = _mm256_add_pd(fjx0,tx);
1101             fjy0             = _mm256_add_pd(fjy0,ty);
1102             fjz0             = _mm256_add_pd(fjz0,tz);
1103
1104             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1105             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1106             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1107             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1108
1109             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1110
1111             /* Inner loop uses 114 flops */
1112         }
1113
1114         /* End of innermost loop */
1115
1116         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1117                                                  f+i_coord_offset,fshift+i_shift_offset);
1118
1119         /* Increment number of inner iterations */
1120         inneriter                  += j_index_end - j_index_start;
1121
1122         /* Outer loop uses 18 flops */
1123     }
1124
1125     /* Increment number of outer iterations */
1126     outeriter        += nri;
1127
1128     /* Update outer/inner flops */
1129
1130     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3_F,outeriter*18 + inneriter*114);
1131 }