src/gromacs/gmxlib/nonbonded/nb_kernel_avx_128_fma_double/nb_kernel_ElecEw_VdwNone_GeomW4P1_avx_128_fma_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_128_fma_double kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_128_fma_double.h"
  48 #include "kernelutil_x86_avx_128_fma_double.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwNone_GeomW4P1_VF_avx_128_fma_double
  52  * Electrostatics interaction: Ewald
  53  * VdW interaction:            None
  54  * Geometry:                   Water4-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecEw_VdwNone_GeomW4P1_VF_avx_128_fma_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 refer to j loop unrolling done with SSE double precision, e.g. for the two 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;
  75     int              j_coord_offsetA,j_coord_offsetB;
  76     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  77     real             rcutoff_scalar;
  78     real             *shiftvec,*fshift,*x,*f;
  79     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  80     int              vdwioffset1;
  81     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  82     int              vdwioffset2;
  83     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  84     int              vdwioffset3;
  85     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  86     int              vdwjidx0A,vdwjidx0B;
  87     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  88     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  89     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  90     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  91     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
  92     real             *charge;
  93     __m128i          ewitab;
  94     __m128d          ewtabscale,eweps,twoeweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
  95     real             *ewtab;
  96     __m128d          dummy_mask,cutoff_mask;
  97     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
  98     __m128d          one     = _mm_set1_pd(1.0);
  99     __m128d          two     = _mm_set1_pd(2.0);
 100     x                = xx[0];
 101     f                = ff[0];
 102
 103     nri              = nlist->nri;
 104     iinr             = nlist->iinr;
 105     jindex           = nlist->jindex;
 106     jjnr             = nlist->jjnr;
 107     shiftidx         = nlist->shift;
 108     gid              = nlist->gid;
 109     shiftvec         = fr->shift_vec[0];
 110     fshift           = fr->fshift[0];
 111     facel            = _mm_set1_pd(fr->epsfac);
 112     charge           = mdatoms->chargeA;
 113
 114     sh_ewald         = _mm_set1_pd(fr->ic->sh_ewald);
 115     ewtab            = fr->ic->tabq_coul_FDV0;
 116     ewtabscale       = _mm_set1_pd(fr->ic->tabq_scale);
 117     ewtabhalfspace   = _mm_set1_pd(0.5/fr->ic->tabq_scale);
 118
 119     /* Setup water-specific parameters */
 120     inr              = nlist->iinr[0];
 121     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 122     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 123     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 124
 125     /* Avoid stupid compiler warnings */
 126     jnrA = jnrB = 0;
 127     j_coord_offsetA = 0;
 128     j_coord_offsetB = 0;
 129
 130     outeriter        = 0;
 131     inneriter        = 0;
 132
 133     /* Start outer loop over neighborlists */
 134     for(iidx=0; iidx<nri; iidx++)
 135     {
 136         /* Load shift vector for this list */
 137         i_shift_offset   = DIM*shiftidx[iidx];
 138
 139         /* Load limits for loop over neighbors */
 140         j_index_start    = jindex[iidx];
 141         j_index_end      = jindex[iidx+1];
 142
 143         /* Get outer coordinate index */
 144         inr              = iinr[iidx];
 145         i_coord_offset   = DIM*inr;
 146
 147         /* Load i particle coords and add shift vector */
 148         gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 149                                                  &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 150
 151         fix1             = _mm_setzero_pd();
 152         fiy1             = _mm_setzero_pd();
 153         fiz1             = _mm_setzero_pd();
 154         fix2             = _mm_setzero_pd();
 155         fiy2             = _mm_setzero_pd();
 156         fiz2             = _mm_setzero_pd();
 157         fix3             = _mm_setzero_pd();
 158         fiy3             = _mm_setzero_pd();
 159         fiz3             = _mm_setzero_pd();
 160
 161         /* Reset potential sums */
 162         velecsum         = _mm_setzero_pd();
 163
 164         /* Start inner kernel loop */
 165         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 166         {
 167
 168             /* Get j neighbor index, and coordinate index */
 169             jnrA             = jjnr[jidx];
 170             jnrB             = jjnr[jidx+1];
 171             j_coord_offsetA  = DIM*jnrA;
 172             j_coord_offsetB  = DIM*jnrB;
 173
 174             /* load j atom coordinates */
 175             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 176                                               &jx0,&jy0,&jz0);
 177
 178             /* Calculate displacement vector */
 179             dx10             = _mm_sub_pd(ix1,jx0);
 180             dy10             = _mm_sub_pd(iy1,jy0);
 181             dz10             = _mm_sub_pd(iz1,jz0);
 182             dx20             = _mm_sub_pd(ix2,jx0);
 183             dy20             = _mm_sub_pd(iy2,jy0);
 184             dz20             = _mm_sub_pd(iz2,jz0);
 185             dx30             = _mm_sub_pd(ix3,jx0);
 186             dy30             = _mm_sub_pd(iy3,jy0);
 187             dz30             = _mm_sub_pd(iz3,jz0);
 188
 189             /* Calculate squared distance and things based on it */
 190             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 191             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 192             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 193
 194             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 195             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 196             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 197
 198             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 199             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 200             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 201
 202             /* Load parameters for j particles */
 203             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 204
 205             fjx0             = _mm_setzero_pd();
 206             fjy0             = _mm_setzero_pd();
 207             fjz0             = _mm_setzero_pd();
 208
 209             /**************************
 210              * CALCULATE INTERACTIONS *
 211              **************************/
 212
 213             r10              = _mm_mul_pd(rsq10,rinv10);
 214
 215             /* Compute parameters for interactions between i and j atoms */
 216             qq10             = _mm_mul_pd(iq1,jq0);
 217
 218             /* EWALD ELECTROSTATICS */
 219
 220             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 221             ewrt             = _mm_mul_pd(r10,ewtabscale);
 222             ewitab           = _mm_cvttpd_epi32(ewrt);
 223 #ifdef __XOP__
 224             eweps            = _mm_frcz_pd(ewrt);
 225 #else
 226             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 227 #endif
 228             twoeweps         = _mm_add_pd(eweps,eweps);
 229             ewitab           = _mm_slli_epi32(ewitab,2);
 230             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 231             ewtabD           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 232             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 233             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 234             ewtabFn          = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,1) +2);
 235             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 236             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 237             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 238             velec            = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
 239             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
 240
 241             /* Update potential sum for this i atom from the interaction with this j atom. */
 242             velecsum         = _mm_add_pd(velecsum,velec);
 243
 244             fscal            = felec;
 245
 246             /* Update vectorial force */
 247             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 248             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 249             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 250
 251             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 252             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 253             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 254
 255             /**************************
 256              * CALCULATE INTERACTIONS *
 257              **************************/
 258
 259             r20              = _mm_mul_pd(rsq20,rinv20);
 260
 261             /* Compute parameters for interactions between i and j atoms */
 262             qq20             = _mm_mul_pd(iq2,jq0);
 263
 264             /* EWALD ELECTROSTATICS */
 265
 266             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 267             ewrt             = _mm_mul_pd(r20,ewtabscale);
 268             ewitab           = _mm_cvttpd_epi32(ewrt);
 269 #ifdef __XOP__
 270             eweps            = _mm_frcz_pd(ewrt);
 271 #else
 272             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 273 #endif
 274             twoeweps         = _mm_add_pd(eweps,eweps);
 275             ewitab           = _mm_slli_epi32(ewitab,2);
 276             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 277             ewtabD           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 278             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 279             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 280             ewtabFn          = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,1) +2);
 281             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 282             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 283             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 284             velec            = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
 285             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
 286
 287             /* Update potential sum for this i atom from the interaction with this j atom. */
 288             velecsum         = _mm_add_pd(velecsum,velec);
 289
 290             fscal            = felec;
 291
 292             /* Update vectorial force */
 293             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 294             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 295             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 296
 297             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 298             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 299             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 300
 301             /**************************
 302              * CALCULATE INTERACTIONS *
 303              **************************/
 304
 305             r30              = _mm_mul_pd(rsq30,rinv30);
 306
 307             /* Compute parameters for interactions between i and j atoms */
 308             qq30             = _mm_mul_pd(iq3,jq0);
 309
 310             /* EWALD ELECTROSTATICS */
 311
 312             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 313             ewrt             = _mm_mul_pd(r30,ewtabscale);
 314             ewitab           = _mm_cvttpd_epi32(ewrt);
 315 #ifdef __XOP__
 316             eweps            = _mm_frcz_pd(ewrt);
 317 #else
 318             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 319 #endif
 320             twoeweps         = _mm_add_pd(eweps,eweps);
 321             ewitab           = _mm_slli_epi32(ewitab,2);
 322             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 323             ewtabD           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 324             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 325             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 326             ewtabFn          = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,1) +2);
 327             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 328             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 329             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 330             velec            = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
 331             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
 332
 333             /* Update potential sum for this i atom from the interaction with this j atom. */
 334             velecsum         = _mm_add_pd(velecsum,velec);
 335
 336             fscal            = felec;
 337
 338             /* Update vectorial force */
 339             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 340             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 341             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 342
 343             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 344             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 345             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 346
 347             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 348
 349             /* Inner loop uses 135 flops */
 350         }
 351
 352         if(jidx<j_index_end)
 353         {
 354
 355             jnrA             = jjnr[jidx];
 356             j_coord_offsetA  = DIM*jnrA;
 357
 358             /* load j atom coordinates */
 359             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 360                                               &jx0,&jy0,&jz0);
 361
 362             /* Calculate displacement vector */
 363             dx10             = _mm_sub_pd(ix1,jx0);
 364             dy10             = _mm_sub_pd(iy1,jy0);
 365             dz10             = _mm_sub_pd(iz1,jz0);
 366             dx20             = _mm_sub_pd(ix2,jx0);
 367             dy20             = _mm_sub_pd(iy2,jy0);
 368             dz20             = _mm_sub_pd(iz2,jz0);
 369             dx30             = _mm_sub_pd(ix3,jx0);
 370             dy30             = _mm_sub_pd(iy3,jy0);
 371             dz30             = _mm_sub_pd(iz3,jz0);
 372
 373             /* Calculate squared distance and things based on it */
 374             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 375             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 376             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 377
 378             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 379             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 380             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 381
 382             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 383             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 384             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 385
 386             /* Load parameters for j particles */
 387             jq0              = _mm_load_sd(charge+jnrA+0);
 388
 389             fjx0             = _mm_setzero_pd();
 390             fjy0             = _mm_setzero_pd();
 391             fjz0             = _mm_setzero_pd();
 392
 393             /**************************
 394              * CALCULATE INTERACTIONS *
 395              **************************/
 396
 397             r10              = _mm_mul_pd(rsq10,rinv10);
 398
 399             /* Compute parameters for interactions between i and j atoms */
 400             qq10             = _mm_mul_pd(iq1,jq0);
 401
 402             /* EWALD ELECTROSTATICS */
 403
 404             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 405             ewrt             = _mm_mul_pd(r10,ewtabscale);
 406             ewitab           = _mm_cvttpd_epi32(ewrt);
 407 #ifdef __XOP__
 408             eweps            = _mm_frcz_pd(ewrt);
 409 #else
 410             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 411 #endif
 412             twoeweps         = _mm_add_pd(eweps,eweps);
 413             ewitab           = _mm_slli_epi32(ewitab,2);
 414             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 415             ewtabD           = _mm_setzero_pd();
 416             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 417             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 418             ewtabFn          = _mm_setzero_pd();
 419             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 420             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 421             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 422             velec            = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
 423             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
 424
 425             /* Update potential sum for this i atom from the interaction with this j atom. */
 426             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 427             velecsum         = _mm_add_pd(velecsum,velec);
 428
 429             fscal            = felec;
 430
 431             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 432
 433             /* Update vectorial force */
 434             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 435             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 436             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 437
 438             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 439             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 440             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 441
 442             /**************************
 443              * CALCULATE INTERACTIONS *
 444              **************************/
 445
 446             r20              = _mm_mul_pd(rsq20,rinv20);
 447
 448             /* Compute parameters for interactions between i and j atoms */
 449             qq20             = _mm_mul_pd(iq2,jq0);
 450
 451             /* EWALD ELECTROSTATICS */
 452
 453             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 454             ewrt             = _mm_mul_pd(r20,ewtabscale);
 455             ewitab           = _mm_cvttpd_epi32(ewrt);
 456 #ifdef __XOP__
 457             eweps            = _mm_frcz_pd(ewrt);
 458 #else
 459             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 460 #endif
 461             twoeweps         = _mm_add_pd(eweps,eweps);
 462             ewitab           = _mm_slli_epi32(ewitab,2);
 463             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 464             ewtabD           = _mm_setzero_pd();
 465             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 466             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 467             ewtabFn          = _mm_setzero_pd();
 468             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 469             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 470             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 471             velec            = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
 472             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
 473
 474             /* Update potential sum for this i atom from the interaction with this j atom. */
 475             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 476             velecsum         = _mm_add_pd(velecsum,velec);
 477
 478             fscal            = felec;
 479
 480             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 481
 482             /* Update vectorial force */
 483             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 484             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 485             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 486
 487             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 488             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 489             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 490
 491             /**************************
 492              * CALCULATE INTERACTIONS *
 493              **************************/
 494
 495             r30              = _mm_mul_pd(rsq30,rinv30);
 496
 497             /* Compute parameters for interactions between i and j atoms */
 498             qq30             = _mm_mul_pd(iq3,jq0);
 499
 500             /* EWALD ELECTROSTATICS */
 501
 502             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 503             ewrt             = _mm_mul_pd(r30,ewtabscale);
 504             ewitab           = _mm_cvttpd_epi32(ewrt);
 505 #ifdef __XOP__
 506             eweps            = _mm_frcz_pd(ewrt);
 507 #else
 508             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 509 #endif
 510             twoeweps         = _mm_add_pd(eweps,eweps);
 511             ewitab           = _mm_slli_epi32(ewitab,2);
 512             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 513             ewtabD           = _mm_setzero_pd();
 514             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 515             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 516             ewtabFn          = _mm_setzero_pd();
 517             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 518             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 519             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 520             velec            = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
 521             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
 522
 523             /* Update potential sum for this i atom from the interaction with this j atom. */
 524             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 525             velecsum         = _mm_add_pd(velecsum,velec);
 526
 527             fscal            = felec;
 528
 529             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 530
 531             /* Update vectorial force */
 532             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 533             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 534             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 535
 536             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 537             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 538             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 539
 540             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 541
 542             /* Inner loop uses 135 flops */
 543         }
 544
 545         /* End of innermost loop */
 546
 547         gmx_mm_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 548                                               f+i_coord_offset+DIM,fshift+i_shift_offset);
 549
 550         ggid                        = gid[iidx];
 551         /* Update potential energies */
 552         gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 553
 554         /* Increment number of inner iterations */
 555         inneriter                  += j_index_end - j_index_start;
 556
 557         /* Outer loop uses 19 flops */
 558     }
 559
 560     /* Increment number of outer iterations */
 561     outeriter        += nri;
 562
 563     /* Update outer/inner flops */
 564
 565     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_VF,outeriter*19 + inneriter*135);
 566 }
 567 /*
 568  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwNone_GeomW4P1_F_avx_128_fma_double
 569  * Electrostatics interaction: Ewald
 570  * VdW interaction:            None
 571  * Geometry:                   Water4-Particle
 572  * Calculate force/pot:        Force
 573  */
 574 void
 575 nb_kernel_ElecEw_VdwNone_GeomW4P1_F_avx_128_fma_double
 576                     (t_nblist                    * gmx_restrict       nlist,
 577                      rvec                        * gmx_restrict          xx,
 578                      rvec                        * gmx_restrict          ff,
 579                      t_forcerec                  * gmx_restrict          fr,
 580                      t_mdatoms                   * gmx_restrict     mdatoms,
 581                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 582                      t_nrnb                      * gmx_restrict        nrnb)
 583 {
 584     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 585      * just 0 for non-waters.
 586      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
 587      * jnr indices corresponding to data put in the four positions in the SIMD register.
 588      */
 589     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 590     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 591     int              jnrA,jnrB;
 592     int              j_coord_offsetA,j_coord_offsetB;
 593     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 594     real             rcutoff_scalar;
 595     real             *shiftvec,*fshift,*x,*f;
 596     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 597     int              vdwioffset1;
 598     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 599     int              vdwioffset2;
 600     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 601     int              vdwioffset3;
 602     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 603     int              vdwjidx0A,vdwjidx0B;
 604     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 605     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 606     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 607     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 608     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 609     real             *charge;
 610     __m128i          ewitab;
 611     __m128d          ewtabscale,eweps,twoeweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 612     real             *ewtab;
 613     __m128d          dummy_mask,cutoff_mask;
 614     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 615     __m128d          one     = _mm_set1_pd(1.0);
 616     __m128d          two     = _mm_set1_pd(2.0);
 617     x                = xx[0];
 618     f                = ff[0];
 619
 620     nri              = nlist->nri;
 621     iinr             = nlist->iinr;
 622     jindex           = nlist->jindex;
 623     jjnr             = nlist->jjnr;
 624     shiftidx         = nlist->shift;
 625     gid              = nlist->gid;
 626     shiftvec         = fr->shift_vec[0];
 627     fshift           = fr->fshift[0];
 628     facel            = _mm_set1_pd(fr->epsfac);
 629     charge           = mdatoms->chargeA;
 630
 631     sh_ewald         = _mm_set1_pd(fr->ic->sh_ewald);
 632     ewtab            = fr->ic->tabq_coul_F;
 633     ewtabscale       = _mm_set1_pd(fr->ic->tabq_scale);
 634     ewtabhalfspace   = _mm_set1_pd(0.5/fr->ic->tabq_scale);
 635
 636     /* Setup water-specific parameters */
 637     inr              = nlist->iinr[0];
 638     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 639     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 640     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 641
 642     /* Avoid stupid compiler warnings */
 643     jnrA = jnrB = 0;
 644     j_coord_offsetA = 0;
 645     j_coord_offsetB = 0;
 646
 647     outeriter        = 0;
 648     inneriter        = 0;
 649
 650     /* Start outer loop over neighborlists */
 651     for(iidx=0; iidx<nri; iidx++)
 652     {
 653         /* Load shift vector for this list */
 654         i_shift_offset   = DIM*shiftidx[iidx];
 655
 656         /* Load limits for loop over neighbors */
 657         j_index_start    = jindex[iidx];
 658         j_index_end      = jindex[iidx+1];
 659
 660         /* Get outer coordinate index */
 661         inr              = iinr[iidx];
 662         i_coord_offset   = DIM*inr;
 663
 664         /* Load i particle coords and add shift vector */
 665         gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 666                                                  &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 667
 668         fix1             = _mm_setzero_pd();
 669         fiy1             = _mm_setzero_pd();
 670         fiz1             = _mm_setzero_pd();
 671         fix2             = _mm_setzero_pd();
 672         fiy2             = _mm_setzero_pd();
 673         fiz2             = _mm_setzero_pd();
 674         fix3             = _mm_setzero_pd();
 675         fiy3             = _mm_setzero_pd();
 676         fiz3             = _mm_setzero_pd();
 677
 678         /* Start inner kernel loop */
 679         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 680         {
 681
 682             /* Get j neighbor index, and coordinate index */
 683             jnrA             = jjnr[jidx];
 684             jnrB             = jjnr[jidx+1];
 685             j_coord_offsetA  = DIM*jnrA;
 686             j_coord_offsetB  = DIM*jnrB;
 687
 688             /* load j atom coordinates */
 689             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 690                                               &jx0,&jy0,&jz0);
 691
 692             /* Calculate displacement vector */
 693             dx10             = _mm_sub_pd(ix1,jx0);
 694             dy10             = _mm_sub_pd(iy1,jy0);
 695             dz10             = _mm_sub_pd(iz1,jz0);
 696             dx20             = _mm_sub_pd(ix2,jx0);
 697             dy20             = _mm_sub_pd(iy2,jy0);
 698             dz20             = _mm_sub_pd(iz2,jz0);
 699             dx30             = _mm_sub_pd(ix3,jx0);
 700             dy30             = _mm_sub_pd(iy3,jy0);
 701             dz30             = _mm_sub_pd(iz3,jz0);
 702
 703             /* Calculate squared distance and things based on it */
 704             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 705             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 706             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 707
 708             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 709             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 710             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 711
 712             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 713             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 714             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 715
 716             /* Load parameters for j particles */
 717             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 718
 719             fjx0             = _mm_setzero_pd();
 720             fjy0             = _mm_setzero_pd();
 721             fjz0             = _mm_setzero_pd();
 722
 723             /**************************
 724              * CALCULATE INTERACTIONS *
 725              **************************/
 726
 727             r10              = _mm_mul_pd(rsq10,rinv10);
 728
 729             /* Compute parameters for interactions between i and j atoms */
 730             qq10             = _mm_mul_pd(iq1,jq0);
 731
 732             /* EWALD ELECTROSTATICS */
 733
 734             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 735             ewrt             = _mm_mul_pd(r10,ewtabscale);
 736             ewitab           = _mm_cvttpd_epi32(ewrt);
 737 #ifdef __XOP__
 738             eweps            = _mm_frcz_pd(ewrt);
 739 #else
 740             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 741 #endif
 742             twoeweps         = _mm_add_pd(eweps,eweps);
 743             gmx_mm_load_2pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),ewtab+_mm_extract_epi32(ewitab,1),
 744                                          &ewtabF,&ewtabFn);
 745             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
 746             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
 747
 748             fscal            = felec;
 749
 750             /* Update vectorial force */
 751             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 752             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 753             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 754
 755             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 756             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 757             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 758
 759             /**************************
 760              * CALCULATE INTERACTIONS *
 761              **************************/
 762
 763             r20              = _mm_mul_pd(rsq20,rinv20);
 764
 765             /* Compute parameters for interactions between i and j atoms */
 766             qq20             = _mm_mul_pd(iq2,jq0);
 767
 768             /* EWALD ELECTROSTATICS */
 769
 770             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 771             ewrt             = _mm_mul_pd(r20,ewtabscale);
 772             ewitab           = _mm_cvttpd_epi32(ewrt);
 773 #ifdef __XOP__
 774             eweps            = _mm_frcz_pd(ewrt);
 775 #else
 776             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 777 #endif
 778             twoeweps         = _mm_add_pd(eweps,eweps);
 779             gmx_mm_load_2pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),ewtab+_mm_extract_epi32(ewitab,1),
 780                                          &ewtabF,&ewtabFn);
 781             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
 782             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
 783
 784             fscal            = felec;
 785
 786             /* Update vectorial force */
 787             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 788             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 789             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 790
 791             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 792             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 793             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 794
 795             /**************************
 796              * CALCULATE INTERACTIONS *
 797              **************************/
 798
 799             r30              = _mm_mul_pd(rsq30,rinv30);
 800
 801             /* Compute parameters for interactions between i and j atoms */
 802             qq30             = _mm_mul_pd(iq3,jq0);
 803
 804             /* EWALD ELECTROSTATICS */
 805
 806             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 807             ewrt             = _mm_mul_pd(r30,ewtabscale);
 808             ewitab           = _mm_cvttpd_epi32(ewrt);
 809 #ifdef __XOP__
 810             eweps            = _mm_frcz_pd(ewrt);
 811 #else
 812             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 813 #endif
 814             twoeweps         = _mm_add_pd(eweps,eweps);
 815             gmx_mm_load_2pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),ewtab+_mm_extract_epi32(ewitab,1),
 816                                          &ewtabF,&ewtabFn);
 817             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
 818             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
 819
 820             fscal            = felec;
 821
 822             /* Update vectorial force */
 823             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 824             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 825             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 826
 827             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 828             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 829             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 830
 831             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 832
 833             /* Inner loop uses 120 flops */
 834         }
 835
 836         if(jidx<j_index_end)
 837         {
 838
 839             jnrA             = jjnr[jidx];
 840             j_coord_offsetA  = DIM*jnrA;
 841
 842             /* load j atom coordinates */
 843             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 844                                               &jx0,&jy0,&jz0);
 845
 846             /* Calculate displacement vector */
 847             dx10             = _mm_sub_pd(ix1,jx0);
 848             dy10             = _mm_sub_pd(iy1,jy0);
 849             dz10             = _mm_sub_pd(iz1,jz0);
 850             dx20             = _mm_sub_pd(ix2,jx0);
 851             dy20             = _mm_sub_pd(iy2,jy0);
 852             dz20             = _mm_sub_pd(iz2,jz0);
 853             dx30             = _mm_sub_pd(ix3,jx0);
 854             dy30             = _mm_sub_pd(iy3,jy0);
 855             dz30             = _mm_sub_pd(iz3,jz0);
 856
 857             /* Calculate squared distance and things based on it */
 858             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 859             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 860             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 861
 862             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 863             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 864             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 865
 866             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 867             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 868             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 869
 870             /* Load parameters for j particles */
 871             jq0              = _mm_load_sd(charge+jnrA+0);
 872
 873             fjx0             = _mm_setzero_pd();
 874             fjy0             = _mm_setzero_pd();
 875             fjz0             = _mm_setzero_pd();
 876
 877             /**************************
 878              * CALCULATE INTERACTIONS *
 879              **************************/
 880
 881             r10              = _mm_mul_pd(rsq10,rinv10);
 882
 883             /* Compute parameters for interactions between i and j atoms */
 884             qq10             = _mm_mul_pd(iq1,jq0);
 885
 886             /* EWALD ELECTROSTATICS */
 887
 888             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 889             ewrt             = _mm_mul_pd(r10,ewtabscale);
 890             ewitab           = _mm_cvttpd_epi32(ewrt);
 891 #ifdef __XOP__
 892             eweps            = _mm_frcz_pd(ewrt);
 893 #else
 894             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 895 #endif
 896             twoeweps         = _mm_add_pd(eweps,eweps);
 897             gmx_mm_load_1pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
 898             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
 899             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
 900
 901             fscal            = felec;
 902
 903             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 904
 905             /* Update vectorial force */
 906             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 907             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 908             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 909
 910             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 911             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 912             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 913
 914             /**************************
 915              * CALCULATE INTERACTIONS *
 916              **************************/
 917
 918             r20              = _mm_mul_pd(rsq20,rinv20);
 919
 920             /* Compute parameters for interactions between i and j atoms */
 921             qq20             = _mm_mul_pd(iq2,jq0);
 922
 923             /* EWALD ELECTROSTATICS */
 924
 925             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 926             ewrt             = _mm_mul_pd(r20,ewtabscale);
 927             ewitab           = _mm_cvttpd_epi32(ewrt);
 928 #ifdef __XOP__
 929             eweps            = _mm_frcz_pd(ewrt);
 930 #else
 931             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 932 #endif
 933             twoeweps         = _mm_add_pd(eweps,eweps);
 934             gmx_mm_load_1pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
 935             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
 936             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
 937
 938             fscal            = felec;
 939
 940             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 941
 942             /* Update vectorial force */
 943             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 944             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 945             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 946
 947             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 948             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 949             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 950
 951             /**************************
 952              * CALCULATE INTERACTIONS *
 953              **************************/
 954
 955             r30              = _mm_mul_pd(rsq30,rinv30);
 956
 957             /* Compute parameters for interactions between i and j atoms */
 958             qq30             = _mm_mul_pd(iq3,jq0);
 959
 960             /* EWALD ELECTROSTATICS */
 961
 962             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 963             ewrt             = _mm_mul_pd(r30,ewtabscale);
 964             ewitab           = _mm_cvttpd_epi32(ewrt);
 965 #ifdef __XOP__
 966             eweps            = _mm_frcz_pd(ewrt);
 967 #else
 968             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 969 #endif
 970             twoeweps         = _mm_add_pd(eweps,eweps);
 971             gmx_mm_load_1pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
 972             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
 973             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
 974
 975             fscal            = felec;
 976
 977             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 978
 979             /* Update vectorial force */
 980             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 981             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 982             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 983
 984             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 985             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 986             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 987
 988             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 989
 990             /* Inner loop uses 120 flops */
 991         }
 992
 993         /* End of innermost loop */
 994
 995         gmx_mm_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 996                                               f+i_coord_offset+DIM,fshift+i_shift_offset);
 997
 998         /* Increment number of inner iterations */
 999         inneriter                  += j_index_end - j_index_start;
1000
1001         /* Outer loop uses 18 flops */
1002     }
1003
1004     /* Increment number of outer iterations */
1005     outeriter        += nri;
1006
1007     /* Update outer/inner flops */
1008
1009     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_F,outeriter*18 + inneriter*120);
1010 }