src/gromacs/gmxlib/nonbonded/nb_kernel_avx_128_fma_double/nb_kernel_ElecCSTab_VdwLJ_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_ElecCSTab_VdwLJ_GeomW4P1_VF_avx_128_fma_double
  52  * Electrostatics interaction: CubicSplineTable
  53  * VdW interaction:            LennardJones
  54  * Geometry:                   Water4-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecCSTab_VdwLJ_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              vdwioffset0;
  81     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  82     int              vdwioffset1;
  83     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  84     int              vdwioffset2;
  85     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  86     int              vdwioffset3;
  87     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  88     int              vdwjidx0A,vdwjidx0B;
  89     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  90     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  91     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  92     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  93     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  94     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
  95     real             *charge;
  96     int              nvdwtype;
  97     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  98     int              *vdwtype;
  99     real             *vdwparam;
 100     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 101     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 102     __m128i          vfitab;
 103     __m128i          ifour       = _mm_set1_epi32(4);
 104     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 105     real             *vftab;
 106     __m128d          dummy_mask,cutoff_mask;
 107     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 108     __m128d          one     = _mm_set1_pd(1.0);
 109     __m128d          two     = _mm_set1_pd(2.0);
 110     x                = xx[0];
 111     f                = ff[0];
 112
 113     nri              = nlist->nri;
 114     iinr             = nlist->iinr;
 115     jindex           = nlist->jindex;
 116     jjnr             = nlist->jjnr;
 117     shiftidx         = nlist->shift;
 118     gid              = nlist->gid;
 119     shiftvec         = fr->shift_vec[0];
 120     fshift           = fr->fshift[0];
 121     facel            = _mm_set1_pd(fr->epsfac);
 122     charge           = mdatoms->chargeA;
 123     nvdwtype         = fr->ntype;
 124     vdwparam         = fr->nbfp;
 125     vdwtype          = mdatoms->typeA;
 126
 127     vftab            = kernel_data->table_elec->data;
 128     vftabscale       = _mm_set1_pd(kernel_data->table_elec->scale);
 129
 130     /* Setup water-specific parameters */
 131     inr              = nlist->iinr[0];
 132     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 133     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 134     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 135     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 136
 137     /* Avoid stupid compiler warnings */
 138     jnrA = jnrB = 0;
 139     j_coord_offsetA = 0;
 140     j_coord_offsetB = 0;
 141
 142     outeriter        = 0;
 143     inneriter        = 0;
 144
 145     /* Start outer loop over neighborlists */
 146     for(iidx=0; iidx<nri; iidx++)
 147     {
 148         /* Load shift vector for this list */
 149         i_shift_offset   = DIM*shiftidx[iidx];
 150
 151         /* Load limits for loop over neighbors */
 152         j_index_start    = jindex[iidx];
 153         j_index_end      = jindex[iidx+1];
 154
 155         /* Get outer coordinate index */
 156         inr              = iinr[iidx];
 157         i_coord_offset   = DIM*inr;
 158
 159         /* Load i particle coords and add shift vector */
 160         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 161                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 162
 163         fix0             = _mm_setzero_pd();
 164         fiy0             = _mm_setzero_pd();
 165         fiz0             = _mm_setzero_pd();
 166         fix1             = _mm_setzero_pd();
 167         fiy1             = _mm_setzero_pd();
 168         fiz1             = _mm_setzero_pd();
 169         fix2             = _mm_setzero_pd();
 170         fiy2             = _mm_setzero_pd();
 171         fiz2             = _mm_setzero_pd();
 172         fix3             = _mm_setzero_pd();
 173         fiy3             = _mm_setzero_pd();
 174         fiz3             = _mm_setzero_pd();
 175
 176         /* Reset potential sums */
 177         velecsum         = _mm_setzero_pd();
 178         vvdwsum          = _mm_setzero_pd();
 179
 180         /* Start inner kernel loop */
 181         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 182         {
 183
 184             /* Get j neighbor index, and coordinate index */
 185             jnrA             = jjnr[jidx];
 186             jnrB             = jjnr[jidx+1];
 187             j_coord_offsetA  = DIM*jnrA;
 188             j_coord_offsetB  = DIM*jnrB;
 189
 190             /* load j atom coordinates */
 191             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 192                                               &jx0,&jy0,&jz0);
 193
 194             /* Calculate displacement vector */
 195             dx00             = _mm_sub_pd(ix0,jx0);
 196             dy00             = _mm_sub_pd(iy0,jy0);
 197             dz00             = _mm_sub_pd(iz0,jz0);
 198             dx10             = _mm_sub_pd(ix1,jx0);
 199             dy10             = _mm_sub_pd(iy1,jy0);
 200             dz10             = _mm_sub_pd(iz1,jz0);
 201             dx20             = _mm_sub_pd(ix2,jx0);
 202             dy20             = _mm_sub_pd(iy2,jy0);
 203             dz20             = _mm_sub_pd(iz2,jz0);
 204             dx30             = _mm_sub_pd(ix3,jx0);
 205             dy30             = _mm_sub_pd(iy3,jy0);
 206             dz30             = _mm_sub_pd(iz3,jz0);
 207
 208             /* Calculate squared distance and things based on it */
 209             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 210             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 211             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 212             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 213
 214             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 215             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 216             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 217
 218             rinvsq00         = gmx_mm_inv_pd(rsq00);
 219
 220             /* Load parameters for j particles */
 221             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 222             vdwjidx0A        = 2*vdwtype[jnrA+0];
 223             vdwjidx0B        = 2*vdwtype[jnrB+0];
 224
 225             fjx0             = _mm_setzero_pd();
 226             fjy0             = _mm_setzero_pd();
 227             fjz0             = _mm_setzero_pd();
 228
 229             /**************************
 230              * CALCULATE INTERACTIONS *
 231              **************************/
 232
 233             /* Compute parameters for interactions between i and j atoms */
 234             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 235                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 236
 237             /* LENNARD-JONES DISPERSION/REPULSION */
 238
 239             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 240             vvdw6            = _mm_mul_pd(c6_00,rinvsix);
 241             vvdw12           = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
 242             vvdw             = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
 243             fvdw             = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
 244
 245             /* Update potential sum for this i atom from the interaction with this j atom. */
 246             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 247
 248             fscal            = fvdw;
 249
 250             /* Update vectorial force */
 251             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 252             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 253             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 254
 255             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 256             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 257             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 258
 259             /**************************
 260              * CALCULATE INTERACTIONS *
 261              **************************/
 262
 263             r10              = _mm_mul_pd(rsq10,rinv10);
 264
 265             /* Compute parameters for interactions between i and j atoms */
 266             qq10             = _mm_mul_pd(iq1,jq0);
 267
 268             /* Calculate table index by multiplying r with table scale and truncate to integer */
 269             rt               = _mm_mul_pd(r10,vftabscale);
 270             vfitab           = _mm_cvttpd_epi32(rt);
 271 #ifdef __XOP__
 272             vfeps            = _mm_frcz_pd(rt);
 273 #else
 274             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 275 #endif
 276             twovfeps         = _mm_add_pd(vfeps,vfeps);
 277             vfitab           = _mm_slli_epi32(vfitab,2);
 278
 279             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 280             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 281             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 282             GMX_MM_TRANSPOSE2_PD(Y,F);
 283             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 284             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 285             GMX_MM_TRANSPOSE2_PD(G,H);
 286             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 287             VV               = _mm_macc_pd(vfeps,Fp,Y);
 288             velec            = _mm_mul_pd(qq10,VV);
 289             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 290             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
 291
 292             /* Update potential sum for this i atom from the interaction with this j atom. */
 293             velecsum         = _mm_add_pd(velecsum,velec);
 294
 295             fscal            = felec;
 296
 297             /* Update vectorial force */
 298             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 299             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 300             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 301
 302             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 303             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 304             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 305
 306             /**************************
 307              * CALCULATE INTERACTIONS *
 308              **************************/
 309
 310             r20              = _mm_mul_pd(rsq20,rinv20);
 311
 312             /* Compute parameters for interactions between i and j atoms */
 313             qq20             = _mm_mul_pd(iq2,jq0);
 314
 315             /* Calculate table index by multiplying r with table scale and truncate to integer */
 316             rt               = _mm_mul_pd(r20,vftabscale);
 317             vfitab           = _mm_cvttpd_epi32(rt);
 318 #ifdef __XOP__
 319             vfeps            = _mm_frcz_pd(rt);
 320 #else
 321             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 322 #endif
 323             twovfeps         = _mm_add_pd(vfeps,vfeps);
 324             vfitab           = _mm_slli_epi32(vfitab,2);
 325
 326             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 327             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 328             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 329             GMX_MM_TRANSPOSE2_PD(Y,F);
 330             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 331             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 332             GMX_MM_TRANSPOSE2_PD(G,H);
 333             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 334             VV               = _mm_macc_pd(vfeps,Fp,Y);
 335             velec            = _mm_mul_pd(qq20,VV);
 336             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 337             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
 338
 339             /* Update potential sum for this i atom from the interaction with this j atom. */
 340             velecsum         = _mm_add_pd(velecsum,velec);
 341
 342             fscal            = felec;
 343
 344             /* Update vectorial force */
 345             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 346             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 347             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 348
 349             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 350             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 351             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 352
 353             /**************************
 354              * CALCULATE INTERACTIONS *
 355              **************************/
 356
 357             r30              = _mm_mul_pd(rsq30,rinv30);
 358
 359             /* Compute parameters for interactions between i and j atoms */
 360             qq30             = _mm_mul_pd(iq3,jq0);
 361
 362             /* Calculate table index by multiplying r with table scale and truncate to integer */
 363             rt               = _mm_mul_pd(r30,vftabscale);
 364             vfitab           = _mm_cvttpd_epi32(rt);
 365 #ifdef __XOP__
 366             vfeps            = _mm_frcz_pd(rt);
 367 #else
 368             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 369 #endif
 370             twovfeps         = _mm_add_pd(vfeps,vfeps);
 371             vfitab           = _mm_slli_epi32(vfitab,2);
 372
 373             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 374             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 375             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 376             GMX_MM_TRANSPOSE2_PD(Y,F);
 377             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 378             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 379             GMX_MM_TRANSPOSE2_PD(G,H);
 380             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 381             VV               = _mm_macc_pd(vfeps,Fp,Y);
 382             velec            = _mm_mul_pd(qq30,VV);
 383             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 384             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
 385
 386             /* Update potential sum for this i atom from the interaction with this j atom. */
 387             velecsum         = _mm_add_pd(velecsum,velec);
 388
 389             fscal            = felec;
 390
 391             /* Update vectorial force */
 392             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 393             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 394             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 395
 396             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 397             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 398             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 399
 400             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 401
 402             /* Inner loop uses 176 flops */
 403         }
 404
 405         if(jidx<j_index_end)
 406         {
 407
 408             jnrA             = jjnr[jidx];
 409             j_coord_offsetA  = DIM*jnrA;
 410
 411             /* load j atom coordinates */
 412             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 413                                               &jx0,&jy0,&jz0);
 414
 415             /* Calculate displacement vector */
 416             dx00             = _mm_sub_pd(ix0,jx0);
 417             dy00             = _mm_sub_pd(iy0,jy0);
 418             dz00             = _mm_sub_pd(iz0,jz0);
 419             dx10             = _mm_sub_pd(ix1,jx0);
 420             dy10             = _mm_sub_pd(iy1,jy0);
 421             dz10             = _mm_sub_pd(iz1,jz0);
 422             dx20             = _mm_sub_pd(ix2,jx0);
 423             dy20             = _mm_sub_pd(iy2,jy0);
 424             dz20             = _mm_sub_pd(iz2,jz0);
 425             dx30             = _mm_sub_pd(ix3,jx0);
 426             dy30             = _mm_sub_pd(iy3,jy0);
 427             dz30             = _mm_sub_pd(iz3,jz0);
 428
 429             /* Calculate squared distance and things based on it */
 430             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 431             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 432             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 433             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 434
 435             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 436             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 437             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 438
 439             rinvsq00         = gmx_mm_inv_pd(rsq00);
 440
 441             /* Load parameters for j particles */
 442             jq0              = _mm_load_sd(charge+jnrA+0);
 443             vdwjidx0A        = 2*vdwtype[jnrA+0];
 444
 445             fjx0             = _mm_setzero_pd();
 446             fjy0             = _mm_setzero_pd();
 447             fjz0             = _mm_setzero_pd();
 448
 449             /**************************
 450              * CALCULATE INTERACTIONS *
 451              **************************/
 452
 453             /* Compute parameters for interactions between i and j atoms */
 454             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
 455
 456             /* LENNARD-JONES DISPERSION/REPULSION */
 457
 458             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 459             vvdw6            = _mm_mul_pd(c6_00,rinvsix);
 460             vvdw12           = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
 461             vvdw             = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
 462             fvdw             = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
 463
 464             /* Update potential sum for this i atom from the interaction with this j atom. */
 465             vvdw             = _mm_unpacklo_pd(vvdw,_mm_setzero_pd());
 466             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 467
 468             fscal            = fvdw;
 469
 470             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 471
 472             /* Update vectorial force */
 473             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 474             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 475             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 476
 477             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 478             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 479             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 480
 481             /**************************
 482              * CALCULATE INTERACTIONS *
 483              **************************/
 484
 485             r10              = _mm_mul_pd(rsq10,rinv10);
 486
 487             /* Compute parameters for interactions between i and j atoms */
 488             qq10             = _mm_mul_pd(iq1,jq0);
 489
 490             /* Calculate table index by multiplying r with table scale and truncate to integer */
 491             rt               = _mm_mul_pd(r10,vftabscale);
 492             vfitab           = _mm_cvttpd_epi32(rt);
 493 #ifdef __XOP__
 494             vfeps            = _mm_frcz_pd(rt);
 495 #else
 496             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 497 #endif
 498             twovfeps         = _mm_add_pd(vfeps,vfeps);
 499             vfitab           = _mm_slli_epi32(vfitab,2);
 500
 501             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 502             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 503             F                = _mm_setzero_pd();
 504             GMX_MM_TRANSPOSE2_PD(Y,F);
 505             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 506             H                = _mm_setzero_pd();
 507             GMX_MM_TRANSPOSE2_PD(G,H);
 508             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 509             VV               = _mm_macc_pd(vfeps,Fp,Y);
 510             velec            = _mm_mul_pd(qq10,VV);
 511             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 512             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
 513
 514             /* Update potential sum for this i atom from the interaction with this j atom. */
 515             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 516             velecsum         = _mm_add_pd(velecsum,velec);
 517
 518             fscal            = felec;
 519
 520             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 521
 522             /* Update vectorial force */
 523             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 524             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 525             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 526
 527             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 528             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 529             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 530
 531             /**************************
 532              * CALCULATE INTERACTIONS *
 533              **************************/
 534
 535             r20              = _mm_mul_pd(rsq20,rinv20);
 536
 537             /* Compute parameters for interactions between i and j atoms */
 538             qq20             = _mm_mul_pd(iq2,jq0);
 539
 540             /* Calculate table index by multiplying r with table scale and truncate to integer */
 541             rt               = _mm_mul_pd(r20,vftabscale);
 542             vfitab           = _mm_cvttpd_epi32(rt);
 543 #ifdef __XOP__
 544             vfeps            = _mm_frcz_pd(rt);
 545 #else
 546             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 547 #endif
 548             twovfeps         = _mm_add_pd(vfeps,vfeps);
 549             vfitab           = _mm_slli_epi32(vfitab,2);
 550
 551             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 552             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 553             F                = _mm_setzero_pd();
 554             GMX_MM_TRANSPOSE2_PD(Y,F);
 555             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 556             H                = _mm_setzero_pd();
 557             GMX_MM_TRANSPOSE2_PD(G,H);
 558             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 559             VV               = _mm_macc_pd(vfeps,Fp,Y);
 560             velec            = _mm_mul_pd(qq20,VV);
 561             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 562             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
 563
 564             /* Update potential sum for this i atom from the interaction with this j atom. */
 565             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 566             velecsum         = _mm_add_pd(velecsum,velec);
 567
 568             fscal            = felec;
 569
 570             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 571
 572             /* Update vectorial force */
 573             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 574             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 575             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 576
 577             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 578             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 579             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 580
 581             /**************************
 582              * CALCULATE INTERACTIONS *
 583              **************************/
 584
 585             r30              = _mm_mul_pd(rsq30,rinv30);
 586
 587             /* Compute parameters for interactions between i and j atoms */
 588             qq30             = _mm_mul_pd(iq3,jq0);
 589
 590             /* Calculate table index by multiplying r with table scale and truncate to integer */
 591             rt               = _mm_mul_pd(r30,vftabscale);
 592             vfitab           = _mm_cvttpd_epi32(rt);
 593 #ifdef __XOP__
 594             vfeps            = _mm_frcz_pd(rt);
 595 #else
 596             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 597 #endif
 598             twovfeps         = _mm_add_pd(vfeps,vfeps);
 599             vfitab           = _mm_slli_epi32(vfitab,2);
 600
 601             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 602             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 603             F                = _mm_setzero_pd();
 604             GMX_MM_TRANSPOSE2_PD(Y,F);
 605             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 606             H                = _mm_setzero_pd();
 607             GMX_MM_TRANSPOSE2_PD(G,H);
 608             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 609             VV               = _mm_macc_pd(vfeps,Fp,Y);
 610             velec            = _mm_mul_pd(qq30,VV);
 611             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 612             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
 613
 614             /* Update potential sum for this i atom from the interaction with this j atom. */
 615             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 616             velecsum         = _mm_add_pd(velecsum,velec);
 617
 618             fscal            = felec;
 619
 620             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 621
 622             /* Update vectorial force */
 623             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 624             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 625             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 626
 627             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 628             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 629             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 630
 631             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 632
 633             /* Inner loop uses 176 flops */
 634         }
 635
 636         /* End of innermost loop */
 637
 638         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 639                                               f+i_coord_offset,fshift+i_shift_offset);
 640
 641         ggid                        = gid[iidx];
 642         /* Update potential energies */
 643         gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 644         gmx_mm_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 645
 646         /* Increment number of inner iterations */
 647         inneriter                  += j_index_end - j_index_start;
 648
 649         /* Outer loop uses 26 flops */
 650     }
 651
 652     /* Increment number of outer iterations */
 653     outeriter        += nri;
 654
 655     /* Update outer/inner flops */
 656
 657     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*176);
 658 }
 659 /*
 660  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_F_avx_128_fma_double
 661  * Electrostatics interaction: CubicSplineTable
 662  * VdW interaction:            LennardJones
 663  * Geometry:                   Water4-Particle
 664  * Calculate force/pot:        Force
 665  */
 666 void
 667 nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_F_avx_128_fma_double
 668                     (t_nblist                    * gmx_restrict       nlist,
 669                      rvec                        * gmx_restrict          xx,
 670                      rvec                        * gmx_restrict          ff,
 671                      t_forcerec                  * gmx_restrict          fr,
 672                      t_mdatoms                   * gmx_restrict     mdatoms,
 673                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 674                      t_nrnb                      * gmx_restrict        nrnb)
 675 {
 676     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 677      * just 0 for non-waters.
 678      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
 679      * jnr indices corresponding to data put in the four positions in the SIMD register.
 680      */
 681     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 682     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 683     int              jnrA,jnrB;
 684     int              j_coord_offsetA,j_coord_offsetB;
 685     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 686     real             rcutoff_scalar;
 687     real             *shiftvec,*fshift,*x,*f;
 688     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 689     int              vdwioffset0;
 690     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 691     int              vdwioffset1;
 692     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 693     int              vdwioffset2;
 694     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 695     int              vdwioffset3;
 696     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 697     int              vdwjidx0A,vdwjidx0B;
 698     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 699     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 700     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 701     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 702     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 703     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 704     real             *charge;
 705     int              nvdwtype;
 706     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 707     int              *vdwtype;
 708     real             *vdwparam;
 709     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 710     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 711     __m128i          vfitab;
 712     __m128i          ifour       = _mm_set1_epi32(4);
 713     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 714     real             *vftab;
 715     __m128d          dummy_mask,cutoff_mask;
 716     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 717     __m128d          one     = _mm_set1_pd(1.0);
 718     __m128d          two     = _mm_set1_pd(2.0);
 719     x                = xx[0];
 720     f                = ff[0];
 721
 722     nri              = nlist->nri;
 723     iinr             = nlist->iinr;
 724     jindex           = nlist->jindex;
 725     jjnr             = nlist->jjnr;
 726     shiftidx         = nlist->shift;
 727     gid              = nlist->gid;
 728     shiftvec         = fr->shift_vec[0];
 729     fshift           = fr->fshift[0];
 730     facel            = _mm_set1_pd(fr->epsfac);
 731     charge           = mdatoms->chargeA;
 732     nvdwtype         = fr->ntype;
 733     vdwparam         = fr->nbfp;
 734     vdwtype          = mdatoms->typeA;
 735
 736     vftab            = kernel_data->table_elec->data;
 737     vftabscale       = _mm_set1_pd(kernel_data->table_elec->scale);
 738
 739     /* Setup water-specific parameters */
 740     inr              = nlist->iinr[0];
 741     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 742     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 743     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 744     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 745
 746     /* Avoid stupid compiler warnings */
 747     jnrA = jnrB = 0;
 748     j_coord_offsetA = 0;
 749     j_coord_offsetB = 0;
 750
 751     outeriter        = 0;
 752     inneriter        = 0;
 753
 754     /* Start outer loop over neighborlists */
 755     for(iidx=0; iidx<nri; iidx++)
 756     {
 757         /* Load shift vector for this list */
 758         i_shift_offset   = DIM*shiftidx[iidx];
 759
 760         /* Load limits for loop over neighbors */
 761         j_index_start    = jindex[iidx];
 762         j_index_end      = jindex[iidx+1];
 763
 764         /* Get outer coordinate index */
 765         inr              = iinr[iidx];
 766         i_coord_offset   = DIM*inr;
 767
 768         /* Load i particle coords and add shift vector */
 769         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 770                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 771
 772         fix0             = _mm_setzero_pd();
 773         fiy0             = _mm_setzero_pd();
 774         fiz0             = _mm_setzero_pd();
 775         fix1             = _mm_setzero_pd();
 776         fiy1             = _mm_setzero_pd();
 777         fiz1             = _mm_setzero_pd();
 778         fix2             = _mm_setzero_pd();
 779         fiy2             = _mm_setzero_pd();
 780         fiz2             = _mm_setzero_pd();
 781         fix3             = _mm_setzero_pd();
 782         fiy3             = _mm_setzero_pd();
 783         fiz3             = _mm_setzero_pd();
 784
 785         /* Start inner kernel loop */
 786         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 787         {
 788
 789             /* Get j neighbor index, and coordinate index */
 790             jnrA             = jjnr[jidx];
 791             jnrB             = jjnr[jidx+1];
 792             j_coord_offsetA  = DIM*jnrA;
 793             j_coord_offsetB  = DIM*jnrB;
 794
 795             /* load j atom coordinates */
 796             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 797                                               &jx0,&jy0,&jz0);
 798
 799             /* Calculate displacement vector */
 800             dx00             = _mm_sub_pd(ix0,jx0);
 801             dy00             = _mm_sub_pd(iy0,jy0);
 802             dz00             = _mm_sub_pd(iz0,jz0);
 803             dx10             = _mm_sub_pd(ix1,jx0);
 804             dy10             = _mm_sub_pd(iy1,jy0);
 805             dz10             = _mm_sub_pd(iz1,jz0);
 806             dx20             = _mm_sub_pd(ix2,jx0);
 807             dy20             = _mm_sub_pd(iy2,jy0);
 808             dz20             = _mm_sub_pd(iz2,jz0);
 809             dx30             = _mm_sub_pd(ix3,jx0);
 810             dy30             = _mm_sub_pd(iy3,jy0);
 811             dz30             = _mm_sub_pd(iz3,jz0);
 812
 813             /* Calculate squared distance and things based on it */
 814             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 815             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 816             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 817             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 818
 819             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 820             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 821             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 822
 823             rinvsq00         = gmx_mm_inv_pd(rsq00);
 824
 825             /* Load parameters for j particles */
 826             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 827             vdwjidx0A        = 2*vdwtype[jnrA+0];
 828             vdwjidx0B        = 2*vdwtype[jnrB+0];
 829
 830             fjx0             = _mm_setzero_pd();
 831             fjy0             = _mm_setzero_pd();
 832             fjz0             = _mm_setzero_pd();
 833
 834             /**************************
 835              * CALCULATE INTERACTIONS *
 836              **************************/
 837
 838             /* Compute parameters for interactions between i and j atoms */
 839             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 840                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 841
 842             /* LENNARD-JONES DISPERSION/REPULSION */
 843
 844             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 845             fvdw             = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
 846
 847             fscal            = fvdw;
 848
 849             /* Update vectorial force */
 850             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 851             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 852             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 853
 854             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 855             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 856             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 857
 858             /**************************
 859              * CALCULATE INTERACTIONS *
 860              **************************/
 861
 862             r10              = _mm_mul_pd(rsq10,rinv10);
 863
 864             /* Compute parameters for interactions between i and j atoms */
 865             qq10             = _mm_mul_pd(iq1,jq0);
 866
 867             /* Calculate table index by multiplying r with table scale and truncate to integer */
 868             rt               = _mm_mul_pd(r10,vftabscale);
 869             vfitab           = _mm_cvttpd_epi32(rt);
 870 #ifdef __XOP__
 871             vfeps            = _mm_frcz_pd(rt);
 872 #else
 873             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 874 #endif
 875             twovfeps         = _mm_add_pd(vfeps,vfeps);
 876             vfitab           = _mm_slli_epi32(vfitab,2);
 877
 878             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 879             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 880             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 881             GMX_MM_TRANSPOSE2_PD(Y,F);
 882             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 883             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 884             GMX_MM_TRANSPOSE2_PD(G,H);
 885             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 886             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 887             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
 888
 889             fscal            = felec;
 890
 891             /* Update vectorial force */
 892             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 893             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 894             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 895
 896             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 897             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 898             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 899
 900             /**************************
 901              * CALCULATE INTERACTIONS *
 902              **************************/
 903
 904             r20              = _mm_mul_pd(rsq20,rinv20);
 905
 906             /* Compute parameters for interactions between i and j atoms */
 907             qq20             = _mm_mul_pd(iq2,jq0);
 908
 909             /* Calculate table index by multiplying r with table scale and truncate to integer */
 910             rt               = _mm_mul_pd(r20,vftabscale);
 911             vfitab           = _mm_cvttpd_epi32(rt);
 912 #ifdef __XOP__
 913             vfeps            = _mm_frcz_pd(rt);
 914 #else
 915             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 916 #endif
 917             twovfeps         = _mm_add_pd(vfeps,vfeps);
 918             vfitab           = _mm_slli_epi32(vfitab,2);
 919
 920             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 921             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 922             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 923             GMX_MM_TRANSPOSE2_PD(Y,F);
 924             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 925             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 926             GMX_MM_TRANSPOSE2_PD(G,H);
 927             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 928             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 929             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
 930
 931             fscal            = felec;
 932
 933             /* Update vectorial force */
 934             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 935             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 936             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 937
 938             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 939             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 940             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 941
 942             /**************************
 943              * CALCULATE INTERACTIONS *
 944              **************************/
 945
 946             r30              = _mm_mul_pd(rsq30,rinv30);
 947
 948             /* Compute parameters for interactions between i and j atoms */
 949             qq30             = _mm_mul_pd(iq3,jq0);
 950
 951             /* Calculate table index by multiplying r with table scale and truncate to integer */
 952             rt               = _mm_mul_pd(r30,vftabscale);
 953             vfitab           = _mm_cvttpd_epi32(rt);
 954 #ifdef __XOP__
 955             vfeps            = _mm_frcz_pd(rt);
 956 #else
 957             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 958 #endif
 959             twovfeps         = _mm_add_pd(vfeps,vfeps);
 960             vfitab           = _mm_slli_epi32(vfitab,2);
 961
 962             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 963             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 964             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 965             GMX_MM_TRANSPOSE2_PD(Y,F);
 966             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 967             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 968             GMX_MM_TRANSPOSE2_PD(G,H);
 969             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 970             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 971             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
 972
 973             fscal            = felec;
 974
 975             /* Update vectorial force */
 976             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 977             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 978             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 979
 980             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 981             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 982             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 983
 984             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 985
 986             /* Inner loop uses 159 flops */
 987         }
 988
 989         if(jidx<j_index_end)
 990         {
 991
 992             jnrA             = jjnr[jidx];
 993             j_coord_offsetA  = DIM*jnrA;
 994
 995             /* load j atom coordinates */
 996             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 997                                               &jx0,&jy0,&jz0);
 998
 999             /* Calculate displacement vector */
1000             dx00             = _mm_sub_pd(ix0,jx0);
1001             dy00             = _mm_sub_pd(iy0,jy0);
1002             dz00             = _mm_sub_pd(iz0,jz0);
1003             dx10             = _mm_sub_pd(ix1,jx0);
1004             dy10             = _mm_sub_pd(iy1,jy0);
1005             dz10             = _mm_sub_pd(iz1,jz0);
1006             dx20             = _mm_sub_pd(ix2,jx0);
1007             dy20             = _mm_sub_pd(iy2,jy0);
1008             dz20             = _mm_sub_pd(iz2,jz0);
1009             dx30             = _mm_sub_pd(ix3,jx0);
1010             dy30             = _mm_sub_pd(iy3,jy0);
1011             dz30             = _mm_sub_pd(iz3,jz0);
1012
1013             /* Calculate squared distance and things based on it */
1014             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
1015             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
1016             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
1017             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
1018
1019             rinv10           = gmx_mm_invsqrt_pd(rsq10);
1020             rinv20           = gmx_mm_invsqrt_pd(rsq20);
1021             rinv30           = gmx_mm_invsqrt_pd(rsq30);
1022
1023             rinvsq00         = gmx_mm_inv_pd(rsq00);
1024
1025             /* Load parameters for j particles */
1026             jq0              = _mm_load_sd(charge+jnrA+0);
1027             vdwjidx0A        = 2*vdwtype[jnrA+0];
1028
1029             fjx0             = _mm_setzero_pd();
1030             fjy0             = _mm_setzero_pd();
1031             fjz0             = _mm_setzero_pd();
1032
1033             /**************************
1034              * CALCULATE INTERACTIONS *
1035              **************************/
1036
1037             /* Compute parameters for interactions between i and j atoms */
1038             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
1039
1040             /* LENNARD-JONES DISPERSION/REPULSION */
1041
1042             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1043             fvdw             = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
1044
1045             fscal            = fvdw;
1046
1047             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1048
1049             /* Update vectorial force */
1050             fix0             = _mm_macc_pd(dx00,fscal,fix0);
1051             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
1052             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
1053
1054             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
1055             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
1056             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
1057
1058             /**************************
1059              * CALCULATE INTERACTIONS *
1060              **************************/
1061
1062             r10              = _mm_mul_pd(rsq10,rinv10);
1063
1064             /* Compute parameters for interactions between i and j atoms */
1065             qq10             = _mm_mul_pd(iq1,jq0);
1066
1067             /* Calculate table index by multiplying r with table scale and truncate to integer */
1068             rt               = _mm_mul_pd(r10,vftabscale);
1069             vfitab           = _mm_cvttpd_epi32(rt);
1070 #ifdef __XOP__
1071             vfeps            = _mm_frcz_pd(rt);
1072 #else
1073             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1074 #endif
1075             twovfeps         = _mm_add_pd(vfeps,vfeps);
1076             vfitab           = _mm_slli_epi32(vfitab,2);
1077
1078             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1079             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1080             F                = _mm_setzero_pd();
1081             GMX_MM_TRANSPOSE2_PD(Y,F);
1082             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1083             H                = _mm_setzero_pd();
1084             GMX_MM_TRANSPOSE2_PD(G,H);
1085             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1086             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1087             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
1088
1089             fscal            = felec;
1090
1091             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1092
1093             /* Update vectorial force */
1094             fix1             = _mm_macc_pd(dx10,fscal,fix1);
1095             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
1096             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
1097
1098             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
1099             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
1100             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
1101
1102             /**************************
1103              * CALCULATE INTERACTIONS *
1104              **************************/
1105
1106             r20              = _mm_mul_pd(rsq20,rinv20);
1107
1108             /* Compute parameters for interactions between i and j atoms */
1109             qq20             = _mm_mul_pd(iq2,jq0);
1110
1111             /* Calculate table index by multiplying r with table scale and truncate to integer */
1112             rt               = _mm_mul_pd(r20,vftabscale);
1113             vfitab           = _mm_cvttpd_epi32(rt);
1114 #ifdef __XOP__
1115             vfeps            = _mm_frcz_pd(rt);
1116 #else
1117             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1118 #endif
1119             twovfeps         = _mm_add_pd(vfeps,vfeps);
1120             vfitab           = _mm_slli_epi32(vfitab,2);
1121
1122             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1123             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1124             F                = _mm_setzero_pd();
1125             GMX_MM_TRANSPOSE2_PD(Y,F);
1126             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1127             H                = _mm_setzero_pd();
1128             GMX_MM_TRANSPOSE2_PD(G,H);
1129             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1130             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1131             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
1132
1133             fscal            = felec;
1134
1135             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1136
1137             /* Update vectorial force */
1138             fix2             = _mm_macc_pd(dx20,fscal,fix2);
1139             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
1140             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
1141
1142             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
1143             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
1144             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
1145
1146             /**************************
1147              * CALCULATE INTERACTIONS *
1148              **************************/
1149
1150             r30              = _mm_mul_pd(rsq30,rinv30);
1151
1152             /* Compute parameters for interactions between i and j atoms */
1153             qq30             = _mm_mul_pd(iq3,jq0);
1154
1155             /* Calculate table index by multiplying r with table scale and truncate to integer */
1156             rt               = _mm_mul_pd(r30,vftabscale);
1157             vfitab           = _mm_cvttpd_epi32(rt);
1158 #ifdef __XOP__
1159             vfeps            = _mm_frcz_pd(rt);
1160 #else
1161             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1162 #endif
1163             twovfeps         = _mm_add_pd(vfeps,vfeps);
1164             vfitab           = _mm_slli_epi32(vfitab,2);
1165
1166             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1167             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1168             F                = _mm_setzero_pd();
1169             GMX_MM_TRANSPOSE2_PD(Y,F);
1170             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1171             H                = _mm_setzero_pd();
1172             GMX_MM_TRANSPOSE2_PD(G,H);
1173             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1174             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1175             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
1176
1177             fscal            = felec;
1178
1179             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1180
1181             /* Update vectorial force */
1182             fix3             = _mm_macc_pd(dx30,fscal,fix3);
1183             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
1184             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
1185
1186             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
1187             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
1188             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
1189
1190             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
1191
1192             /* Inner loop uses 159 flops */
1193         }
1194
1195         /* End of innermost loop */
1196
1197         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1198                                               f+i_coord_offset,fshift+i_shift_offset);
1199
1200         /* Increment number of inner iterations */
1201         inneriter                  += j_index_end - j_index_start;
1202
1203         /* Outer loop uses 24 flops */
1204     }
1205
1206     /* Increment number of outer iterations */
1207     outeriter        += nri;
1208
1209     /* Update outer/inner flops */
1210
1211     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*159);
1212 }