src/gmxlib/nonbonded/nb_kernel_avx_128_fma_double/nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_avx_128_fma_double.c

   1 /*
   2  * Note: this file was generated by the Gromacs avx_128_fma_double kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 #include "gmx_math_x86_avx_128_fma_double.h"
  34 #include "kernelutil_x86_avx_128_fma_double.h"
  35
  36 /*
  37  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_VF_avx_128_fma_double
  38  * Electrostatics interaction: CubicSplineTable
  39  * VdW interaction:            LennardJones
  40  * Geometry:                   Water4-Particle
  41  * Calculate force/pot:        PotentialAndForce
  42  */
  43 void
  44 nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_VF_avx_128_fma_double
  45                     (t_nblist * gmx_restrict                nlist,
  46                      rvec * gmx_restrict                    xx,
  47                      rvec * gmx_restrict                    ff,
  48                      t_forcerec * gmx_restrict              fr,
  49                      t_mdatoms * gmx_restrict               mdatoms,
  50                      nb_kernel_data_t * gmx_restrict        kernel_data,
  51                      t_nrnb * gmx_restrict                  nrnb)
  52 {
  53     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  54      * just 0 for non-waters.
  55      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
  56      * jnr indices corresponding to data put in the four positions in the SIMD register.
  57      */
  58     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  59     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  60     int              jnrA,jnrB;
  61     int              j_coord_offsetA,j_coord_offsetB;
  62     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  63     real             rcutoff_scalar;
  64     real             *shiftvec,*fshift,*x,*f;
  65     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  66     int              vdwioffset0;
  67     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  68     int              vdwioffset1;
  69     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  70     int              vdwioffset2;
  71     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  72     int              vdwioffset3;
  73     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  74     int              vdwjidx0A,vdwjidx0B;
  75     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  76     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  77     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  78     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  79     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  80     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
  81     real             *charge;
  82     int              nvdwtype;
  83     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  84     int              *vdwtype;
  85     real             *vdwparam;
  86     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
  87     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
  88     __m128i          vfitab;
  89     __m128i          ifour       = _mm_set1_epi32(4);
  90     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
  91     real             *vftab;
  92     __m128d          dummy_mask,cutoff_mask;
  93     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
  94     __m128d          one     = _mm_set1_pd(1.0);
  95     __m128d          two     = _mm_set1_pd(2.0);
  96     x                = xx[0];
  97     f                = ff[0];
  98
  99     nri              = nlist->nri;
 100     iinr             = nlist->iinr;
 101     jindex           = nlist->jindex;
 102     jjnr             = nlist->jjnr;
 103     shiftidx         = nlist->shift;
 104     gid              = nlist->gid;
 105     shiftvec         = fr->shift_vec[0];
 106     fshift           = fr->fshift[0];
 107     facel            = _mm_set1_pd(fr->epsfac);
 108     charge           = mdatoms->chargeA;
 109     nvdwtype         = fr->ntype;
 110     vdwparam         = fr->nbfp;
 111     vdwtype          = mdatoms->typeA;
 112
 113     vftab            = kernel_data->table_elec->data;
 114     vftabscale       = _mm_set1_pd(kernel_data->table_elec->scale);
 115
 116     /* Setup water-specific parameters */
 117     inr              = nlist->iinr[0];
 118     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 119     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 120     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 121     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 122
 123     /* Avoid stupid compiler warnings */
 124     jnrA = jnrB = 0;
 125     j_coord_offsetA = 0;
 126     j_coord_offsetB = 0;
 127
 128     outeriter        = 0;
 129     inneriter        = 0;
 130
 131     /* Start outer loop over neighborlists */
 132     for(iidx=0; iidx<nri; iidx++)
 133     {
 134         /* Load shift vector for this list */
 135         i_shift_offset   = DIM*shiftidx[iidx];
 136
 137         /* Load limits for loop over neighbors */
 138         j_index_start    = jindex[iidx];
 139         j_index_end      = jindex[iidx+1];
 140
 141         /* Get outer coordinate index */
 142         inr              = iinr[iidx];
 143         i_coord_offset   = DIM*inr;
 144
 145         /* Load i particle coords and add shift vector */
 146         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 147                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 148
 149         fix0             = _mm_setzero_pd();
 150         fiy0             = _mm_setzero_pd();
 151         fiz0             = _mm_setzero_pd();
 152         fix1             = _mm_setzero_pd();
 153         fiy1             = _mm_setzero_pd();
 154         fiz1             = _mm_setzero_pd();
 155         fix2             = _mm_setzero_pd();
 156         fiy2             = _mm_setzero_pd();
 157         fiz2             = _mm_setzero_pd();
 158         fix3             = _mm_setzero_pd();
 159         fiy3             = _mm_setzero_pd();
 160         fiz3             = _mm_setzero_pd();
 161
 162         /* Reset potential sums */
 163         velecsum         = _mm_setzero_pd();
 164         vvdwsum          = _mm_setzero_pd();
 165
 166         /* Start inner kernel loop */
 167         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 168         {
 169
 170             /* Get j neighbor index, and coordinate index */
 171             jnrA             = jjnr[jidx];
 172             jnrB             = jjnr[jidx+1];
 173             j_coord_offsetA  = DIM*jnrA;
 174             j_coord_offsetB  = DIM*jnrB;
 175
 176             /* load j atom coordinates */
 177             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 178                                               &jx0,&jy0,&jz0);
 179
 180             /* Calculate displacement vector */
 181             dx00             = _mm_sub_pd(ix0,jx0);
 182             dy00             = _mm_sub_pd(iy0,jy0);
 183             dz00             = _mm_sub_pd(iz0,jz0);
 184             dx10             = _mm_sub_pd(ix1,jx0);
 185             dy10             = _mm_sub_pd(iy1,jy0);
 186             dz10             = _mm_sub_pd(iz1,jz0);
 187             dx20             = _mm_sub_pd(ix2,jx0);
 188             dy20             = _mm_sub_pd(iy2,jy0);
 189             dz20             = _mm_sub_pd(iz2,jz0);
 190             dx30             = _mm_sub_pd(ix3,jx0);
 191             dy30             = _mm_sub_pd(iy3,jy0);
 192             dz30             = _mm_sub_pd(iz3,jz0);
 193
 194             /* Calculate squared distance and things based on it */
 195             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 196             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 197             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 198             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 199
 200             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 201             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 202             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 203
 204             rinvsq00         = gmx_mm_inv_pd(rsq00);
 205
 206             /* Load parameters for j particles */
 207             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 208             vdwjidx0A        = 2*vdwtype[jnrA+0];
 209             vdwjidx0B        = 2*vdwtype[jnrB+0];
 210
 211             fjx0             = _mm_setzero_pd();
 212             fjy0             = _mm_setzero_pd();
 213             fjz0             = _mm_setzero_pd();
 214
 215             /**************************
 216              * CALCULATE INTERACTIONS *
 217              **************************/
 218
 219             /* Compute parameters for interactions between i and j atoms */
 220             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 221                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 222
 223             /* LENNARD-JONES DISPERSION/REPULSION */
 224
 225             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 226             vvdw6            = _mm_mul_pd(c6_00,rinvsix);
 227             vvdw12           = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
 228             vvdw             = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
 229             fvdw             = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
 230
 231             /* Update potential sum for this i atom from the interaction with this j atom. */
 232             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 233
 234             fscal            = fvdw;
 235
 236             /* Update vectorial force */
 237             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 238             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 239             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 240
 241             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 242             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 243             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 244
 245             /**************************
 246              * CALCULATE INTERACTIONS *
 247              **************************/
 248
 249             r10              = _mm_mul_pd(rsq10,rinv10);
 250
 251             /* Compute parameters for interactions between i and j atoms */
 252             qq10             = _mm_mul_pd(iq1,jq0);
 253
 254             /* Calculate table index by multiplying r with table scale and truncate to integer */
 255             rt               = _mm_mul_pd(r10,vftabscale);
 256             vfitab           = _mm_cvttpd_epi32(rt);
 257 #ifdef __XOP__
 258             vfeps            = _mm_frcz_pd(rt);
 259 #else
 260             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 261 #endif
 262             twovfeps         = _mm_add_pd(vfeps,vfeps);
 263             vfitab           = _mm_slli_epi32(vfitab,2);
 264
 265             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 266             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 267             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 268             GMX_MM_TRANSPOSE2_PD(Y,F);
 269             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 270             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 271             GMX_MM_TRANSPOSE2_PD(G,H);
 272             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 273             VV               = _mm_macc_pd(vfeps,Fp,Y);
 274             velec            = _mm_mul_pd(qq10,VV);
 275             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 276             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
 277
 278             /* Update potential sum for this i atom from the interaction with this j atom. */
 279             velecsum         = _mm_add_pd(velecsum,velec);
 280
 281             fscal            = felec;
 282
 283             /* Update vectorial force */
 284             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 285             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 286             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 287
 288             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 289             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 290             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 291
 292             /**************************
 293              * CALCULATE INTERACTIONS *
 294              **************************/
 295
 296             r20              = _mm_mul_pd(rsq20,rinv20);
 297
 298             /* Compute parameters for interactions between i and j atoms */
 299             qq20             = _mm_mul_pd(iq2,jq0);
 300
 301             /* Calculate table index by multiplying r with table scale and truncate to integer */
 302             rt               = _mm_mul_pd(r20,vftabscale);
 303             vfitab           = _mm_cvttpd_epi32(rt);
 304 #ifdef __XOP__
 305             vfeps            = _mm_frcz_pd(rt);
 306 #else
 307             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 308 #endif
 309             twovfeps         = _mm_add_pd(vfeps,vfeps);
 310             vfitab           = _mm_slli_epi32(vfitab,2);
 311
 312             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 313             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 314             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 315             GMX_MM_TRANSPOSE2_PD(Y,F);
 316             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 317             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 318             GMX_MM_TRANSPOSE2_PD(G,H);
 319             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 320             VV               = _mm_macc_pd(vfeps,Fp,Y);
 321             velec            = _mm_mul_pd(qq20,VV);
 322             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 323             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
 324
 325             /* Update potential sum for this i atom from the interaction with this j atom. */
 326             velecsum         = _mm_add_pd(velecsum,velec);
 327
 328             fscal            = felec;
 329
 330             /* Update vectorial force */
 331             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 332             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 333             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 334
 335             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 336             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 337             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 338
 339             /**************************
 340              * CALCULATE INTERACTIONS *
 341              **************************/
 342
 343             r30              = _mm_mul_pd(rsq30,rinv30);
 344
 345             /* Compute parameters for interactions between i and j atoms */
 346             qq30             = _mm_mul_pd(iq3,jq0);
 347
 348             /* Calculate table index by multiplying r with table scale and truncate to integer */
 349             rt               = _mm_mul_pd(r30,vftabscale);
 350             vfitab           = _mm_cvttpd_epi32(rt);
 351 #ifdef __XOP__
 352             vfeps            = _mm_frcz_pd(rt);
 353 #else
 354             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 355 #endif
 356             twovfeps         = _mm_add_pd(vfeps,vfeps);
 357             vfitab           = _mm_slli_epi32(vfitab,2);
 358
 359             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 360             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 361             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 362             GMX_MM_TRANSPOSE2_PD(Y,F);
 363             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 364             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 365             GMX_MM_TRANSPOSE2_PD(G,H);
 366             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 367             VV               = _mm_macc_pd(vfeps,Fp,Y);
 368             velec            = _mm_mul_pd(qq30,VV);
 369             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 370             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
 371
 372             /* Update potential sum for this i atom from the interaction with this j atom. */
 373             velecsum         = _mm_add_pd(velecsum,velec);
 374
 375             fscal            = felec;
 376
 377             /* Update vectorial force */
 378             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 379             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 380             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 381
 382             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 383             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 384             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 385
 386             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 387
 388             /* Inner loop uses 176 flops */
 389         }
 390
 391         if(jidx<j_index_end)
 392         {
 393
 394             jnrA             = jjnr[jidx];
 395             j_coord_offsetA  = DIM*jnrA;
 396
 397             /* load j atom coordinates */
 398             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 399                                               &jx0,&jy0,&jz0);
 400
 401             /* Calculate displacement vector */
 402             dx00             = _mm_sub_pd(ix0,jx0);
 403             dy00             = _mm_sub_pd(iy0,jy0);
 404             dz00             = _mm_sub_pd(iz0,jz0);
 405             dx10             = _mm_sub_pd(ix1,jx0);
 406             dy10             = _mm_sub_pd(iy1,jy0);
 407             dz10             = _mm_sub_pd(iz1,jz0);
 408             dx20             = _mm_sub_pd(ix2,jx0);
 409             dy20             = _mm_sub_pd(iy2,jy0);
 410             dz20             = _mm_sub_pd(iz2,jz0);
 411             dx30             = _mm_sub_pd(ix3,jx0);
 412             dy30             = _mm_sub_pd(iy3,jy0);
 413             dz30             = _mm_sub_pd(iz3,jz0);
 414
 415             /* Calculate squared distance and things based on it */
 416             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 417             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 418             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 419             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 420
 421             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 422             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 423             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 424
 425             rinvsq00         = gmx_mm_inv_pd(rsq00);
 426
 427             /* Load parameters for j particles */
 428             jq0              = _mm_load_sd(charge+jnrA+0);
 429             vdwjidx0A        = 2*vdwtype[jnrA+0];
 430
 431             fjx0             = _mm_setzero_pd();
 432             fjy0             = _mm_setzero_pd();
 433             fjz0             = _mm_setzero_pd();
 434
 435             /**************************
 436              * CALCULATE INTERACTIONS *
 437              **************************/
 438
 439             /* Compute parameters for interactions between i and j atoms */
 440             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
 441
 442             /* LENNARD-JONES DISPERSION/REPULSION */
 443
 444             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 445             vvdw6            = _mm_mul_pd(c6_00,rinvsix);
 446             vvdw12           = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
 447             vvdw             = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
 448             fvdw             = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
 449
 450             /* Update potential sum for this i atom from the interaction with this j atom. */
 451             vvdw             = _mm_unpacklo_pd(vvdw,_mm_setzero_pd());
 452             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 453
 454             fscal            = fvdw;
 455
 456             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 457
 458             /* Update vectorial force */
 459             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 460             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 461             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 462
 463             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 464             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 465             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 466
 467             /**************************
 468              * CALCULATE INTERACTIONS *
 469              **************************/
 470
 471             r10              = _mm_mul_pd(rsq10,rinv10);
 472
 473             /* Compute parameters for interactions between i and j atoms */
 474             qq10             = _mm_mul_pd(iq1,jq0);
 475
 476             /* Calculate table index by multiplying r with table scale and truncate to integer */
 477             rt               = _mm_mul_pd(r10,vftabscale);
 478             vfitab           = _mm_cvttpd_epi32(rt);
 479 #ifdef __XOP__
 480             vfeps            = _mm_frcz_pd(rt);
 481 #else
 482             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 483 #endif
 484             twovfeps         = _mm_add_pd(vfeps,vfeps);
 485             vfitab           = _mm_slli_epi32(vfitab,2);
 486
 487             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 488             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 489             F                = _mm_setzero_pd();
 490             GMX_MM_TRANSPOSE2_PD(Y,F);
 491             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 492             H                = _mm_setzero_pd();
 493             GMX_MM_TRANSPOSE2_PD(G,H);
 494             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 495             VV               = _mm_macc_pd(vfeps,Fp,Y);
 496             velec            = _mm_mul_pd(qq10,VV);
 497             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 498             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
 499
 500             /* Update potential sum for this i atom from the interaction with this j atom. */
 501             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 502             velecsum         = _mm_add_pd(velecsum,velec);
 503
 504             fscal            = felec;
 505
 506             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 507
 508             /* Update vectorial force */
 509             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 510             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 511             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 512
 513             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 514             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 515             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 516
 517             /**************************
 518              * CALCULATE INTERACTIONS *
 519              **************************/
 520
 521             r20              = _mm_mul_pd(rsq20,rinv20);
 522
 523             /* Compute parameters for interactions between i and j atoms */
 524             qq20             = _mm_mul_pd(iq2,jq0);
 525
 526             /* Calculate table index by multiplying r with table scale and truncate to integer */
 527             rt               = _mm_mul_pd(r20,vftabscale);
 528             vfitab           = _mm_cvttpd_epi32(rt);
 529 #ifdef __XOP__
 530             vfeps            = _mm_frcz_pd(rt);
 531 #else
 532             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 533 #endif
 534             twovfeps         = _mm_add_pd(vfeps,vfeps);
 535             vfitab           = _mm_slli_epi32(vfitab,2);
 536
 537             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 538             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 539             F                = _mm_setzero_pd();
 540             GMX_MM_TRANSPOSE2_PD(Y,F);
 541             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 542             H                = _mm_setzero_pd();
 543             GMX_MM_TRANSPOSE2_PD(G,H);
 544             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 545             VV               = _mm_macc_pd(vfeps,Fp,Y);
 546             velec            = _mm_mul_pd(qq20,VV);
 547             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 548             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
 549
 550             /* Update potential sum for this i atom from the interaction with this j atom. */
 551             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 552             velecsum         = _mm_add_pd(velecsum,velec);
 553
 554             fscal            = felec;
 555
 556             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 557
 558             /* Update vectorial force */
 559             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 560             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 561             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 562
 563             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 564             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 565             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 566
 567             /**************************
 568              * CALCULATE INTERACTIONS *
 569              **************************/
 570
 571             r30              = _mm_mul_pd(rsq30,rinv30);
 572
 573             /* Compute parameters for interactions between i and j atoms */
 574             qq30             = _mm_mul_pd(iq3,jq0);
 575
 576             /* Calculate table index by multiplying r with table scale and truncate to integer */
 577             rt               = _mm_mul_pd(r30,vftabscale);
 578             vfitab           = _mm_cvttpd_epi32(rt);
 579 #ifdef __XOP__
 580             vfeps            = _mm_frcz_pd(rt);
 581 #else
 582             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 583 #endif
 584             twovfeps         = _mm_add_pd(vfeps,vfeps);
 585             vfitab           = _mm_slli_epi32(vfitab,2);
 586
 587             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 588             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 589             F                = _mm_setzero_pd();
 590             GMX_MM_TRANSPOSE2_PD(Y,F);
 591             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 592             H                = _mm_setzero_pd();
 593             GMX_MM_TRANSPOSE2_PD(G,H);
 594             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 595             VV               = _mm_macc_pd(vfeps,Fp,Y);
 596             velec            = _mm_mul_pd(qq30,VV);
 597             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 598             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
 599
 600             /* Update potential sum for this i atom from the interaction with this j atom. */
 601             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 602             velecsum         = _mm_add_pd(velecsum,velec);
 603
 604             fscal            = felec;
 605
 606             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 607
 608             /* Update vectorial force */
 609             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 610             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 611             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 612
 613             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 614             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 615             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 616
 617             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 618
 619             /* Inner loop uses 176 flops */
 620         }
 621
 622         /* End of innermost loop */
 623
 624         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 625                                               f+i_coord_offset,fshift+i_shift_offset);
 626
 627         ggid                        = gid[iidx];
 628         /* Update potential energies */
 629         gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 630         gmx_mm_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 631
 632         /* Increment number of inner iterations */
 633         inneriter                  += j_index_end - j_index_start;
 634
 635         /* Outer loop uses 26 flops */
 636     }
 637
 638     /* Increment number of outer iterations */
 639     outeriter        += nri;
 640
 641     /* Update outer/inner flops */
 642
 643     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*176);
 644 }
 645 /*
 646  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_F_avx_128_fma_double
 647  * Electrostatics interaction: CubicSplineTable
 648  * VdW interaction:            LennardJones
 649  * Geometry:                   Water4-Particle
 650  * Calculate force/pot:        Force
 651  */
 652 void
 653 nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_F_avx_128_fma_double
 654                     (t_nblist * gmx_restrict                nlist,
 655                      rvec * gmx_restrict                    xx,
 656                      rvec * gmx_restrict                    ff,
 657                      t_forcerec * gmx_restrict              fr,
 658                      t_mdatoms * gmx_restrict               mdatoms,
 659                      nb_kernel_data_t * gmx_restrict        kernel_data,
 660                      t_nrnb * gmx_restrict                  nrnb)
 661 {
 662     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 663      * just 0 for non-waters.
 664      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
 665      * jnr indices corresponding to data put in the four positions in the SIMD register.
 666      */
 667     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 668     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 669     int              jnrA,jnrB;
 670     int              j_coord_offsetA,j_coord_offsetB;
 671     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 672     real             rcutoff_scalar;
 673     real             *shiftvec,*fshift,*x,*f;
 674     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 675     int              vdwioffset0;
 676     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 677     int              vdwioffset1;
 678     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 679     int              vdwioffset2;
 680     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 681     int              vdwioffset3;
 682     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 683     int              vdwjidx0A,vdwjidx0B;
 684     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 685     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 686     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 687     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 688     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 689     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 690     real             *charge;
 691     int              nvdwtype;
 692     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 693     int              *vdwtype;
 694     real             *vdwparam;
 695     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 696     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 697     __m128i          vfitab;
 698     __m128i          ifour       = _mm_set1_epi32(4);
 699     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 700     real             *vftab;
 701     __m128d          dummy_mask,cutoff_mask;
 702     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 703     __m128d          one     = _mm_set1_pd(1.0);
 704     __m128d          two     = _mm_set1_pd(2.0);
 705     x                = xx[0];
 706     f                = ff[0];
 707
 708     nri              = nlist->nri;
 709     iinr             = nlist->iinr;
 710     jindex           = nlist->jindex;
 711     jjnr             = nlist->jjnr;
 712     shiftidx         = nlist->shift;
 713     gid              = nlist->gid;
 714     shiftvec         = fr->shift_vec[0];
 715     fshift           = fr->fshift[0];
 716     facel            = _mm_set1_pd(fr->epsfac);
 717     charge           = mdatoms->chargeA;
 718     nvdwtype         = fr->ntype;
 719     vdwparam         = fr->nbfp;
 720     vdwtype          = mdatoms->typeA;
 721
 722     vftab            = kernel_data->table_elec->data;
 723     vftabscale       = _mm_set1_pd(kernel_data->table_elec->scale);
 724
 725     /* Setup water-specific parameters */
 726     inr              = nlist->iinr[0];
 727     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 728     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 729     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 730     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 731
 732     /* Avoid stupid compiler warnings */
 733     jnrA = jnrB = 0;
 734     j_coord_offsetA = 0;
 735     j_coord_offsetB = 0;
 736
 737     outeriter        = 0;
 738     inneriter        = 0;
 739
 740     /* Start outer loop over neighborlists */
 741     for(iidx=0; iidx<nri; iidx++)
 742     {
 743         /* Load shift vector for this list */
 744         i_shift_offset   = DIM*shiftidx[iidx];
 745
 746         /* Load limits for loop over neighbors */
 747         j_index_start    = jindex[iidx];
 748         j_index_end      = jindex[iidx+1];
 749
 750         /* Get outer coordinate index */
 751         inr              = iinr[iidx];
 752         i_coord_offset   = DIM*inr;
 753
 754         /* Load i particle coords and add shift vector */
 755         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 756                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 757
 758         fix0             = _mm_setzero_pd();
 759         fiy0             = _mm_setzero_pd();
 760         fiz0             = _mm_setzero_pd();
 761         fix1             = _mm_setzero_pd();
 762         fiy1             = _mm_setzero_pd();
 763         fiz1             = _mm_setzero_pd();
 764         fix2             = _mm_setzero_pd();
 765         fiy2             = _mm_setzero_pd();
 766         fiz2             = _mm_setzero_pd();
 767         fix3             = _mm_setzero_pd();
 768         fiy3             = _mm_setzero_pd();
 769         fiz3             = _mm_setzero_pd();
 770
 771         /* Start inner kernel loop */
 772         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 773         {
 774
 775             /* Get j neighbor index, and coordinate index */
 776             jnrA             = jjnr[jidx];
 777             jnrB             = jjnr[jidx+1];
 778             j_coord_offsetA  = DIM*jnrA;
 779             j_coord_offsetB  = DIM*jnrB;
 780
 781             /* load j atom coordinates */
 782             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 783                                               &jx0,&jy0,&jz0);
 784
 785             /* Calculate displacement vector */
 786             dx00             = _mm_sub_pd(ix0,jx0);
 787             dy00             = _mm_sub_pd(iy0,jy0);
 788             dz00             = _mm_sub_pd(iz0,jz0);
 789             dx10             = _mm_sub_pd(ix1,jx0);
 790             dy10             = _mm_sub_pd(iy1,jy0);
 791             dz10             = _mm_sub_pd(iz1,jz0);
 792             dx20             = _mm_sub_pd(ix2,jx0);
 793             dy20             = _mm_sub_pd(iy2,jy0);
 794             dz20             = _mm_sub_pd(iz2,jz0);
 795             dx30             = _mm_sub_pd(ix3,jx0);
 796             dy30             = _mm_sub_pd(iy3,jy0);
 797             dz30             = _mm_sub_pd(iz3,jz0);
 798
 799             /* Calculate squared distance and things based on it */
 800             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 801             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 802             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 803             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 804
 805             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 806             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 807             rinv30           = gmx_mm_invsqrt_pd(rsq30);
 808
 809             rinvsq00         = gmx_mm_inv_pd(rsq00);
 810
 811             /* Load parameters for j particles */
 812             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 813             vdwjidx0A        = 2*vdwtype[jnrA+0];
 814             vdwjidx0B        = 2*vdwtype[jnrB+0];
 815
 816             fjx0             = _mm_setzero_pd();
 817             fjy0             = _mm_setzero_pd();
 818             fjz0             = _mm_setzero_pd();
 819
 820             /**************************
 821              * CALCULATE INTERACTIONS *
 822              **************************/
 823
 824             /* Compute parameters for interactions between i and j atoms */
 825             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 826                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 827
 828             /* LENNARD-JONES DISPERSION/REPULSION */
 829
 830             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 831             fvdw             = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
 832
 833             fscal            = fvdw;
 834
 835             /* Update vectorial force */
 836             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 837             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 838             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 839
 840             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 841             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 842             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 843
 844             /**************************
 845              * CALCULATE INTERACTIONS *
 846              **************************/
 847
 848             r10              = _mm_mul_pd(rsq10,rinv10);
 849
 850             /* Compute parameters for interactions between i and j atoms */
 851             qq10             = _mm_mul_pd(iq1,jq0);
 852
 853             /* Calculate table index by multiplying r with table scale and truncate to integer */
 854             rt               = _mm_mul_pd(r10,vftabscale);
 855             vfitab           = _mm_cvttpd_epi32(rt);
 856 #ifdef __XOP__
 857             vfeps            = _mm_frcz_pd(rt);
 858 #else
 859             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 860 #endif
 861             twovfeps         = _mm_add_pd(vfeps,vfeps);
 862             vfitab           = _mm_slli_epi32(vfitab,2);
 863
 864             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 865             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 866             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 867             GMX_MM_TRANSPOSE2_PD(Y,F);
 868             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 869             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 870             GMX_MM_TRANSPOSE2_PD(G,H);
 871             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 872             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 873             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
 874
 875             fscal            = felec;
 876
 877             /* Update vectorial force */
 878             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 879             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 880             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 881
 882             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 883             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 884             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 885
 886             /**************************
 887              * CALCULATE INTERACTIONS *
 888              **************************/
 889
 890             r20              = _mm_mul_pd(rsq20,rinv20);
 891
 892             /* Compute parameters for interactions between i and j atoms */
 893             qq20             = _mm_mul_pd(iq2,jq0);
 894
 895             /* Calculate table index by multiplying r with table scale and truncate to integer */
 896             rt               = _mm_mul_pd(r20,vftabscale);
 897             vfitab           = _mm_cvttpd_epi32(rt);
 898 #ifdef __XOP__
 899             vfeps            = _mm_frcz_pd(rt);
 900 #else
 901             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 902 #endif
 903             twovfeps         = _mm_add_pd(vfeps,vfeps);
 904             vfitab           = _mm_slli_epi32(vfitab,2);
 905
 906             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 907             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 908             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 909             GMX_MM_TRANSPOSE2_PD(Y,F);
 910             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 911             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 912             GMX_MM_TRANSPOSE2_PD(G,H);
 913             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 914             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 915             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
 916
 917             fscal            = felec;
 918
 919             /* Update vectorial force */
 920             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 921             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 922             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 923
 924             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 925             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 926             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 927
 928             /**************************
 929              * CALCULATE INTERACTIONS *
 930              **************************/
 931
 932             r30              = _mm_mul_pd(rsq30,rinv30);
 933
 934             /* Compute parameters for interactions between i and j atoms */
 935             qq30             = _mm_mul_pd(iq3,jq0);
 936
 937             /* Calculate table index by multiplying r with table scale and truncate to integer */
 938             rt               = _mm_mul_pd(r30,vftabscale);
 939             vfitab           = _mm_cvttpd_epi32(rt);
 940 #ifdef __XOP__
 941             vfeps            = _mm_frcz_pd(rt);
 942 #else
 943             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 944 #endif
 945             twovfeps         = _mm_add_pd(vfeps,vfeps);
 946             vfitab           = _mm_slli_epi32(vfitab,2);
 947
 948             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 949             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 950             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 951             GMX_MM_TRANSPOSE2_PD(Y,F);
 952             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 953             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 954             GMX_MM_TRANSPOSE2_PD(G,H);
 955             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
 956             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
 957             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
 958
 959             fscal            = felec;
 960
 961             /* Update vectorial force */
 962             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 963             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 964             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 965
 966             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 967             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 968             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 969
 970             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 971
 972             /* Inner loop uses 159 flops */
 973         }
 974
 975         if(jidx<j_index_end)
 976         {
 977
 978             jnrA             = jjnr[jidx];
 979             j_coord_offsetA  = DIM*jnrA;
 980
 981             /* load j atom coordinates */
 982             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 983                                               &jx0,&jy0,&jz0);
 984
 985             /* Calculate displacement vector */
 986             dx00             = _mm_sub_pd(ix0,jx0);
 987             dy00             = _mm_sub_pd(iy0,jy0);
 988             dz00             = _mm_sub_pd(iz0,jz0);
 989             dx10             = _mm_sub_pd(ix1,jx0);
 990             dy10             = _mm_sub_pd(iy1,jy0);
 991             dz10             = _mm_sub_pd(iz1,jz0);
 992             dx20             = _mm_sub_pd(ix2,jx0);
 993             dy20             = _mm_sub_pd(iy2,jy0);
 994             dz20             = _mm_sub_pd(iz2,jz0);
 995             dx30             = _mm_sub_pd(ix3,jx0);
 996             dy30             = _mm_sub_pd(iy3,jy0);
 997             dz30             = _mm_sub_pd(iz3,jz0);
 998
 999             /* Calculate squared distance and things based on it */
1000             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
1001             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
1002             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
1003             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
1004
1005             rinv10           = gmx_mm_invsqrt_pd(rsq10);
1006             rinv20           = gmx_mm_invsqrt_pd(rsq20);
1007             rinv30           = gmx_mm_invsqrt_pd(rsq30);
1008
1009             rinvsq00         = gmx_mm_inv_pd(rsq00);
1010
1011             /* Load parameters for j particles */
1012             jq0              = _mm_load_sd(charge+jnrA+0);
1013             vdwjidx0A        = 2*vdwtype[jnrA+0];
1014
1015             fjx0             = _mm_setzero_pd();
1016             fjy0             = _mm_setzero_pd();
1017             fjz0             = _mm_setzero_pd();
1018
1019             /**************************
1020              * CALCULATE INTERACTIONS *
1021              **************************/
1022
1023             /* Compute parameters for interactions between i and j atoms */
1024             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
1025
1026             /* LENNARD-JONES DISPERSION/REPULSION */
1027
1028             rinvsix          = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1029             fvdw             = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
1030
1031             fscal            = fvdw;
1032
1033             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1034
1035             /* Update vectorial force */
1036             fix0             = _mm_macc_pd(dx00,fscal,fix0);
1037             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
1038             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
1039
1040             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
1041             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
1042             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
1043
1044             /**************************
1045              * CALCULATE INTERACTIONS *
1046              **************************/
1047
1048             r10              = _mm_mul_pd(rsq10,rinv10);
1049
1050             /* Compute parameters for interactions between i and j atoms */
1051             qq10             = _mm_mul_pd(iq1,jq0);
1052
1053             /* Calculate table index by multiplying r with table scale and truncate to integer */
1054             rt               = _mm_mul_pd(r10,vftabscale);
1055             vfitab           = _mm_cvttpd_epi32(rt);
1056 #ifdef __XOP__
1057             vfeps            = _mm_frcz_pd(rt);
1058 #else
1059             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1060 #endif
1061             twovfeps         = _mm_add_pd(vfeps,vfeps);
1062             vfitab           = _mm_slli_epi32(vfitab,2);
1063
1064             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1065             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1066             F                = _mm_setzero_pd();
1067             GMX_MM_TRANSPOSE2_PD(Y,F);
1068             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1069             H                = _mm_setzero_pd();
1070             GMX_MM_TRANSPOSE2_PD(G,H);
1071             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1072             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1073             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
1074
1075             fscal            = felec;
1076
1077             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1078
1079             /* Update vectorial force */
1080             fix1             = _mm_macc_pd(dx10,fscal,fix1);
1081             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
1082             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
1083
1084             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
1085             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
1086             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
1087
1088             /**************************
1089              * CALCULATE INTERACTIONS *
1090              **************************/
1091
1092             r20              = _mm_mul_pd(rsq20,rinv20);
1093
1094             /* Compute parameters for interactions between i and j atoms */
1095             qq20             = _mm_mul_pd(iq2,jq0);
1096
1097             /* Calculate table index by multiplying r with table scale and truncate to integer */
1098             rt               = _mm_mul_pd(r20,vftabscale);
1099             vfitab           = _mm_cvttpd_epi32(rt);
1100 #ifdef __XOP__
1101             vfeps            = _mm_frcz_pd(rt);
1102 #else
1103             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1104 #endif
1105             twovfeps         = _mm_add_pd(vfeps,vfeps);
1106             vfitab           = _mm_slli_epi32(vfitab,2);
1107
1108             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1109             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1110             F                = _mm_setzero_pd();
1111             GMX_MM_TRANSPOSE2_PD(Y,F);
1112             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1113             H                = _mm_setzero_pd();
1114             GMX_MM_TRANSPOSE2_PD(G,H);
1115             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1116             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1117             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
1118
1119             fscal            = felec;
1120
1121             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1122
1123             /* Update vectorial force */
1124             fix2             = _mm_macc_pd(dx20,fscal,fix2);
1125             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
1126             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
1127
1128             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
1129             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
1130             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
1131
1132             /**************************
1133              * CALCULATE INTERACTIONS *
1134              **************************/
1135
1136             r30              = _mm_mul_pd(rsq30,rinv30);
1137
1138             /* Compute parameters for interactions between i and j atoms */
1139             qq30             = _mm_mul_pd(iq3,jq0);
1140
1141             /* Calculate table index by multiplying r with table scale and truncate to integer */
1142             rt               = _mm_mul_pd(r30,vftabscale);
1143             vfitab           = _mm_cvttpd_epi32(rt);
1144 #ifdef __XOP__
1145             vfeps            = _mm_frcz_pd(rt);
1146 #else
1147             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1148 #endif
1149             twovfeps         = _mm_add_pd(vfeps,vfeps);
1150             vfitab           = _mm_slli_epi32(vfitab,2);
1151
1152             /* CUBIC SPLINE TABLE ELECTROSTATICS */
1153             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1154             F                = _mm_setzero_pd();
1155             GMX_MM_TRANSPOSE2_PD(Y,F);
1156             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1157             H                = _mm_setzero_pd();
1158             GMX_MM_TRANSPOSE2_PD(G,H);
1159             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1160             FF               = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1161             felec            = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq30,FF),_mm_mul_pd(vftabscale,rinv30)));
1162
1163             fscal            = felec;
1164
1165             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1166
1167             /* Update vectorial force */
1168             fix3             = _mm_macc_pd(dx30,fscal,fix3);
1169             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
1170             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
1171
1172             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
1173             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
1174             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
1175
1176             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
1177
1178             /* Inner loop uses 159 flops */
1179         }
1180
1181         /* End of innermost loop */
1182
1183         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1184                                               f+i_coord_offset,fshift+i_shift_offset);
1185
1186         /* Increment number of inner iterations */
1187         inneriter                  += j_index_end - j_index_start;
1188
1189         /* Outer loop uses 24 flops */
1190     }
1191
1192     /* Increment number of outer iterations */
1193     outeriter        += nri;
1194
1195     /* Update outer/inner flops */
1196
1197     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*159);
1198 }