src/gromacs/gmxlib/nonbonded/nb_kernel_avx_128_fma_double/nb_kernel_ElecRF_VdwCSTab_GeomW3W3_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_ElecRF_VdwCSTab_GeomW3W3_VF_avx_128_fma_double
  52  * Electrostatics interaction: ReactionField
  53  * VdW interaction:            CubicSplineTable
  54  * Geometry:                   Water3-Water3
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecRF_VdwCSTab_GeomW3W3_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              vdwjidx0A,vdwjidx0B;
  87     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  88     int              vdwjidx1A,vdwjidx1B;
  89     __m128d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  90     int              vdwjidx2A,vdwjidx2B;
  91     __m128d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  92     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  93     __m128d          dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
  94     __m128d          dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
  95     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  96     __m128d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
  97     __m128d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
  98     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  99     __m128d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 100     __m128d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 101     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 102     real             *charge;
 103     int              nvdwtype;
 104     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 105     int              *vdwtype;
 106     real             *vdwparam;
 107     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 108     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 109     __m128i          vfitab;
 110     __m128i          ifour       = _mm_set1_epi32(4);
 111     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 112     real             *vftab;
 113     __m128d          dummy_mask,cutoff_mask;
 114     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 115     __m128d          one     = _mm_set1_pd(1.0);
 116     __m128d          two     = _mm_set1_pd(2.0);
 117     x                = xx[0];
 118     f                = ff[0];
 119
 120     nri              = nlist->nri;
 121     iinr             = nlist->iinr;
 122     jindex           = nlist->jindex;
 123     jjnr             = nlist->jjnr;
 124     shiftidx         = nlist->shift;
 125     gid              = nlist->gid;
 126     shiftvec         = fr->shift_vec[0];
 127     fshift           = fr->fshift[0];
 128     facel            = _mm_set1_pd(fr->epsfac);
 129     charge           = mdatoms->chargeA;
 130     krf              = _mm_set1_pd(fr->ic->k_rf);
 131     krf2             = _mm_set1_pd(fr->ic->k_rf*2.0);
 132     crf              = _mm_set1_pd(fr->ic->c_rf);
 133     nvdwtype         = fr->ntype;
 134     vdwparam         = fr->nbfp;
 135     vdwtype          = mdatoms->typeA;
 136
 137     vftab            = kernel_data->table_vdw->data;
 138     vftabscale       = _mm_set1_pd(kernel_data->table_vdw->scale);
 139
 140     /* Setup water-specific parameters */
 141     inr              = nlist->iinr[0];
 142     iq0              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+0]));
 143     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 144     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 145     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 146
 147     jq0              = _mm_set1_pd(charge[inr+0]);
 148     jq1              = _mm_set1_pd(charge[inr+1]);
 149     jq2              = _mm_set1_pd(charge[inr+2]);
 150     vdwjidx0A        = 2*vdwtype[inr+0];
 151     qq00             = _mm_mul_pd(iq0,jq0);
 152     c6_00            = _mm_set1_pd(vdwparam[vdwioffset0+vdwjidx0A]);
 153     c12_00           = _mm_set1_pd(vdwparam[vdwioffset0+vdwjidx0A+1]);
 154     qq01             = _mm_mul_pd(iq0,jq1);
 155     qq02             = _mm_mul_pd(iq0,jq2);
 156     qq10             = _mm_mul_pd(iq1,jq0);
 157     qq11             = _mm_mul_pd(iq1,jq1);
 158     qq12             = _mm_mul_pd(iq1,jq2);
 159     qq20             = _mm_mul_pd(iq2,jq0);
 160     qq21             = _mm_mul_pd(iq2,jq1);
 161     qq22             = _mm_mul_pd(iq2,jq2);
 162
 163     /* Avoid stupid compiler warnings */
 164     jnrA = jnrB = 0;
 165     j_coord_offsetA = 0;
 166     j_coord_offsetB = 0;
 167
 168     outeriter        = 0;
 169     inneriter        = 0;
 170
 171     /* Start outer loop over neighborlists */
 172     for(iidx=0; iidx<nri; iidx++)
 173     {
 174         /* Load shift vector for this list */
 175         i_shift_offset   = DIM*shiftidx[iidx];
 176
 177         /* Load limits for loop over neighbors */
 178         j_index_start    = jindex[iidx];
 179         j_index_end      = jindex[iidx+1];
 180
 181         /* Get outer coordinate index */
 182         inr              = iinr[iidx];
 183         i_coord_offset   = DIM*inr;
 184
 185         /* Load i particle coords and add shift vector */
 186         gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 187                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 188
 189         fix0             = _mm_setzero_pd();
 190         fiy0             = _mm_setzero_pd();
 191         fiz0             = _mm_setzero_pd();
 192         fix1             = _mm_setzero_pd();
 193         fiy1             = _mm_setzero_pd();
 194         fiz1             = _mm_setzero_pd();
 195         fix2             = _mm_setzero_pd();
 196         fiy2             = _mm_setzero_pd();
 197         fiz2             = _mm_setzero_pd();
 198
 199         /* Reset potential sums */
 200         velecsum         = _mm_setzero_pd();
 201         vvdwsum          = _mm_setzero_pd();
 202
 203         /* Start inner kernel loop */
 204         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 205         {
 206
 207             /* Get j neighbor index, and coordinate index */
 208             jnrA             = jjnr[jidx];
 209             jnrB             = jjnr[jidx+1];
 210             j_coord_offsetA  = DIM*jnrA;
 211             j_coord_offsetB  = DIM*jnrB;
 212
 213             /* load j atom coordinates */
 214             gmx_mm_load_3rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 215                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 216
 217             /* Calculate displacement vector */
 218             dx00             = _mm_sub_pd(ix0,jx0);
 219             dy00             = _mm_sub_pd(iy0,jy0);
 220             dz00             = _mm_sub_pd(iz0,jz0);
 221             dx01             = _mm_sub_pd(ix0,jx1);
 222             dy01             = _mm_sub_pd(iy0,jy1);
 223             dz01             = _mm_sub_pd(iz0,jz1);
 224             dx02             = _mm_sub_pd(ix0,jx2);
 225             dy02             = _mm_sub_pd(iy0,jy2);
 226             dz02             = _mm_sub_pd(iz0,jz2);
 227             dx10             = _mm_sub_pd(ix1,jx0);
 228             dy10             = _mm_sub_pd(iy1,jy0);
 229             dz10             = _mm_sub_pd(iz1,jz0);
 230             dx11             = _mm_sub_pd(ix1,jx1);
 231             dy11             = _mm_sub_pd(iy1,jy1);
 232             dz11             = _mm_sub_pd(iz1,jz1);
 233             dx12             = _mm_sub_pd(ix1,jx2);
 234             dy12             = _mm_sub_pd(iy1,jy2);
 235             dz12             = _mm_sub_pd(iz1,jz2);
 236             dx20             = _mm_sub_pd(ix2,jx0);
 237             dy20             = _mm_sub_pd(iy2,jy0);
 238             dz20             = _mm_sub_pd(iz2,jz0);
 239             dx21             = _mm_sub_pd(ix2,jx1);
 240             dy21             = _mm_sub_pd(iy2,jy1);
 241             dz21             = _mm_sub_pd(iz2,jz1);
 242             dx22             = _mm_sub_pd(ix2,jx2);
 243             dy22             = _mm_sub_pd(iy2,jy2);
 244             dz22             = _mm_sub_pd(iz2,jz2);
 245
 246             /* Calculate squared distance and things based on it */
 247             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 248             rsq01            = gmx_mm_calc_rsq_pd(dx01,dy01,dz01);
 249             rsq02            = gmx_mm_calc_rsq_pd(dx02,dy02,dz02);
 250             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 251             rsq11            = gmx_mm_calc_rsq_pd(dx11,dy11,dz11);
 252             rsq12            = gmx_mm_calc_rsq_pd(dx12,dy12,dz12);
 253             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 254             rsq21            = gmx_mm_calc_rsq_pd(dx21,dy21,dz21);
 255             rsq22            = gmx_mm_calc_rsq_pd(dx22,dy22,dz22);
 256
 257             rinv00           = gmx_mm_invsqrt_pd(rsq00);
 258             rinv01           = gmx_mm_invsqrt_pd(rsq01);
 259             rinv02           = gmx_mm_invsqrt_pd(rsq02);
 260             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 261             rinv11           = gmx_mm_invsqrt_pd(rsq11);
 262             rinv12           = gmx_mm_invsqrt_pd(rsq12);
 263             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 264             rinv21           = gmx_mm_invsqrt_pd(rsq21);
 265             rinv22           = gmx_mm_invsqrt_pd(rsq22);
 266
 267             rinvsq00         = _mm_mul_pd(rinv00,rinv00);
 268             rinvsq01         = _mm_mul_pd(rinv01,rinv01);
 269             rinvsq02         = _mm_mul_pd(rinv02,rinv02);
 270             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 271             rinvsq11         = _mm_mul_pd(rinv11,rinv11);
 272             rinvsq12         = _mm_mul_pd(rinv12,rinv12);
 273             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 274             rinvsq21         = _mm_mul_pd(rinv21,rinv21);
 275             rinvsq22         = _mm_mul_pd(rinv22,rinv22);
 276
 277             fjx0             = _mm_setzero_pd();
 278             fjy0             = _mm_setzero_pd();
 279             fjz0             = _mm_setzero_pd();
 280             fjx1             = _mm_setzero_pd();
 281             fjy1             = _mm_setzero_pd();
 282             fjz1             = _mm_setzero_pd();
 283             fjx2             = _mm_setzero_pd();
 284             fjy2             = _mm_setzero_pd();
 285             fjz2             = _mm_setzero_pd();
 286
 287             /**************************
 288              * CALCULATE INTERACTIONS *
 289              **************************/
 290
 291             r00              = _mm_mul_pd(rsq00,rinv00);
 292
 293             /* Calculate table index by multiplying r with table scale and truncate to integer */
 294             rt               = _mm_mul_pd(r00,vftabscale);
 295             vfitab           = _mm_cvttpd_epi32(rt);
 296 #ifdef __XOP__
 297             vfeps            = _mm_frcz_pd(rt);
 298 #else
 299             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 300 #endif
 301             twovfeps         = _mm_add_pd(vfeps,vfeps);
 302             vfitab           = _mm_slli_epi32(vfitab,3);
 303
 304             /* REACTION-FIELD ELECTROSTATICS */
 305             velec            = _mm_mul_pd(qq00,_mm_sub_pd(_mm_macc_pd(krf,rsq00,rinv00),crf));
 306             felec            = _mm_mul_pd(qq00,_mm_msub_pd(rinv00,rinvsq00,krf2));
 307
 308             /* CUBIC SPLINE TABLE DISPERSION */
 309             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 310             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 311             GMX_MM_TRANSPOSE2_PD(Y,F);
 312             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 313             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 314             GMX_MM_TRANSPOSE2_PD(G,H);
 315             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 316             VV               = _mm_macc_pd(vfeps,Fp,Y);
 317             vvdw6            = _mm_mul_pd(c6_00,VV);
 318             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 319             fvdw6            = _mm_mul_pd(c6_00,FF);
 320
 321             /* CUBIC SPLINE TABLE REPULSION */
 322             vfitab           = _mm_add_epi32(vfitab,ifour);
 323             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 324             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 325             GMX_MM_TRANSPOSE2_PD(Y,F);
 326             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 327             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 328             GMX_MM_TRANSPOSE2_PD(G,H);
 329             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 330             VV               = _mm_macc_pd(vfeps,Fp,Y);
 331             vvdw12           = _mm_mul_pd(c12_00,VV);
 332             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 333             fvdw12           = _mm_mul_pd(c12_00,FF);
 334             vvdw             = _mm_add_pd(vvdw12,vvdw6);
 335             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
 336
 337             /* Update potential sum for this i atom from the interaction with this j atom. */
 338             velecsum         = _mm_add_pd(velecsum,velec);
 339             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 340
 341             fscal            = _mm_add_pd(felec,fvdw);
 342
 343             /* Update vectorial force */
 344             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 345             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 346             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 347
 348             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 349             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 350             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 351
 352             /**************************
 353              * CALCULATE INTERACTIONS *
 354              **************************/
 355
 356             /* REACTION-FIELD ELECTROSTATICS */
 357             velec            = _mm_mul_pd(qq01,_mm_sub_pd(_mm_macc_pd(krf,rsq01,rinv01),crf));
 358             felec            = _mm_mul_pd(qq01,_mm_msub_pd(rinv01,rinvsq01,krf2));
 359
 360             /* Update potential sum for this i atom from the interaction with this j atom. */
 361             velecsum         = _mm_add_pd(velecsum,velec);
 362
 363             fscal            = felec;
 364
 365             /* Update vectorial force */
 366             fix0             = _mm_macc_pd(dx01,fscal,fix0);
 367             fiy0             = _mm_macc_pd(dy01,fscal,fiy0);
 368             fiz0             = _mm_macc_pd(dz01,fscal,fiz0);
 369
 370             fjx1             = _mm_macc_pd(dx01,fscal,fjx1);
 371             fjy1             = _mm_macc_pd(dy01,fscal,fjy1);
 372             fjz1             = _mm_macc_pd(dz01,fscal,fjz1);
 373
 374             /**************************
 375              * CALCULATE INTERACTIONS *
 376              **************************/
 377
 378             /* REACTION-FIELD ELECTROSTATICS */
 379             velec            = _mm_mul_pd(qq02,_mm_sub_pd(_mm_macc_pd(krf,rsq02,rinv02),crf));
 380             felec            = _mm_mul_pd(qq02,_mm_msub_pd(rinv02,rinvsq02,krf2));
 381
 382             /* Update potential sum for this i atom from the interaction with this j atom. */
 383             velecsum         = _mm_add_pd(velecsum,velec);
 384
 385             fscal            = felec;
 386
 387             /* Update vectorial force */
 388             fix0             = _mm_macc_pd(dx02,fscal,fix0);
 389             fiy0             = _mm_macc_pd(dy02,fscal,fiy0);
 390             fiz0             = _mm_macc_pd(dz02,fscal,fiz0);
 391
 392             fjx2             = _mm_macc_pd(dx02,fscal,fjx2);
 393             fjy2             = _mm_macc_pd(dy02,fscal,fjy2);
 394             fjz2             = _mm_macc_pd(dz02,fscal,fjz2);
 395
 396             /**************************
 397              * CALCULATE INTERACTIONS *
 398              **************************/
 399
 400             /* REACTION-FIELD ELECTROSTATICS */
 401             velec            = _mm_mul_pd(qq10,_mm_sub_pd(_mm_macc_pd(krf,rsq10,rinv10),crf));
 402             felec            = _mm_mul_pd(qq10,_mm_msub_pd(rinv10,rinvsq10,krf2));
 403
 404             /* Update potential sum for this i atom from the interaction with this j atom. */
 405             velecsum         = _mm_add_pd(velecsum,velec);
 406
 407             fscal            = felec;
 408
 409             /* Update vectorial force */
 410             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 411             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 412             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 413
 414             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 415             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 416             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 417
 418             /**************************
 419              * CALCULATE INTERACTIONS *
 420              **************************/
 421
 422             /* REACTION-FIELD ELECTROSTATICS */
 423             velec            = _mm_mul_pd(qq11,_mm_sub_pd(_mm_macc_pd(krf,rsq11,rinv11),crf));
 424             felec            = _mm_mul_pd(qq11,_mm_msub_pd(rinv11,rinvsq11,krf2));
 425
 426             /* Update potential sum for this i atom from the interaction with this j atom. */
 427             velecsum         = _mm_add_pd(velecsum,velec);
 428
 429             fscal            = felec;
 430
 431             /* Update vectorial force */
 432             fix1             = _mm_macc_pd(dx11,fscal,fix1);
 433             fiy1             = _mm_macc_pd(dy11,fscal,fiy1);
 434             fiz1             = _mm_macc_pd(dz11,fscal,fiz1);
 435
 436             fjx1             = _mm_macc_pd(dx11,fscal,fjx1);
 437             fjy1             = _mm_macc_pd(dy11,fscal,fjy1);
 438             fjz1             = _mm_macc_pd(dz11,fscal,fjz1);
 439
 440             /**************************
 441              * CALCULATE INTERACTIONS *
 442              **************************/
 443
 444             /* REACTION-FIELD ELECTROSTATICS */
 445             velec            = _mm_mul_pd(qq12,_mm_sub_pd(_mm_macc_pd(krf,rsq12,rinv12),crf));
 446             felec            = _mm_mul_pd(qq12,_mm_msub_pd(rinv12,rinvsq12,krf2));
 447
 448             /* Update potential sum for this i atom from the interaction with this j atom. */
 449             velecsum         = _mm_add_pd(velecsum,velec);
 450
 451             fscal            = felec;
 452
 453             /* Update vectorial force */
 454             fix1             = _mm_macc_pd(dx12,fscal,fix1);
 455             fiy1             = _mm_macc_pd(dy12,fscal,fiy1);
 456             fiz1             = _mm_macc_pd(dz12,fscal,fiz1);
 457
 458             fjx2             = _mm_macc_pd(dx12,fscal,fjx2);
 459             fjy2             = _mm_macc_pd(dy12,fscal,fjy2);
 460             fjz2             = _mm_macc_pd(dz12,fscal,fjz2);
 461
 462             /**************************
 463              * CALCULATE INTERACTIONS *
 464              **************************/
 465
 466             /* REACTION-FIELD ELECTROSTATICS */
 467             velec            = _mm_mul_pd(qq20,_mm_sub_pd(_mm_macc_pd(krf,rsq20,rinv20),crf));
 468             felec            = _mm_mul_pd(qq20,_mm_msub_pd(rinv20,rinvsq20,krf2));
 469
 470             /* Update potential sum for this i atom from the interaction with this j atom. */
 471             velecsum         = _mm_add_pd(velecsum,velec);
 472
 473             fscal            = felec;
 474
 475             /* Update vectorial force */
 476             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 477             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 478             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 479
 480             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 481             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 482             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 483
 484             /**************************
 485              * CALCULATE INTERACTIONS *
 486              **************************/
 487
 488             /* REACTION-FIELD ELECTROSTATICS */
 489             velec            = _mm_mul_pd(qq21,_mm_sub_pd(_mm_macc_pd(krf,rsq21,rinv21),crf));
 490             felec            = _mm_mul_pd(qq21,_mm_msub_pd(rinv21,rinvsq21,krf2));
 491
 492             /* Update potential sum for this i atom from the interaction with this j atom. */
 493             velecsum         = _mm_add_pd(velecsum,velec);
 494
 495             fscal            = felec;
 496
 497             /* Update vectorial force */
 498             fix2             = _mm_macc_pd(dx21,fscal,fix2);
 499             fiy2             = _mm_macc_pd(dy21,fscal,fiy2);
 500             fiz2             = _mm_macc_pd(dz21,fscal,fiz2);
 501
 502             fjx1             = _mm_macc_pd(dx21,fscal,fjx1);
 503             fjy1             = _mm_macc_pd(dy21,fscal,fjy1);
 504             fjz1             = _mm_macc_pd(dz21,fscal,fjz1);
 505
 506             /**************************
 507              * CALCULATE INTERACTIONS *
 508              **************************/
 509
 510             /* REACTION-FIELD ELECTROSTATICS */
 511             velec            = _mm_mul_pd(qq22,_mm_sub_pd(_mm_macc_pd(krf,rsq22,rinv22),crf));
 512             felec            = _mm_mul_pd(qq22,_mm_msub_pd(rinv22,rinvsq22,krf2));
 513
 514             /* Update potential sum for this i atom from the interaction with this j atom. */
 515             velecsum         = _mm_add_pd(velecsum,velec);
 516
 517             fscal            = felec;
 518
 519             /* Update vectorial force */
 520             fix2             = _mm_macc_pd(dx22,fscal,fix2);
 521             fiy2             = _mm_macc_pd(dy22,fscal,fiy2);
 522             fiz2             = _mm_macc_pd(dz22,fscal,fiz2);
 523
 524             fjx2             = _mm_macc_pd(dx22,fscal,fjx2);
 525             fjy2             = _mm_macc_pd(dy22,fscal,fjy2);
 526             fjz2             = _mm_macc_pd(dz22,fscal,fjz2);
 527
 528             gmx_mm_decrement_3rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 529
 530             /* Inner loop uses 350 flops */
 531         }
 532
 533         if(jidx<j_index_end)
 534         {
 535
 536             jnrA             = jjnr[jidx];
 537             j_coord_offsetA  = DIM*jnrA;
 538
 539             /* load j atom coordinates */
 540             gmx_mm_load_3rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 541                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 542
 543             /* Calculate displacement vector */
 544             dx00             = _mm_sub_pd(ix0,jx0);
 545             dy00             = _mm_sub_pd(iy0,jy0);
 546             dz00             = _mm_sub_pd(iz0,jz0);
 547             dx01             = _mm_sub_pd(ix0,jx1);
 548             dy01             = _mm_sub_pd(iy0,jy1);
 549             dz01             = _mm_sub_pd(iz0,jz1);
 550             dx02             = _mm_sub_pd(ix0,jx2);
 551             dy02             = _mm_sub_pd(iy0,jy2);
 552             dz02             = _mm_sub_pd(iz0,jz2);
 553             dx10             = _mm_sub_pd(ix1,jx0);
 554             dy10             = _mm_sub_pd(iy1,jy0);
 555             dz10             = _mm_sub_pd(iz1,jz0);
 556             dx11             = _mm_sub_pd(ix1,jx1);
 557             dy11             = _mm_sub_pd(iy1,jy1);
 558             dz11             = _mm_sub_pd(iz1,jz1);
 559             dx12             = _mm_sub_pd(ix1,jx2);
 560             dy12             = _mm_sub_pd(iy1,jy2);
 561             dz12             = _mm_sub_pd(iz1,jz2);
 562             dx20             = _mm_sub_pd(ix2,jx0);
 563             dy20             = _mm_sub_pd(iy2,jy0);
 564             dz20             = _mm_sub_pd(iz2,jz0);
 565             dx21             = _mm_sub_pd(ix2,jx1);
 566             dy21             = _mm_sub_pd(iy2,jy1);
 567             dz21             = _mm_sub_pd(iz2,jz1);
 568             dx22             = _mm_sub_pd(ix2,jx2);
 569             dy22             = _mm_sub_pd(iy2,jy2);
 570             dz22             = _mm_sub_pd(iz2,jz2);
 571
 572             /* Calculate squared distance and things based on it */
 573             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 574             rsq01            = gmx_mm_calc_rsq_pd(dx01,dy01,dz01);
 575             rsq02            = gmx_mm_calc_rsq_pd(dx02,dy02,dz02);
 576             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 577             rsq11            = gmx_mm_calc_rsq_pd(dx11,dy11,dz11);
 578             rsq12            = gmx_mm_calc_rsq_pd(dx12,dy12,dz12);
 579             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 580             rsq21            = gmx_mm_calc_rsq_pd(dx21,dy21,dz21);
 581             rsq22            = gmx_mm_calc_rsq_pd(dx22,dy22,dz22);
 582
 583             rinv00           = gmx_mm_invsqrt_pd(rsq00);
 584             rinv01           = gmx_mm_invsqrt_pd(rsq01);
 585             rinv02           = gmx_mm_invsqrt_pd(rsq02);
 586             rinv10           = gmx_mm_invsqrt_pd(rsq10);
 587             rinv11           = gmx_mm_invsqrt_pd(rsq11);
 588             rinv12           = gmx_mm_invsqrt_pd(rsq12);
 589             rinv20           = gmx_mm_invsqrt_pd(rsq20);
 590             rinv21           = gmx_mm_invsqrt_pd(rsq21);
 591             rinv22           = gmx_mm_invsqrt_pd(rsq22);
 592
 593             rinvsq00         = _mm_mul_pd(rinv00,rinv00);
 594             rinvsq01         = _mm_mul_pd(rinv01,rinv01);
 595             rinvsq02         = _mm_mul_pd(rinv02,rinv02);
 596             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 597             rinvsq11         = _mm_mul_pd(rinv11,rinv11);
 598             rinvsq12         = _mm_mul_pd(rinv12,rinv12);
 599             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 600             rinvsq21         = _mm_mul_pd(rinv21,rinv21);
 601             rinvsq22         = _mm_mul_pd(rinv22,rinv22);
 602
 603             fjx0             = _mm_setzero_pd();
 604             fjy0             = _mm_setzero_pd();
 605             fjz0             = _mm_setzero_pd();
 606             fjx1             = _mm_setzero_pd();
 607             fjy1             = _mm_setzero_pd();
 608             fjz1             = _mm_setzero_pd();
 609             fjx2             = _mm_setzero_pd();
 610             fjy2             = _mm_setzero_pd();
 611             fjz2             = _mm_setzero_pd();
 612
 613             /**************************
 614              * CALCULATE INTERACTIONS *
 615              **************************/
 616
 617             r00              = _mm_mul_pd(rsq00,rinv00);
 618
 619             /* Calculate table index by multiplying r with table scale and truncate to integer */
 620             rt               = _mm_mul_pd(r00,vftabscale);
 621             vfitab           = _mm_cvttpd_epi32(rt);
 622 #ifdef __XOP__
 623             vfeps            = _mm_frcz_pd(rt);
 624 #else
 625             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 626 #endif
 627             twovfeps         = _mm_add_pd(vfeps,vfeps);
 628             vfitab           = _mm_slli_epi32(vfitab,3);
 629
 630             /* REACTION-FIELD ELECTROSTATICS */
 631             velec            = _mm_mul_pd(qq00,_mm_sub_pd(_mm_macc_pd(krf,rsq00,rinv00),crf));
 632             felec            = _mm_mul_pd(qq00,_mm_msub_pd(rinv00,rinvsq00,krf2));
 633
 634             /* CUBIC SPLINE TABLE DISPERSION */
 635             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 636             F                = _mm_setzero_pd();
 637             GMX_MM_TRANSPOSE2_PD(Y,F);
 638             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 639             H                = _mm_setzero_pd();
 640             GMX_MM_TRANSPOSE2_PD(G,H);
 641             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 642             VV               = _mm_macc_pd(vfeps,Fp,Y);
 643             vvdw6            = _mm_mul_pd(c6_00,VV);
 644             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 645             fvdw6            = _mm_mul_pd(c6_00,FF);
 646
 647             /* CUBIC SPLINE TABLE REPULSION */
 648             vfitab           = _mm_add_epi32(vfitab,ifour);
 649             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 650             F                = _mm_setzero_pd();
 651             GMX_MM_TRANSPOSE2_PD(Y,F);
 652             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 653             H                = _mm_setzero_pd();
 654             GMX_MM_TRANSPOSE2_PD(G,H);
 655             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 656             VV               = _mm_macc_pd(vfeps,Fp,Y);
 657             vvdw12           = _mm_mul_pd(c12_00,VV);
 658             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 659             fvdw12           = _mm_mul_pd(c12_00,FF);
 660             vvdw             = _mm_add_pd(vvdw12,vvdw6);
 661             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
 662
 663             /* Update potential sum for this i atom from the interaction with this j atom. */
 664             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 665             velecsum         = _mm_add_pd(velecsum,velec);
 666             vvdw             = _mm_unpacklo_pd(vvdw,_mm_setzero_pd());
 667             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 668
 669             fscal            = _mm_add_pd(felec,fvdw);
 670
 671             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 672
 673             /* Update vectorial force */
 674             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 675             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 676             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 677
 678             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 679             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 680             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 681
 682             /**************************
 683              * CALCULATE INTERACTIONS *
 684              **************************/
 685
 686             /* REACTION-FIELD ELECTROSTATICS */
 687             velec            = _mm_mul_pd(qq01,_mm_sub_pd(_mm_macc_pd(krf,rsq01,rinv01),crf));
 688             felec            = _mm_mul_pd(qq01,_mm_msub_pd(rinv01,rinvsq01,krf2));
 689
 690             /* Update potential sum for this i atom from the interaction with this j atom. */
 691             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 692             velecsum         = _mm_add_pd(velecsum,velec);
 693
 694             fscal            = felec;
 695
 696             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 697
 698             /* Update vectorial force */
 699             fix0             = _mm_macc_pd(dx01,fscal,fix0);
 700             fiy0             = _mm_macc_pd(dy01,fscal,fiy0);
 701             fiz0             = _mm_macc_pd(dz01,fscal,fiz0);
 702
 703             fjx1             = _mm_macc_pd(dx01,fscal,fjx1);
 704             fjy1             = _mm_macc_pd(dy01,fscal,fjy1);
 705             fjz1             = _mm_macc_pd(dz01,fscal,fjz1);
 706
 707             /**************************
 708              * CALCULATE INTERACTIONS *
 709              **************************/
 710
 711             /* REACTION-FIELD ELECTROSTATICS */
 712             velec            = _mm_mul_pd(qq02,_mm_sub_pd(_mm_macc_pd(krf,rsq02,rinv02),crf));
 713             felec            = _mm_mul_pd(qq02,_mm_msub_pd(rinv02,rinvsq02,krf2));
 714
 715             /* Update potential sum for this i atom from the interaction with this j atom. */
 716             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 717             velecsum         = _mm_add_pd(velecsum,velec);
 718
 719             fscal            = felec;
 720
 721             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 722
 723             /* Update vectorial force */
 724             fix0             = _mm_macc_pd(dx02,fscal,fix0);
 725             fiy0             = _mm_macc_pd(dy02,fscal,fiy0);
 726             fiz0             = _mm_macc_pd(dz02,fscal,fiz0);
 727
 728             fjx2             = _mm_macc_pd(dx02,fscal,fjx2);
 729             fjy2             = _mm_macc_pd(dy02,fscal,fjy2);
 730             fjz2             = _mm_macc_pd(dz02,fscal,fjz2);
 731
 732             /**************************
 733              * CALCULATE INTERACTIONS *
 734              **************************/
 735
 736             /* REACTION-FIELD ELECTROSTATICS */
 737             velec            = _mm_mul_pd(qq10,_mm_sub_pd(_mm_macc_pd(krf,rsq10,rinv10),crf));
 738             felec            = _mm_mul_pd(qq10,_mm_msub_pd(rinv10,rinvsq10,krf2));
 739
 740             /* Update potential sum for this i atom from the interaction with this j atom. */
 741             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 742             velecsum         = _mm_add_pd(velecsum,velec);
 743
 744             fscal            = felec;
 745
 746             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 747
 748             /* Update vectorial force */
 749             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 750             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 751             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 752
 753             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 754             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 755             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 756
 757             /**************************
 758              * CALCULATE INTERACTIONS *
 759              **************************/
 760
 761             /* REACTION-FIELD ELECTROSTATICS */
 762             velec            = _mm_mul_pd(qq11,_mm_sub_pd(_mm_macc_pd(krf,rsq11,rinv11),crf));
 763             felec            = _mm_mul_pd(qq11,_mm_msub_pd(rinv11,rinvsq11,krf2));
 764
 765             /* Update potential sum for this i atom from the interaction with this j atom. */
 766             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 767             velecsum         = _mm_add_pd(velecsum,velec);
 768
 769             fscal            = felec;
 770
 771             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 772
 773             /* Update vectorial force */
 774             fix1             = _mm_macc_pd(dx11,fscal,fix1);
 775             fiy1             = _mm_macc_pd(dy11,fscal,fiy1);
 776             fiz1             = _mm_macc_pd(dz11,fscal,fiz1);
 777
 778             fjx1             = _mm_macc_pd(dx11,fscal,fjx1);
 779             fjy1             = _mm_macc_pd(dy11,fscal,fjy1);
 780             fjz1             = _mm_macc_pd(dz11,fscal,fjz1);
 781
 782             /**************************
 783              * CALCULATE INTERACTIONS *
 784              **************************/
 785
 786             /* REACTION-FIELD ELECTROSTATICS */
 787             velec            = _mm_mul_pd(qq12,_mm_sub_pd(_mm_macc_pd(krf,rsq12,rinv12),crf));
 788             felec            = _mm_mul_pd(qq12,_mm_msub_pd(rinv12,rinvsq12,krf2));
 789
 790             /* Update potential sum for this i atom from the interaction with this j atom. */
 791             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 792             velecsum         = _mm_add_pd(velecsum,velec);
 793
 794             fscal            = felec;
 795
 796             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 797
 798             /* Update vectorial force */
 799             fix1             = _mm_macc_pd(dx12,fscal,fix1);
 800             fiy1             = _mm_macc_pd(dy12,fscal,fiy1);
 801             fiz1             = _mm_macc_pd(dz12,fscal,fiz1);
 802
 803             fjx2             = _mm_macc_pd(dx12,fscal,fjx2);
 804             fjy2             = _mm_macc_pd(dy12,fscal,fjy2);
 805             fjz2             = _mm_macc_pd(dz12,fscal,fjz2);
 806
 807             /**************************
 808              * CALCULATE INTERACTIONS *
 809              **************************/
 810
 811             /* REACTION-FIELD ELECTROSTATICS */
 812             velec            = _mm_mul_pd(qq20,_mm_sub_pd(_mm_macc_pd(krf,rsq20,rinv20),crf));
 813             felec            = _mm_mul_pd(qq20,_mm_msub_pd(rinv20,rinvsq20,krf2));
 814
 815             /* Update potential sum for this i atom from the interaction with this j atom. */
 816             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 817             velecsum         = _mm_add_pd(velecsum,velec);
 818
 819             fscal            = felec;
 820
 821             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 822
 823             /* Update vectorial force */
 824             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 825             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 826             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 827
 828             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 829             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 830             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 831
 832             /**************************
 833              * CALCULATE INTERACTIONS *
 834              **************************/
 835
 836             /* REACTION-FIELD ELECTROSTATICS */
 837             velec            = _mm_mul_pd(qq21,_mm_sub_pd(_mm_macc_pd(krf,rsq21,rinv21),crf));
 838             felec            = _mm_mul_pd(qq21,_mm_msub_pd(rinv21,rinvsq21,krf2));
 839
 840             /* Update potential sum for this i atom from the interaction with this j atom. */
 841             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 842             velecsum         = _mm_add_pd(velecsum,velec);
 843
 844             fscal            = felec;
 845
 846             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 847
 848             /* Update vectorial force */
 849             fix2             = _mm_macc_pd(dx21,fscal,fix2);
 850             fiy2             = _mm_macc_pd(dy21,fscal,fiy2);
 851             fiz2             = _mm_macc_pd(dz21,fscal,fiz2);
 852
 853             fjx1             = _mm_macc_pd(dx21,fscal,fjx1);
 854             fjy1             = _mm_macc_pd(dy21,fscal,fjy1);
 855             fjz1             = _mm_macc_pd(dz21,fscal,fjz1);
 856
 857             /**************************
 858              * CALCULATE INTERACTIONS *
 859              **************************/
 860
 861             /* REACTION-FIELD ELECTROSTATICS */
 862             velec            = _mm_mul_pd(qq22,_mm_sub_pd(_mm_macc_pd(krf,rsq22,rinv22),crf));
 863             felec            = _mm_mul_pd(qq22,_mm_msub_pd(rinv22,rinvsq22,krf2));
 864
 865             /* Update potential sum for this i atom from the interaction with this j atom. */
 866             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 867             velecsum         = _mm_add_pd(velecsum,velec);
 868
 869             fscal            = felec;
 870
 871             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 872
 873             /* Update vectorial force */
 874             fix2             = _mm_macc_pd(dx22,fscal,fix2);
 875             fiy2             = _mm_macc_pd(dy22,fscal,fiy2);
 876             fiz2             = _mm_macc_pd(dz22,fscal,fiz2);
 877
 878             fjx2             = _mm_macc_pd(dx22,fscal,fjx2);
 879             fjy2             = _mm_macc_pd(dy22,fscal,fjy2);
 880             fjz2             = _mm_macc_pd(dz22,fscal,fjz2);
 881
 882             gmx_mm_decrement_3rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 883
 884             /* Inner loop uses 350 flops */
 885         }
 886
 887         /* End of innermost loop */
 888
 889         gmx_mm_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
 890                                               f+i_coord_offset,fshift+i_shift_offset);
 891
 892         ggid                        = gid[iidx];
 893         /* Update potential energies */
 894         gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 895         gmx_mm_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 896
 897         /* Increment number of inner iterations */
 898         inneriter                  += j_index_end - j_index_start;
 899
 900         /* Outer loop uses 20 flops */
 901     }
 902
 903     /* Increment number of outer iterations */
 904     outeriter        += nri;
 905
 906     /* Update outer/inner flops */
 907
 908     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*20 + inneriter*350);
 909 }
 910 /*
 911  * Gromacs nonbonded kernel:   nb_kernel_ElecRF_VdwCSTab_GeomW3W3_F_avx_128_fma_double
 912  * Electrostatics interaction: ReactionField
 913  * VdW interaction:            CubicSplineTable
 914  * Geometry:                   Water3-Water3
 915  * Calculate force/pot:        Force
 916  */
 917 void
 918 nb_kernel_ElecRF_VdwCSTab_GeomW3W3_F_avx_128_fma_double
 919                     (t_nblist                    * gmx_restrict       nlist,
 920                      rvec                        * gmx_restrict          xx,
 921                      rvec                        * gmx_restrict          ff,
 922                      t_forcerec                  * gmx_restrict          fr,
 923                      t_mdatoms                   * gmx_restrict     mdatoms,
 924                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 925                      t_nrnb                      * gmx_restrict        nrnb)
 926 {
 927     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 928      * just 0 for non-waters.
 929      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
 930      * jnr indices corresponding to data put in the four positions in the SIMD register.
 931      */
 932     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 933     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 934     int              jnrA,jnrB;
 935     int              j_coord_offsetA,j_coord_offsetB;
 936     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 937     real             rcutoff_scalar;
 938     real             *shiftvec,*fshift,*x,*f;
 939     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 940     int              vdwioffset0;
 941     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 942     int              vdwioffset1;
 943     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 944     int              vdwioffset2;
 945     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 946     int              vdwjidx0A,vdwjidx0B;
 947     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 948     int              vdwjidx1A,vdwjidx1B;
 949     __m128d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
 950     int              vdwjidx2A,vdwjidx2B;
 951     __m128d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
 952     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 953     __m128d          dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
 954     __m128d          dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
 955     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 956     __m128d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
 957     __m128d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
 958     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 959     __m128d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 960     __m128d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 961     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 962     real             *charge;
 963     int              nvdwtype;
 964     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 965     int              *vdwtype;
 966     real             *vdwparam;
 967     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 968     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 969     __m128i          vfitab;
 970     __m128i          ifour       = _mm_set1_epi32(4);
 971     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 972     real             *vftab;
 973     __m128d          dummy_mask,cutoff_mask;
 974     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 975     __m128d          one     = _mm_set1_pd(1.0);
 976     __m128d          two     = _mm_set1_pd(2.0);
 977     x                = xx[0];
 978     f                = ff[0];
 979
 980     nri              = nlist->nri;
 981     iinr             = nlist->iinr;
 982     jindex           = nlist->jindex;
 983     jjnr             = nlist->jjnr;
 984     shiftidx         = nlist->shift;
 985     gid              = nlist->gid;
 986     shiftvec         = fr->shift_vec[0];
 987     fshift           = fr->fshift[0];
 988     facel            = _mm_set1_pd(fr->epsfac);
 989     charge           = mdatoms->chargeA;
 990     krf              = _mm_set1_pd(fr->ic->k_rf);
 991     krf2             = _mm_set1_pd(fr->ic->k_rf*2.0);
 992     crf              = _mm_set1_pd(fr->ic->c_rf);
 993     nvdwtype         = fr->ntype;
 994     vdwparam         = fr->nbfp;
 995     vdwtype          = mdatoms->typeA;
 996
 997     vftab            = kernel_data->table_vdw->data;
 998     vftabscale       = _mm_set1_pd(kernel_data->table_vdw->scale);
 999
1000     /* Setup water-specific parameters */
1001     inr              = nlist->iinr[0];
1002     iq0              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+0]));
1003     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
1004     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
1005     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
1006
1007     jq0              = _mm_set1_pd(charge[inr+0]);
1008     jq1              = _mm_set1_pd(charge[inr+1]);
1009     jq2              = _mm_set1_pd(charge[inr+2]);
1010     vdwjidx0A        = 2*vdwtype[inr+0];
1011     qq00             = _mm_mul_pd(iq0,jq0);
1012     c6_00            = _mm_set1_pd(vdwparam[vdwioffset0+vdwjidx0A]);
1013     c12_00           = _mm_set1_pd(vdwparam[vdwioffset0+vdwjidx0A+1]);
1014     qq01             = _mm_mul_pd(iq0,jq1);
1015     qq02             = _mm_mul_pd(iq0,jq2);
1016     qq10             = _mm_mul_pd(iq1,jq0);
1017     qq11             = _mm_mul_pd(iq1,jq1);
1018     qq12             = _mm_mul_pd(iq1,jq2);
1019     qq20             = _mm_mul_pd(iq2,jq0);
1020     qq21             = _mm_mul_pd(iq2,jq1);
1021     qq22             = _mm_mul_pd(iq2,jq2);
1022
1023     /* Avoid stupid compiler warnings */
1024     jnrA = jnrB = 0;
1025     j_coord_offsetA = 0;
1026     j_coord_offsetB = 0;
1027
1028     outeriter        = 0;
1029     inneriter        = 0;
1030
1031     /* Start outer loop over neighborlists */
1032     for(iidx=0; iidx<nri; iidx++)
1033     {
1034         /* Load shift vector for this list */
1035         i_shift_offset   = DIM*shiftidx[iidx];
1036
1037         /* Load limits for loop over neighbors */
1038         j_index_start    = jindex[iidx];
1039         j_index_end      = jindex[iidx+1];
1040
1041         /* Get outer coordinate index */
1042         inr              = iinr[iidx];
1043         i_coord_offset   = DIM*inr;
1044
1045         /* Load i particle coords and add shift vector */
1046         gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
1047                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
1048
1049         fix0             = _mm_setzero_pd();
1050         fiy0             = _mm_setzero_pd();
1051         fiz0             = _mm_setzero_pd();
1052         fix1             = _mm_setzero_pd();
1053         fiy1             = _mm_setzero_pd();
1054         fiz1             = _mm_setzero_pd();
1055         fix2             = _mm_setzero_pd();
1056         fiy2             = _mm_setzero_pd();
1057         fiz2             = _mm_setzero_pd();
1058
1059         /* Start inner kernel loop */
1060         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
1061         {
1062
1063             /* Get j neighbor index, and coordinate index */
1064             jnrA             = jjnr[jidx];
1065             jnrB             = jjnr[jidx+1];
1066             j_coord_offsetA  = DIM*jnrA;
1067             j_coord_offsetB  = DIM*jnrB;
1068
1069             /* load j atom coordinates */
1070             gmx_mm_load_3rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1071                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1072
1073             /* Calculate displacement vector */
1074             dx00             = _mm_sub_pd(ix0,jx0);
1075             dy00             = _mm_sub_pd(iy0,jy0);
1076             dz00             = _mm_sub_pd(iz0,jz0);
1077             dx01             = _mm_sub_pd(ix0,jx1);
1078             dy01             = _mm_sub_pd(iy0,jy1);
1079             dz01             = _mm_sub_pd(iz0,jz1);
1080             dx02             = _mm_sub_pd(ix0,jx2);
1081             dy02             = _mm_sub_pd(iy0,jy2);
1082             dz02             = _mm_sub_pd(iz0,jz2);
1083             dx10             = _mm_sub_pd(ix1,jx0);
1084             dy10             = _mm_sub_pd(iy1,jy0);
1085             dz10             = _mm_sub_pd(iz1,jz0);
1086             dx11             = _mm_sub_pd(ix1,jx1);
1087             dy11             = _mm_sub_pd(iy1,jy1);
1088             dz11             = _mm_sub_pd(iz1,jz1);
1089             dx12             = _mm_sub_pd(ix1,jx2);
1090             dy12             = _mm_sub_pd(iy1,jy2);
1091             dz12             = _mm_sub_pd(iz1,jz2);
1092             dx20             = _mm_sub_pd(ix2,jx0);
1093             dy20             = _mm_sub_pd(iy2,jy0);
1094             dz20             = _mm_sub_pd(iz2,jz0);
1095             dx21             = _mm_sub_pd(ix2,jx1);
1096             dy21             = _mm_sub_pd(iy2,jy1);
1097             dz21             = _mm_sub_pd(iz2,jz1);
1098             dx22             = _mm_sub_pd(ix2,jx2);
1099             dy22             = _mm_sub_pd(iy2,jy2);
1100             dz22             = _mm_sub_pd(iz2,jz2);
1101
1102             /* Calculate squared distance and things based on it */
1103             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
1104             rsq01            = gmx_mm_calc_rsq_pd(dx01,dy01,dz01);
1105             rsq02            = gmx_mm_calc_rsq_pd(dx02,dy02,dz02);
1106             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
1107             rsq11            = gmx_mm_calc_rsq_pd(dx11,dy11,dz11);
1108             rsq12            = gmx_mm_calc_rsq_pd(dx12,dy12,dz12);
1109             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
1110             rsq21            = gmx_mm_calc_rsq_pd(dx21,dy21,dz21);
1111             rsq22            = gmx_mm_calc_rsq_pd(dx22,dy22,dz22);
1112
1113             rinv00           = gmx_mm_invsqrt_pd(rsq00);
1114             rinv01           = gmx_mm_invsqrt_pd(rsq01);
1115             rinv02           = gmx_mm_invsqrt_pd(rsq02);
1116             rinv10           = gmx_mm_invsqrt_pd(rsq10);
1117             rinv11           = gmx_mm_invsqrt_pd(rsq11);
1118             rinv12           = gmx_mm_invsqrt_pd(rsq12);
1119             rinv20           = gmx_mm_invsqrt_pd(rsq20);
1120             rinv21           = gmx_mm_invsqrt_pd(rsq21);
1121             rinv22           = gmx_mm_invsqrt_pd(rsq22);
1122
1123             rinvsq00         = _mm_mul_pd(rinv00,rinv00);
1124             rinvsq01         = _mm_mul_pd(rinv01,rinv01);
1125             rinvsq02         = _mm_mul_pd(rinv02,rinv02);
1126             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
1127             rinvsq11         = _mm_mul_pd(rinv11,rinv11);
1128             rinvsq12         = _mm_mul_pd(rinv12,rinv12);
1129             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
1130             rinvsq21         = _mm_mul_pd(rinv21,rinv21);
1131             rinvsq22         = _mm_mul_pd(rinv22,rinv22);
1132
1133             fjx0             = _mm_setzero_pd();
1134             fjy0             = _mm_setzero_pd();
1135             fjz0             = _mm_setzero_pd();
1136             fjx1             = _mm_setzero_pd();
1137             fjy1             = _mm_setzero_pd();
1138             fjz1             = _mm_setzero_pd();
1139             fjx2             = _mm_setzero_pd();
1140             fjy2             = _mm_setzero_pd();
1141             fjz2             = _mm_setzero_pd();
1142
1143             /**************************
1144              * CALCULATE INTERACTIONS *
1145              **************************/
1146
1147             r00              = _mm_mul_pd(rsq00,rinv00);
1148
1149             /* Calculate table index by multiplying r with table scale and truncate to integer */
1150             rt               = _mm_mul_pd(r00,vftabscale);
1151             vfitab           = _mm_cvttpd_epi32(rt);
1152 #ifdef __XOP__
1153             vfeps            = _mm_frcz_pd(rt);
1154 #else
1155             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1156 #endif
1157             twovfeps         = _mm_add_pd(vfeps,vfeps);
1158             vfitab           = _mm_slli_epi32(vfitab,3);
1159
1160             /* REACTION-FIELD ELECTROSTATICS */
1161             felec            = _mm_mul_pd(qq00,_mm_msub_pd(rinv00,rinvsq00,krf2));
1162
1163             /* CUBIC SPLINE TABLE DISPERSION */
1164             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1165             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
1166             GMX_MM_TRANSPOSE2_PD(Y,F);
1167             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1168             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
1169             GMX_MM_TRANSPOSE2_PD(G,H);
1170             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
1171             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
1172             fvdw6            = _mm_mul_pd(c6_00,FF);
1173
1174             /* CUBIC SPLINE TABLE REPULSION */
1175             vfitab           = _mm_add_epi32(vfitab,ifour);
1176             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1177             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
1178             GMX_MM_TRANSPOSE2_PD(Y,F);
1179             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1180             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
1181             GMX_MM_TRANSPOSE2_PD(G,H);
1182             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
1183             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
1184             fvdw12           = _mm_mul_pd(c12_00,FF);
1185             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
1186
1187             fscal            = _mm_add_pd(felec,fvdw);
1188
1189             /* Update vectorial force */
1190             fix0             = _mm_macc_pd(dx00,fscal,fix0);
1191             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
1192             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
1193
1194             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
1195             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
1196             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
1197
1198             /**************************
1199              * CALCULATE INTERACTIONS *
1200              **************************/
1201
1202             /* REACTION-FIELD ELECTROSTATICS */
1203             felec            = _mm_mul_pd(qq01,_mm_msub_pd(rinv01,rinvsq01,krf2));
1204
1205             fscal            = felec;
1206
1207             /* Update vectorial force */
1208             fix0             = _mm_macc_pd(dx01,fscal,fix0);
1209             fiy0             = _mm_macc_pd(dy01,fscal,fiy0);
1210             fiz0             = _mm_macc_pd(dz01,fscal,fiz0);
1211
1212             fjx1             = _mm_macc_pd(dx01,fscal,fjx1);
1213             fjy1             = _mm_macc_pd(dy01,fscal,fjy1);
1214             fjz1             = _mm_macc_pd(dz01,fscal,fjz1);
1215
1216             /**************************
1217              * CALCULATE INTERACTIONS *
1218              **************************/
1219
1220             /* REACTION-FIELD ELECTROSTATICS */
1221             felec            = _mm_mul_pd(qq02,_mm_msub_pd(rinv02,rinvsq02,krf2));
1222
1223             fscal            = felec;
1224
1225             /* Update vectorial force */
1226             fix0             = _mm_macc_pd(dx02,fscal,fix0);
1227             fiy0             = _mm_macc_pd(dy02,fscal,fiy0);
1228             fiz0             = _mm_macc_pd(dz02,fscal,fiz0);
1229
1230             fjx2             = _mm_macc_pd(dx02,fscal,fjx2);
1231             fjy2             = _mm_macc_pd(dy02,fscal,fjy2);
1232             fjz2             = _mm_macc_pd(dz02,fscal,fjz2);
1233
1234             /**************************
1235              * CALCULATE INTERACTIONS *
1236              **************************/
1237
1238             /* REACTION-FIELD ELECTROSTATICS */
1239             felec            = _mm_mul_pd(qq10,_mm_msub_pd(rinv10,rinvsq10,krf2));
1240
1241             fscal            = felec;
1242
1243             /* Update vectorial force */
1244             fix1             = _mm_macc_pd(dx10,fscal,fix1);
1245             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
1246             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
1247
1248             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
1249             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
1250             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
1251
1252             /**************************
1253              * CALCULATE INTERACTIONS *
1254              **************************/
1255
1256             /* REACTION-FIELD ELECTROSTATICS */
1257             felec            = _mm_mul_pd(qq11,_mm_msub_pd(rinv11,rinvsq11,krf2));
1258
1259             fscal            = felec;
1260
1261             /* Update vectorial force */
1262             fix1             = _mm_macc_pd(dx11,fscal,fix1);
1263             fiy1             = _mm_macc_pd(dy11,fscal,fiy1);
1264             fiz1             = _mm_macc_pd(dz11,fscal,fiz1);
1265
1266             fjx1             = _mm_macc_pd(dx11,fscal,fjx1);
1267             fjy1             = _mm_macc_pd(dy11,fscal,fjy1);
1268             fjz1             = _mm_macc_pd(dz11,fscal,fjz1);
1269
1270             /**************************
1271              * CALCULATE INTERACTIONS *
1272              **************************/
1273
1274             /* REACTION-FIELD ELECTROSTATICS */
1275             felec            = _mm_mul_pd(qq12,_mm_msub_pd(rinv12,rinvsq12,krf2));
1276
1277             fscal            = felec;
1278
1279             /* Update vectorial force */
1280             fix1             = _mm_macc_pd(dx12,fscal,fix1);
1281             fiy1             = _mm_macc_pd(dy12,fscal,fiy1);
1282             fiz1             = _mm_macc_pd(dz12,fscal,fiz1);
1283
1284             fjx2             = _mm_macc_pd(dx12,fscal,fjx2);
1285             fjy2             = _mm_macc_pd(dy12,fscal,fjy2);
1286             fjz2             = _mm_macc_pd(dz12,fscal,fjz2);
1287
1288             /**************************
1289              * CALCULATE INTERACTIONS *
1290              **************************/
1291
1292             /* REACTION-FIELD ELECTROSTATICS */
1293             felec            = _mm_mul_pd(qq20,_mm_msub_pd(rinv20,rinvsq20,krf2));
1294
1295             fscal            = felec;
1296
1297             /* Update vectorial force */
1298             fix2             = _mm_macc_pd(dx20,fscal,fix2);
1299             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
1300             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
1301
1302             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
1303             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
1304             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
1305
1306             /**************************
1307              * CALCULATE INTERACTIONS *
1308              **************************/
1309
1310             /* REACTION-FIELD ELECTROSTATICS */
1311             felec            = _mm_mul_pd(qq21,_mm_msub_pd(rinv21,rinvsq21,krf2));
1312
1313             fscal            = felec;
1314
1315             /* Update vectorial force */
1316             fix2             = _mm_macc_pd(dx21,fscal,fix2);
1317             fiy2             = _mm_macc_pd(dy21,fscal,fiy2);
1318             fiz2             = _mm_macc_pd(dz21,fscal,fiz2);
1319
1320             fjx1             = _mm_macc_pd(dx21,fscal,fjx1);
1321             fjy1             = _mm_macc_pd(dy21,fscal,fjy1);
1322             fjz1             = _mm_macc_pd(dz21,fscal,fjz1);
1323
1324             /**************************
1325              * CALCULATE INTERACTIONS *
1326              **************************/
1327
1328             /* REACTION-FIELD ELECTROSTATICS */
1329             felec            = _mm_mul_pd(qq22,_mm_msub_pd(rinv22,rinvsq22,krf2));
1330
1331             fscal            = felec;
1332
1333             /* Update vectorial force */
1334             fix2             = _mm_macc_pd(dx22,fscal,fix2);
1335             fiy2             = _mm_macc_pd(dy22,fscal,fiy2);
1336             fiz2             = _mm_macc_pd(dz22,fscal,fiz2);
1337
1338             fjx2             = _mm_macc_pd(dx22,fscal,fjx2);
1339             fjy2             = _mm_macc_pd(dy22,fscal,fjy2);
1340             fjz2             = _mm_macc_pd(dz22,fscal,fjz2);
1341
1342             gmx_mm_decrement_3rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1343
1344             /* Inner loop uses 297 flops */
1345         }
1346
1347         if(jidx<j_index_end)
1348         {
1349
1350             jnrA             = jjnr[jidx];
1351             j_coord_offsetA  = DIM*jnrA;
1352
1353             /* load j atom coordinates */
1354             gmx_mm_load_3rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
1355                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1356
1357             /* Calculate displacement vector */
1358             dx00             = _mm_sub_pd(ix0,jx0);
1359             dy00             = _mm_sub_pd(iy0,jy0);
1360             dz00             = _mm_sub_pd(iz0,jz0);
1361             dx01             = _mm_sub_pd(ix0,jx1);
1362             dy01             = _mm_sub_pd(iy0,jy1);
1363             dz01             = _mm_sub_pd(iz0,jz1);
1364             dx02             = _mm_sub_pd(ix0,jx2);
1365             dy02             = _mm_sub_pd(iy0,jy2);
1366             dz02             = _mm_sub_pd(iz0,jz2);
1367             dx10             = _mm_sub_pd(ix1,jx0);
1368             dy10             = _mm_sub_pd(iy1,jy0);
1369             dz10             = _mm_sub_pd(iz1,jz0);
1370             dx11             = _mm_sub_pd(ix1,jx1);
1371             dy11             = _mm_sub_pd(iy1,jy1);
1372             dz11             = _mm_sub_pd(iz1,jz1);
1373             dx12             = _mm_sub_pd(ix1,jx2);
1374             dy12             = _mm_sub_pd(iy1,jy2);
1375             dz12             = _mm_sub_pd(iz1,jz2);
1376             dx20             = _mm_sub_pd(ix2,jx0);
1377             dy20             = _mm_sub_pd(iy2,jy0);
1378             dz20             = _mm_sub_pd(iz2,jz0);
1379             dx21             = _mm_sub_pd(ix2,jx1);
1380             dy21             = _mm_sub_pd(iy2,jy1);
1381             dz21             = _mm_sub_pd(iz2,jz1);
1382             dx22             = _mm_sub_pd(ix2,jx2);
1383             dy22             = _mm_sub_pd(iy2,jy2);
1384             dz22             = _mm_sub_pd(iz2,jz2);
1385
1386             /* Calculate squared distance and things based on it */
1387             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
1388             rsq01            = gmx_mm_calc_rsq_pd(dx01,dy01,dz01);
1389             rsq02            = gmx_mm_calc_rsq_pd(dx02,dy02,dz02);
1390             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
1391             rsq11            = gmx_mm_calc_rsq_pd(dx11,dy11,dz11);
1392             rsq12            = gmx_mm_calc_rsq_pd(dx12,dy12,dz12);
1393             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
1394             rsq21            = gmx_mm_calc_rsq_pd(dx21,dy21,dz21);
1395             rsq22            = gmx_mm_calc_rsq_pd(dx22,dy22,dz22);
1396
1397             rinv00           = gmx_mm_invsqrt_pd(rsq00);
1398             rinv01           = gmx_mm_invsqrt_pd(rsq01);
1399             rinv02           = gmx_mm_invsqrt_pd(rsq02);
1400             rinv10           = gmx_mm_invsqrt_pd(rsq10);
1401             rinv11           = gmx_mm_invsqrt_pd(rsq11);
1402             rinv12           = gmx_mm_invsqrt_pd(rsq12);
1403             rinv20           = gmx_mm_invsqrt_pd(rsq20);
1404             rinv21           = gmx_mm_invsqrt_pd(rsq21);
1405             rinv22           = gmx_mm_invsqrt_pd(rsq22);
1406
1407             rinvsq00         = _mm_mul_pd(rinv00,rinv00);
1408             rinvsq01         = _mm_mul_pd(rinv01,rinv01);
1409             rinvsq02         = _mm_mul_pd(rinv02,rinv02);
1410             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
1411             rinvsq11         = _mm_mul_pd(rinv11,rinv11);
1412             rinvsq12         = _mm_mul_pd(rinv12,rinv12);
1413             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
1414             rinvsq21         = _mm_mul_pd(rinv21,rinv21);
1415             rinvsq22         = _mm_mul_pd(rinv22,rinv22);
1416
1417             fjx0             = _mm_setzero_pd();
1418             fjy0             = _mm_setzero_pd();
1419             fjz0             = _mm_setzero_pd();
1420             fjx1             = _mm_setzero_pd();
1421             fjy1             = _mm_setzero_pd();
1422             fjz1             = _mm_setzero_pd();
1423             fjx2             = _mm_setzero_pd();
1424             fjy2             = _mm_setzero_pd();
1425             fjz2             = _mm_setzero_pd();
1426
1427             /**************************
1428              * CALCULATE INTERACTIONS *
1429              **************************/
1430
1431             r00              = _mm_mul_pd(rsq00,rinv00);
1432
1433             /* Calculate table index by multiplying r with table scale and truncate to integer */
1434             rt               = _mm_mul_pd(r00,vftabscale);
1435             vfitab           = _mm_cvttpd_epi32(rt);
1436 #ifdef __XOP__
1437             vfeps            = _mm_frcz_pd(rt);
1438 #else
1439             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1440 #endif
1441             twovfeps         = _mm_add_pd(vfeps,vfeps);
1442             vfitab           = _mm_slli_epi32(vfitab,3);
1443
1444             /* REACTION-FIELD ELECTROSTATICS */
1445             felec            = _mm_mul_pd(qq00,_mm_msub_pd(rinv00,rinvsq00,krf2));
1446
1447             /* CUBIC SPLINE TABLE DISPERSION */
1448             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1449             F                = _mm_setzero_pd();
1450             GMX_MM_TRANSPOSE2_PD(Y,F);
1451             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1452             H                = _mm_setzero_pd();
1453             GMX_MM_TRANSPOSE2_PD(G,H);
1454             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
1455             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
1456             fvdw6            = _mm_mul_pd(c6_00,FF);
1457
1458             /* CUBIC SPLINE TABLE REPULSION */
1459             vfitab           = _mm_add_epi32(vfitab,ifour);
1460             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1461             F                = _mm_setzero_pd();
1462             GMX_MM_TRANSPOSE2_PD(Y,F);
1463             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1464             H                = _mm_setzero_pd();
1465             GMX_MM_TRANSPOSE2_PD(G,H);
1466             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
1467             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
1468             fvdw12           = _mm_mul_pd(c12_00,FF);
1469             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
1470
1471             fscal            = _mm_add_pd(felec,fvdw);
1472
1473             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1474
1475             /* Update vectorial force */
1476             fix0             = _mm_macc_pd(dx00,fscal,fix0);
1477             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
1478             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
1479
1480             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
1481             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
1482             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
1483
1484             /**************************
1485              * CALCULATE INTERACTIONS *
1486              **************************/
1487
1488             /* REACTION-FIELD ELECTROSTATICS */
1489             felec            = _mm_mul_pd(qq01,_mm_msub_pd(rinv01,rinvsq01,krf2));
1490
1491             fscal            = felec;
1492
1493             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1494
1495             /* Update vectorial force */
1496             fix0             = _mm_macc_pd(dx01,fscal,fix0);
1497             fiy0             = _mm_macc_pd(dy01,fscal,fiy0);
1498             fiz0             = _mm_macc_pd(dz01,fscal,fiz0);
1499
1500             fjx1             = _mm_macc_pd(dx01,fscal,fjx1);
1501             fjy1             = _mm_macc_pd(dy01,fscal,fjy1);
1502             fjz1             = _mm_macc_pd(dz01,fscal,fjz1);
1503
1504             /**************************
1505              * CALCULATE INTERACTIONS *
1506              **************************/
1507
1508             /* REACTION-FIELD ELECTROSTATICS */
1509             felec            = _mm_mul_pd(qq02,_mm_msub_pd(rinv02,rinvsq02,krf2));
1510
1511             fscal            = felec;
1512
1513             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1514
1515             /* Update vectorial force */
1516             fix0             = _mm_macc_pd(dx02,fscal,fix0);
1517             fiy0             = _mm_macc_pd(dy02,fscal,fiy0);
1518             fiz0             = _mm_macc_pd(dz02,fscal,fiz0);
1519
1520             fjx2             = _mm_macc_pd(dx02,fscal,fjx2);
1521             fjy2             = _mm_macc_pd(dy02,fscal,fjy2);
1522             fjz2             = _mm_macc_pd(dz02,fscal,fjz2);
1523
1524             /**************************
1525              * CALCULATE INTERACTIONS *
1526              **************************/
1527
1528             /* REACTION-FIELD ELECTROSTATICS */
1529             felec            = _mm_mul_pd(qq10,_mm_msub_pd(rinv10,rinvsq10,krf2));
1530
1531             fscal            = felec;
1532
1533             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1534
1535             /* Update vectorial force */
1536             fix1             = _mm_macc_pd(dx10,fscal,fix1);
1537             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
1538             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
1539
1540             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
1541             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
1542             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
1543
1544             /**************************
1545              * CALCULATE INTERACTIONS *
1546              **************************/
1547
1548             /* REACTION-FIELD ELECTROSTATICS */
1549             felec            = _mm_mul_pd(qq11,_mm_msub_pd(rinv11,rinvsq11,krf2));
1550
1551             fscal            = felec;
1552
1553             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1554
1555             /* Update vectorial force */
1556             fix1             = _mm_macc_pd(dx11,fscal,fix1);
1557             fiy1             = _mm_macc_pd(dy11,fscal,fiy1);
1558             fiz1             = _mm_macc_pd(dz11,fscal,fiz1);
1559
1560             fjx1             = _mm_macc_pd(dx11,fscal,fjx1);
1561             fjy1             = _mm_macc_pd(dy11,fscal,fjy1);
1562             fjz1             = _mm_macc_pd(dz11,fscal,fjz1);
1563
1564             /**************************
1565              * CALCULATE INTERACTIONS *
1566              **************************/
1567
1568             /* REACTION-FIELD ELECTROSTATICS */
1569             felec            = _mm_mul_pd(qq12,_mm_msub_pd(rinv12,rinvsq12,krf2));
1570
1571             fscal            = felec;
1572
1573             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1574
1575             /* Update vectorial force */
1576             fix1             = _mm_macc_pd(dx12,fscal,fix1);
1577             fiy1             = _mm_macc_pd(dy12,fscal,fiy1);
1578             fiz1             = _mm_macc_pd(dz12,fscal,fiz1);
1579
1580             fjx2             = _mm_macc_pd(dx12,fscal,fjx2);
1581             fjy2             = _mm_macc_pd(dy12,fscal,fjy2);
1582             fjz2             = _mm_macc_pd(dz12,fscal,fjz2);
1583
1584             /**************************
1585              * CALCULATE INTERACTIONS *
1586              **************************/
1587
1588             /* REACTION-FIELD ELECTROSTATICS */
1589             felec            = _mm_mul_pd(qq20,_mm_msub_pd(rinv20,rinvsq20,krf2));
1590
1591             fscal            = felec;
1592
1593             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1594
1595             /* Update vectorial force */
1596             fix2             = _mm_macc_pd(dx20,fscal,fix2);
1597             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
1598             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
1599
1600             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
1601             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
1602             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
1603
1604             /**************************
1605              * CALCULATE INTERACTIONS *
1606              **************************/
1607
1608             /* REACTION-FIELD ELECTROSTATICS */
1609             felec            = _mm_mul_pd(qq21,_mm_msub_pd(rinv21,rinvsq21,krf2));
1610
1611             fscal            = felec;
1612
1613             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1614
1615             /* Update vectorial force */
1616             fix2             = _mm_macc_pd(dx21,fscal,fix2);
1617             fiy2             = _mm_macc_pd(dy21,fscal,fiy2);
1618             fiz2             = _mm_macc_pd(dz21,fscal,fiz2);
1619
1620             fjx1             = _mm_macc_pd(dx21,fscal,fjx1);
1621             fjy1             = _mm_macc_pd(dy21,fscal,fjy1);
1622             fjz1             = _mm_macc_pd(dz21,fscal,fjz1);
1623
1624             /**************************
1625              * CALCULATE INTERACTIONS *
1626              **************************/
1627
1628             /* REACTION-FIELD ELECTROSTATICS */
1629             felec            = _mm_mul_pd(qq22,_mm_msub_pd(rinv22,rinvsq22,krf2));
1630
1631             fscal            = felec;
1632
1633             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1634
1635             /* Update vectorial force */
1636             fix2             = _mm_macc_pd(dx22,fscal,fix2);
1637             fiy2             = _mm_macc_pd(dy22,fscal,fiy2);
1638             fiz2             = _mm_macc_pd(dz22,fscal,fiz2);
1639
1640             fjx2             = _mm_macc_pd(dx22,fscal,fjx2);
1641             fjy2             = _mm_macc_pd(dy22,fscal,fjy2);
1642             fjz2             = _mm_macc_pd(dz22,fscal,fjz2);
1643
1644             gmx_mm_decrement_3rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1645
1646             /* Inner loop uses 297 flops */
1647         }
1648
1649         /* End of innermost loop */
1650
1651         gmx_mm_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1652                                               f+i_coord_offset,fshift+i_shift_offset);
1653
1654         /* Increment number of inner iterations */
1655         inneriter                  += j_index_end - j_index_start;
1656
1657         /* Outer loop uses 18 flops */
1658     }
1659
1660     /* Increment number of outer iterations */
1661     outeriter        += nri;
1662
1663     /* Update outer/inner flops */
1664
1665     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*18 + inneriter*297);
1666 }