src/gromacs/gmxlib/nonbonded/nb_kernel_sse4_1_single/nb_kernel_ElecGB_VdwNone_GeomP1P1_sse4_1_single.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013, 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 sse4_1_single kernel generator.
  37  */
  38 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "types/simple.h"
  46 #include "vec.h"
  47 #include "nrnb.h"
  48
  49 #include "gromacs/simd/math_x86_sse4_1_single.h"
  50 #include "kernelutil_x86_sse4_1_single.h"
  51
  52 /*
  53  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwNone_GeomP1P1_VF_sse4_1_single
  54  * Electrostatics interaction: GeneralizedBorn
  55  * VdW interaction:            None
  56  * Geometry:                   Particle-Particle
  57  * Calculate force/pot:        PotentialAndForce
  58  */
  59 void
  60 nb_kernel_ElecGB_VdwNone_GeomP1P1_VF_sse4_1_single
  61                     (t_nblist                    * gmx_restrict       nlist,
  62                      rvec                        * gmx_restrict          xx,
  63                      rvec                        * gmx_restrict          ff,
  64                      t_forcerec                  * gmx_restrict          fr,
  65                      t_mdatoms                   * gmx_restrict     mdatoms,
  66                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  67                      t_nrnb                      * gmx_restrict        nrnb)
  68 {
  69     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  70      * just 0 for non-waters.
  71      * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
  72      * jnr indices corresponding to data put in the four positions in the SIMD register.
  73      */
  74     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  75     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  76     int              jnrA,jnrB,jnrC,jnrD;
  77     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  78     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  79     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  80     real             rcutoff_scalar;
  81     real             *shiftvec,*fshift,*x,*f;
  82     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  83     real             scratch[4*DIM];
  84     __m128           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  85     int              vdwioffset0;
  86     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  87     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  88     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  89     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  90     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
  91     real             *charge;
  92     __m128i          gbitab;
  93     __m128           vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,gbeps,dvdatmp;
  94     __m128           minushalf = _mm_set1_ps(-0.5);
  95     real             *invsqrta,*dvda,*gbtab;
  96     __m128i          vfitab;
  97     __m128i          ifour       = _mm_set1_epi32(4);
  98     __m128           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
  99     real             *vftab;
 100     __m128           dummy_mask,cutoff_mask;
 101     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 102     __m128           one     = _mm_set1_ps(1.0);
 103     __m128           two     = _mm_set1_ps(2.0);
 104     x                = xx[0];
 105     f                = ff[0];
 106
 107     nri              = nlist->nri;
 108     iinr             = nlist->iinr;
 109     jindex           = nlist->jindex;
 110     jjnr             = nlist->jjnr;
 111     shiftidx         = nlist->shift;
 112     gid              = nlist->gid;
 113     shiftvec         = fr->shift_vec[0];
 114     fshift           = fr->fshift[0];
 115     facel            = _mm_set1_ps(fr->epsfac);
 116     charge           = mdatoms->chargeA;
 117
 118     invsqrta         = fr->invsqrta;
 119     dvda             = fr->dvda;
 120     gbtabscale       = _mm_set1_ps(fr->gbtab.scale);
 121     gbtab            = fr->gbtab.data;
 122     gbinvepsdiff     = _mm_set1_ps((1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent));
 123
 124     /* Avoid stupid compiler warnings */
 125     jnrA = jnrB = jnrC = jnrD = 0;
 126     j_coord_offsetA = 0;
 127     j_coord_offsetB = 0;
 128     j_coord_offsetC = 0;
 129     j_coord_offsetD = 0;
 130
 131     outeriter        = 0;
 132     inneriter        = 0;
 133
 134     for(iidx=0;iidx<4*DIM;iidx++)
 135     {
 136         scratch[iidx] = 0.0;
 137     }
 138
 139     /* Start outer loop over neighborlists */
 140     for(iidx=0; iidx<nri; iidx++)
 141     {
 142         /* Load shift vector for this list */
 143         i_shift_offset   = DIM*shiftidx[iidx];
 144
 145         /* Load limits for loop over neighbors */
 146         j_index_start    = jindex[iidx];
 147         j_index_end      = jindex[iidx+1];
 148
 149         /* Get outer coordinate index */
 150         inr              = iinr[iidx];
 151         i_coord_offset   = DIM*inr;
 152
 153         /* Load i particle coords and add shift vector */
 154         gmx_mm_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 155
 156         fix0             = _mm_setzero_ps();
 157         fiy0             = _mm_setzero_ps();
 158         fiz0             = _mm_setzero_ps();
 159
 160         /* Load parameters for i particles */
 161         iq0              = _mm_mul_ps(facel,_mm_load1_ps(charge+inr+0));
 162         isai0            = _mm_load1_ps(invsqrta+inr+0);
 163
 164         /* Reset potential sums */
 165         velecsum         = _mm_setzero_ps();
 166         vgbsum           = _mm_setzero_ps();
 167         dvdasum          = _mm_setzero_ps();
 168
 169         /* Start inner kernel loop */
 170         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 171         {
 172
 173             /* Get j neighbor index, and coordinate index */
 174             jnrA             = jjnr[jidx];
 175             jnrB             = jjnr[jidx+1];
 176             jnrC             = jjnr[jidx+2];
 177             jnrD             = jjnr[jidx+3];
 178             j_coord_offsetA  = DIM*jnrA;
 179             j_coord_offsetB  = DIM*jnrB;
 180             j_coord_offsetC  = DIM*jnrC;
 181             j_coord_offsetD  = DIM*jnrD;
 182
 183             /* load j atom coordinates */
 184             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 185                                               x+j_coord_offsetC,x+j_coord_offsetD,
 186                                               &jx0,&jy0,&jz0);
 187
 188             /* Calculate displacement vector */
 189             dx00             = _mm_sub_ps(ix0,jx0);
 190             dy00             = _mm_sub_ps(iy0,jy0);
 191             dz00             = _mm_sub_ps(iz0,jz0);
 192
 193             /* Calculate squared distance and things based on it */
 194             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 195
 196             rinv00           = gmx_mm_invsqrt_ps(rsq00);
 197
 198             /* Load parameters for j particles */
 199             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 200                                                               charge+jnrC+0,charge+jnrD+0);
 201             isaj0            = gmx_mm_load_4real_swizzle_ps(invsqrta+jnrA+0,invsqrta+jnrB+0,
 202                                                               invsqrta+jnrC+0,invsqrta+jnrD+0);
 203
 204             /**************************
 205              * CALCULATE INTERACTIONS *
 206              **************************/
 207
 208             r00              = _mm_mul_ps(rsq00,rinv00);
 209
 210             /* Compute parameters for interactions between i and j atoms */
 211             qq00             = _mm_mul_ps(iq0,jq0);
 212
 213             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 214             isaprod          = _mm_mul_ps(isai0,isaj0);
 215             gbqqfactor       = _mm_xor_ps(signbit,_mm_mul_ps(qq00,_mm_mul_ps(isaprod,gbinvepsdiff)));
 216             gbscale          = _mm_mul_ps(isaprod,gbtabscale);
 217
 218             /* Calculate generalized born table index - this is a separate table from the normal one,
 219              * but we use the same procedure by multiplying r with scale and truncating to integer.
 220              */
 221             rt               = _mm_mul_ps(r00,gbscale);
 222             gbitab           = _mm_cvttps_epi32(rt);
 223             gbeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 224             gbitab           = _mm_slli_epi32(gbitab,2);
 225             Y                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,0) );
 226             F                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,1) );
 227             G                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,2) );
 228             H                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,3) );
 229             _MM_TRANSPOSE4_PS(Y,F,G,H);
 230             Heps             = _mm_mul_ps(gbeps,H);
 231             Fp               = _mm_add_ps(F,_mm_mul_ps(gbeps,_mm_add_ps(G,Heps)));
 232             VV               = _mm_add_ps(Y,_mm_mul_ps(gbeps,Fp));
 233             vgb              = _mm_mul_ps(gbqqfactor,VV);
 234
 235             FF               = _mm_add_ps(Fp,_mm_mul_ps(gbeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 236             fgb              = _mm_mul_ps(gbqqfactor,_mm_mul_ps(FF,gbscale));
 237             dvdatmp          = _mm_mul_ps(minushalf,_mm_add_ps(vgb,_mm_mul_ps(fgb,r00)));
 238             dvdasum          = _mm_add_ps(dvdasum,dvdatmp);
 239             fjptrA           = dvda+jnrA;
 240             fjptrB           = dvda+jnrB;
 241             fjptrC           = dvda+jnrC;
 242             fjptrD           = dvda+jnrD;
 243             gmx_mm_increment_4real_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,_mm_mul_ps(dvdatmp,_mm_mul_ps(isaj0,isaj0)));
 244             velec            = _mm_mul_ps(qq00,rinv00);
 245             felec            = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(velec,rinv00),fgb),rinv00);
 246
 247             /* Update potential sum for this i atom from the interaction with this j atom. */
 248             velecsum         = _mm_add_ps(velecsum,velec);
 249             vgbsum           = _mm_add_ps(vgbsum,vgb);
 250
 251             fscal            = felec;
 252
 253             /* Calculate temporary vectorial force */
 254             tx               = _mm_mul_ps(fscal,dx00);
 255             ty               = _mm_mul_ps(fscal,dy00);
 256             tz               = _mm_mul_ps(fscal,dz00);
 257
 258             /* Update vectorial force */
 259             fix0             = _mm_add_ps(fix0,tx);
 260             fiy0             = _mm_add_ps(fiy0,ty);
 261             fiz0             = _mm_add_ps(fiz0,tz);
 262
 263             fjptrA             = f+j_coord_offsetA;
 264             fjptrB             = f+j_coord_offsetB;
 265             fjptrC             = f+j_coord_offsetC;
 266             fjptrD             = f+j_coord_offsetD;
 267             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 268
 269             /* Inner loop uses 58 flops */
 270         }
 271
 272         if(jidx<j_index_end)
 273         {
 274
 275             /* Get j neighbor index, and coordinate index */
 276             jnrlistA         = jjnr[jidx];
 277             jnrlistB         = jjnr[jidx+1];
 278             jnrlistC         = jjnr[jidx+2];
 279             jnrlistD         = jjnr[jidx+3];
 280             /* Sign of each element will be negative for non-real atoms.
 281              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 282              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 283              */
 284             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 285             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 286             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 287             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 288             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 289             j_coord_offsetA  = DIM*jnrA;
 290             j_coord_offsetB  = DIM*jnrB;
 291             j_coord_offsetC  = DIM*jnrC;
 292             j_coord_offsetD  = DIM*jnrD;
 293
 294             /* load j atom coordinates */
 295             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 296                                               x+j_coord_offsetC,x+j_coord_offsetD,
 297                                               &jx0,&jy0,&jz0);
 298
 299             /* Calculate displacement vector */
 300             dx00             = _mm_sub_ps(ix0,jx0);
 301             dy00             = _mm_sub_ps(iy0,jy0);
 302             dz00             = _mm_sub_ps(iz0,jz0);
 303
 304             /* Calculate squared distance and things based on it */
 305             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 306
 307             rinv00           = gmx_mm_invsqrt_ps(rsq00);
 308
 309             /* Load parameters for j particles */
 310             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 311                                                               charge+jnrC+0,charge+jnrD+0);
 312             isaj0            = gmx_mm_load_4real_swizzle_ps(invsqrta+jnrA+0,invsqrta+jnrB+0,
 313                                                               invsqrta+jnrC+0,invsqrta+jnrD+0);
 314
 315             /**************************
 316              * CALCULATE INTERACTIONS *
 317              **************************/
 318
 319             r00              = _mm_mul_ps(rsq00,rinv00);
 320             r00              = _mm_andnot_ps(dummy_mask,r00);
 321
 322             /* Compute parameters for interactions between i and j atoms */
 323             qq00             = _mm_mul_ps(iq0,jq0);
 324
 325             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 326             isaprod          = _mm_mul_ps(isai0,isaj0);
 327             gbqqfactor       = _mm_xor_ps(signbit,_mm_mul_ps(qq00,_mm_mul_ps(isaprod,gbinvepsdiff)));
 328             gbscale          = _mm_mul_ps(isaprod,gbtabscale);
 329
 330             /* Calculate generalized born table index - this is a separate table from the normal one,
 331              * but we use the same procedure by multiplying r with scale and truncating to integer.
 332              */
 333             rt               = _mm_mul_ps(r00,gbscale);
 334             gbitab           = _mm_cvttps_epi32(rt);
 335             gbeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 336             gbitab           = _mm_slli_epi32(gbitab,2);
 337             Y                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,0) );
 338             F                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,1) );
 339             G                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,2) );
 340             H                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,3) );
 341             _MM_TRANSPOSE4_PS(Y,F,G,H);
 342             Heps             = _mm_mul_ps(gbeps,H);
 343             Fp               = _mm_add_ps(F,_mm_mul_ps(gbeps,_mm_add_ps(G,Heps)));
 344             VV               = _mm_add_ps(Y,_mm_mul_ps(gbeps,Fp));
 345             vgb              = _mm_mul_ps(gbqqfactor,VV);
 346
 347             FF               = _mm_add_ps(Fp,_mm_mul_ps(gbeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 348             fgb              = _mm_mul_ps(gbqqfactor,_mm_mul_ps(FF,gbscale));
 349             dvdatmp          = _mm_mul_ps(minushalf,_mm_add_ps(vgb,_mm_mul_ps(fgb,r00)));
 350             dvdatmp          = _mm_andnot_ps(dummy_mask,dvdatmp);
 351             dvdasum          = _mm_add_ps(dvdasum,dvdatmp);
 352             /* The pointers to scratch make sure that this code with compilers that take gmx_restrict seriously (e.g. icc 13) really can't screw things up. */
 353             fjptrA             = (jnrlistA>=0) ? dvda+jnrA : scratch;
 354             fjptrB             = (jnrlistB>=0) ? dvda+jnrB : scratch;
 355             fjptrC             = (jnrlistC>=0) ? dvda+jnrC : scratch;
 356             fjptrD             = (jnrlistD>=0) ? dvda+jnrD : scratch;
 357             gmx_mm_increment_4real_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,_mm_mul_ps(dvdatmp,_mm_mul_ps(isaj0,isaj0)));
 358             velec            = _mm_mul_ps(qq00,rinv00);
 359             felec            = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(velec,rinv00),fgb),rinv00);
 360
 361             /* Update potential sum for this i atom from the interaction with this j atom. */
 362             velec            = _mm_andnot_ps(dummy_mask,velec);
 363             velecsum         = _mm_add_ps(velecsum,velec);
 364             vgb              = _mm_andnot_ps(dummy_mask,vgb);
 365             vgbsum           = _mm_add_ps(vgbsum,vgb);
 366
 367             fscal            = felec;
 368
 369             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 370
 371             /* Calculate temporary vectorial force */
 372             tx               = _mm_mul_ps(fscal,dx00);
 373             ty               = _mm_mul_ps(fscal,dy00);
 374             tz               = _mm_mul_ps(fscal,dz00);
 375
 376             /* Update vectorial force */
 377             fix0             = _mm_add_ps(fix0,tx);
 378             fiy0             = _mm_add_ps(fiy0,ty);
 379             fiz0             = _mm_add_ps(fiz0,tz);
 380
 381             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 382             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 383             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 384             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 385             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 386
 387             /* Inner loop uses 59 flops */
 388         }
 389
 390         /* End of innermost loop */
 391
 392         gmx_mm_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
 393                                               f+i_coord_offset,fshift+i_shift_offset);
 394
 395         ggid                        = gid[iidx];
 396         /* Update potential energies */
 397         gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 398         gmx_mm_update_1pot_ps(vgbsum,kernel_data->energygrp_polarization+ggid);
 399         dvdasum = _mm_mul_ps(dvdasum, _mm_mul_ps(isai0,isai0));
 400         gmx_mm_update_1pot_ps(dvdasum,dvda+inr);
 401
 402         /* Increment number of inner iterations */
 403         inneriter                  += j_index_end - j_index_start;
 404
 405         /* Outer loop uses 9 flops */
 406     }
 407
 408     /* Increment number of outer iterations */
 409     outeriter        += nri;
 410
 411     /* Update outer/inner flops */
 412
 413     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VF,outeriter*9 + inneriter*59);
 414 }
 415 /*
 416  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwNone_GeomP1P1_F_sse4_1_single
 417  * Electrostatics interaction: GeneralizedBorn
 418  * VdW interaction:            None
 419  * Geometry:                   Particle-Particle
 420  * Calculate force/pot:        Force
 421  */
 422 void
 423 nb_kernel_ElecGB_VdwNone_GeomP1P1_F_sse4_1_single
 424                     (t_nblist                    * gmx_restrict       nlist,
 425                      rvec                        * gmx_restrict          xx,
 426                      rvec                        * gmx_restrict          ff,
 427                      t_forcerec                  * gmx_restrict          fr,
 428                      t_mdatoms                   * gmx_restrict     mdatoms,
 429                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 430                      t_nrnb                      * gmx_restrict        nrnb)
 431 {
 432     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 433      * just 0 for non-waters.
 434      * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
 435      * jnr indices corresponding to data put in the four positions in the SIMD register.
 436      */
 437     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 438     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 439     int              jnrA,jnrB,jnrC,jnrD;
 440     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 441     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 442     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 443     real             rcutoff_scalar;
 444     real             *shiftvec,*fshift,*x,*f;
 445     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 446     real             scratch[4*DIM];
 447     __m128           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 448     int              vdwioffset0;
 449     __m128           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 450     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 451     __m128           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 452     __m128           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 453     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
 454     real             *charge;
 455     __m128i          gbitab;
 456     __m128           vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,gbeps,dvdatmp;
 457     __m128           minushalf = _mm_set1_ps(-0.5);
 458     real             *invsqrta,*dvda,*gbtab;
 459     __m128i          vfitab;
 460     __m128i          ifour       = _mm_set1_epi32(4);
 461     __m128           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 462     real             *vftab;
 463     __m128           dummy_mask,cutoff_mask;
 464     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 465     __m128           one     = _mm_set1_ps(1.0);
 466     __m128           two     = _mm_set1_ps(2.0);
 467     x                = xx[0];
 468     f                = ff[0];
 469
 470     nri              = nlist->nri;
 471     iinr             = nlist->iinr;
 472     jindex           = nlist->jindex;
 473     jjnr             = nlist->jjnr;
 474     shiftidx         = nlist->shift;
 475     gid              = nlist->gid;
 476     shiftvec         = fr->shift_vec[0];
 477     fshift           = fr->fshift[0];
 478     facel            = _mm_set1_ps(fr->epsfac);
 479     charge           = mdatoms->chargeA;
 480
 481     invsqrta         = fr->invsqrta;
 482     dvda             = fr->dvda;
 483     gbtabscale       = _mm_set1_ps(fr->gbtab.scale);
 484     gbtab            = fr->gbtab.data;
 485     gbinvepsdiff     = _mm_set1_ps((1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent));
 486
 487     /* Avoid stupid compiler warnings */
 488     jnrA = jnrB = jnrC = jnrD = 0;
 489     j_coord_offsetA = 0;
 490     j_coord_offsetB = 0;
 491     j_coord_offsetC = 0;
 492     j_coord_offsetD = 0;
 493
 494     outeriter        = 0;
 495     inneriter        = 0;
 496
 497     for(iidx=0;iidx<4*DIM;iidx++)
 498     {
 499         scratch[iidx] = 0.0;
 500     }
 501
 502     /* Start outer loop over neighborlists */
 503     for(iidx=0; iidx<nri; iidx++)
 504     {
 505         /* Load shift vector for this list */
 506         i_shift_offset   = DIM*shiftidx[iidx];
 507
 508         /* Load limits for loop over neighbors */
 509         j_index_start    = jindex[iidx];
 510         j_index_end      = jindex[iidx+1];
 511
 512         /* Get outer coordinate index */
 513         inr              = iinr[iidx];
 514         i_coord_offset   = DIM*inr;
 515
 516         /* Load i particle coords and add shift vector */
 517         gmx_mm_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 518
 519         fix0             = _mm_setzero_ps();
 520         fiy0             = _mm_setzero_ps();
 521         fiz0             = _mm_setzero_ps();
 522
 523         /* Load parameters for i particles */
 524         iq0              = _mm_mul_ps(facel,_mm_load1_ps(charge+inr+0));
 525         isai0            = _mm_load1_ps(invsqrta+inr+0);
 526
 527         dvdasum          = _mm_setzero_ps();
 528
 529         /* Start inner kernel loop */
 530         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 531         {
 532
 533             /* Get j neighbor index, and coordinate index */
 534             jnrA             = jjnr[jidx];
 535             jnrB             = jjnr[jidx+1];
 536             jnrC             = jjnr[jidx+2];
 537             jnrD             = jjnr[jidx+3];
 538             j_coord_offsetA  = DIM*jnrA;
 539             j_coord_offsetB  = DIM*jnrB;
 540             j_coord_offsetC  = DIM*jnrC;
 541             j_coord_offsetD  = DIM*jnrD;
 542
 543             /* load j atom coordinates */
 544             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 545                                               x+j_coord_offsetC,x+j_coord_offsetD,
 546                                               &jx0,&jy0,&jz0);
 547
 548             /* Calculate displacement vector */
 549             dx00             = _mm_sub_ps(ix0,jx0);
 550             dy00             = _mm_sub_ps(iy0,jy0);
 551             dz00             = _mm_sub_ps(iz0,jz0);
 552
 553             /* Calculate squared distance and things based on it */
 554             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 555
 556             rinv00           = gmx_mm_invsqrt_ps(rsq00);
 557
 558             /* Load parameters for j particles */
 559             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 560                                                               charge+jnrC+0,charge+jnrD+0);
 561             isaj0            = gmx_mm_load_4real_swizzle_ps(invsqrta+jnrA+0,invsqrta+jnrB+0,
 562                                                               invsqrta+jnrC+0,invsqrta+jnrD+0);
 563
 564             /**************************
 565              * CALCULATE INTERACTIONS *
 566              **************************/
 567
 568             r00              = _mm_mul_ps(rsq00,rinv00);
 569
 570             /* Compute parameters for interactions between i and j atoms */
 571             qq00             = _mm_mul_ps(iq0,jq0);
 572
 573             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 574             isaprod          = _mm_mul_ps(isai0,isaj0);
 575             gbqqfactor       = _mm_xor_ps(signbit,_mm_mul_ps(qq00,_mm_mul_ps(isaprod,gbinvepsdiff)));
 576             gbscale          = _mm_mul_ps(isaprod,gbtabscale);
 577
 578             /* Calculate generalized born table index - this is a separate table from the normal one,
 579              * but we use the same procedure by multiplying r with scale and truncating to integer.
 580              */
 581             rt               = _mm_mul_ps(r00,gbscale);
 582             gbitab           = _mm_cvttps_epi32(rt);
 583             gbeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 584             gbitab           = _mm_slli_epi32(gbitab,2);
 585             Y                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,0) );
 586             F                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,1) );
 587             G                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,2) );
 588             H                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,3) );
 589             _MM_TRANSPOSE4_PS(Y,F,G,H);
 590             Heps             = _mm_mul_ps(gbeps,H);
 591             Fp               = _mm_add_ps(F,_mm_mul_ps(gbeps,_mm_add_ps(G,Heps)));
 592             VV               = _mm_add_ps(Y,_mm_mul_ps(gbeps,Fp));
 593             vgb              = _mm_mul_ps(gbqqfactor,VV);
 594
 595             FF               = _mm_add_ps(Fp,_mm_mul_ps(gbeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 596             fgb              = _mm_mul_ps(gbqqfactor,_mm_mul_ps(FF,gbscale));
 597             dvdatmp          = _mm_mul_ps(minushalf,_mm_add_ps(vgb,_mm_mul_ps(fgb,r00)));
 598             dvdasum          = _mm_add_ps(dvdasum,dvdatmp);
 599             fjptrA           = dvda+jnrA;
 600             fjptrB           = dvda+jnrB;
 601             fjptrC           = dvda+jnrC;
 602             fjptrD           = dvda+jnrD;
 603             gmx_mm_increment_4real_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,_mm_mul_ps(dvdatmp,_mm_mul_ps(isaj0,isaj0)));
 604             velec            = _mm_mul_ps(qq00,rinv00);
 605             felec            = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(velec,rinv00),fgb),rinv00);
 606
 607             fscal            = felec;
 608
 609             /* Calculate temporary vectorial force */
 610             tx               = _mm_mul_ps(fscal,dx00);
 611             ty               = _mm_mul_ps(fscal,dy00);
 612             tz               = _mm_mul_ps(fscal,dz00);
 613
 614             /* Update vectorial force */
 615             fix0             = _mm_add_ps(fix0,tx);
 616             fiy0             = _mm_add_ps(fiy0,ty);
 617             fiz0             = _mm_add_ps(fiz0,tz);
 618
 619             fjptrA             = f+j_coord_offsetA;
 620             fjptrB             = f+j_coord_offsetB;
 621             fjptrC             = f+j_coord_offsetC;
 622             fjptrD             = f+j_coord_offsetD;
 623             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 624
 625             /* Inner loop uses 56 flops */
 626         }
 627
 628         if(jidx<j_index_end)
 629         {
 630
 631             /* Get j neighbor index, and coordinate index */
 632             jnrlistA         = jjnr[jidx];
 633             jnrlistB         = jjnr[jidx+1];
 634             jnrlistC         = jjnr[jidx+2];
 635             jnrlistD         = jjnr[jidx+3];
 636             /* Sign of each element will be negative for non-real atoms.
 637              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 638              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 639              */
 640             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 641             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 642             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 643             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 644             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 645             j_coord_offsetA  = DIM*jnrA;
 646             j_coord_offsetB  = DIM*jnrB;
 647             j_coord_offsetC  = DIM*jnrC;
 648             j_coord_offsetD  = DIM*jnrD;
 649
 650             /* load j atom coordinates */
 651             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 652                                               x+j_coord_offsetC,x+j_coord_offsetD,
 653                                               &jx0,&jy0,&jz0);
 654
 655             /* Calculate displacement vector */
 656             dx00             = _mm_sub_ps(ix0,jx0);
 657             dy00             = _mm_sub_ps(iy0,jy0);
 658             dz00             = _mm_sub_ps(iz0,jz0);
 659
 660             /* Calculate squared distance and things based on it */
 661             rsq00            = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
 662
 663             rinv00           = gmx_mm_invsqrt_ps(rsq00);
 664
 665             /* Load parameters for j particles */
 666             jq0              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 667                                                               charge+jnrC+0,charge+jnrD+0);
 668             isaj0            = gmx_mm_load_4real_swizzle_ps(invsqrta+jnrA+0,invsqrta+jnrB+0,
 669                                                               invsqrta+jnrC+0,invsqrta+jnrD+0);
 670
 671             /**************************
 672              * CALCULATE INTERACTIONS *
 673              **************************/
 674
 675             r00              = _mm_mul_ps(rsq00,rinv00);
 676             r00              = _mm_andnot_ps(dummy_mask,r00);
 677
 678             /* Compute parameters for interactions between i and j atoms */
 679             qq00             = _mm_mul_ps(iq0,jq0);
 680
 681             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 682             isaprod          = _mm_mul_ps(isai0,isaj0);
 683             gbqqfactor       = _mm_xor_ps(signbit,_mm_mul_ps(qq00,_mm_mul_ps(isaprod,gbinvepsdiff)));
 684             gbscale          = _mm_mul_ps(isaprod,gbtabscale);
 685
 686             /* Calculate generalized born table index - this is a separate table from the normal one,
 687              * but we use the same procedure by multiplying r with scale and truncating to integer.
 688              */
 689             rt               = _mm_mul_ps(r00,gbscale);
 690             gbitab           = _mm_cvttps_epi32(rt);
 691             gbeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 692             gbitab           = _mm_slli_epi32(gbitab,2);
 693             Y                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,0) );
 694             F                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,1) );
 695             G                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,2) );
 696             H                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,3) );
 697             _MM_TRANSPOSE4_PS(Y,F,G,H);
 698             Heps             = _mm_mul_ps(gbeps,H);
 699             Fp               = _mm_add_ps(F,_mm_mul_ps(gbeps,_mm_add_ps(G,Heps)));
 700             VV               = _mm_add_ps(Y,_mm_mul_ps(gbeps,Fp));
 701             vgb              = _mm_mul_ps(gbqqfactor,VV);
 702
 703             FF               = _mm_add_ps(Fp,_mm_mul_ps(gbeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 704             fgb              = _mm_mul_ps(gbqqfactor,_mm_mul_ps(FF,gbscale));
 705             dvdatmp          = _mm_mul_ps(minushalf,_mm_add_ps(vgb,_mm_mul_ps(fgb,r00)));
 706             dvdatmp          = _mm_andnot_ps(dummy_mask,dvdatmp);
 707             dvdasum          = _mm_add_ps(dvdasum,dvdatmp);
 708             /* The pointers to scratch make sure that this code with compilers that take gmx_restrict seriously (e.g. icc 13) really can't screw things up. */
 709             fjptrA             = (jnrlistA>=0) ? dvda+jnrA : scratch;
 710             fjptrB             = (jnrlistB>=0) ? dvda+jnrB : scratch;
 711             fjptrC             = (jnrlistC>=0) ? dvda+jnrC : scratch;
 712             fjptrD             = (jnrlistD>=0) ? dvda+jnrD : scratch;
 713             gmx_mm_increment_4real_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,_mm_mul_ps(dvdatmp,_mm_mul_ps(isaj0,isaj0)));
 714             velec            = _mm_mul_ps(qq00,rinv00);
 715             felec            = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(velec,rinv00),fgb),rinv00);
 716
 717             fscal            = felec;
 718
 719             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 720
 721             /* Calculate temporary vectorial force */
 722             tx               = _mm_mul_ps(fscal,dx00);
 723             ty               = _mm_mul_ps(fscal,dy00);
 724             tz               = _mm_mul_ps(fscal,dz00);
 725
 726             /* Update vectorial force */
 727             fix0             = _mm_add_ps(fix0,tx);
 728             fiy0             = _mm_add_ps(fiy0,ty);
 729             fiz0             = _mm_add_ps(fiz0,tz);
 730
 731             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 732             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 733             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 734             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 735             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 736
 737             /* Inner loop uses 57 flops */
 738         }
 739
 740         /* End of innermost loop */
 741
 742         gmx_mm_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
 743                                               f+i_coord_offset,fshift+i_shift_offset);
 744
 745         dvdasum = _mm_mul_ps(dvdasum, _mm_mul_ps(isai0,isai0));
 746         gmx_mm_update_1pot_ps(dvdasum,dvda+inr);
 747
 748         /* Increment number of inner iterations */
 749         inneriter                  += j_index_end - j_index_start;
 750
 751         /* Outer loop uses 7 flops */
 752     }
 753
 754     /* Increment number of outer iterations */
 755     outeriter        += nri;
 756
 757     /* Update outer/inner flops */
 758
 759     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_F,outeriter*7 + inneriter*57);
 760 }