| File: | gromacs/gmxlib/nonbonded/nb_kernel_sse4_1_single/nb_kernel_ElecRFCut_VdwLJSw_GeomW3P1_sse4_1_single.c |
| Location: | line 156, column 5 |
| Description: | Value stored to 'j_coord_offsetA' is never read |
| 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 sse4_1_single kernel generator. |
| 37 | */ |
| 38 | #ifdef HAVE_CONFIG_H1 |
| 39 | #include <config.h> |
| 40 | #endif |
| 41 | |
| 42 | #include <math.h> |
| 43 | |
| 44 | #include "../nb_kernel.h" |
| 45 | #include "types/simple.h" |
| 46 | #include "gromacs/math/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_ElecRFCut_VdwLJSw_GeomW3P1_VF_sse4_1_single |
| 54 | * Electrostatics interaction: ReactionField |
| 55 | * VdW interaction: LennardJones |
| 56 | * Geometry: Water3-Particle |
| 57 | * Calculate force/pot: PotentialAndForce |
| 58 | */ |
| 59 | void |
| 60 | nb_kernel_ElecRFCut_VdwLJSw_GeomW3P1_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__attribute__ ((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*DIM3]; |
| 84 | __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall; |
| 85 | int vdwioffset0; |
| 86 | __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
| 87 | int vdwioffset1; |
| 88 | __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
| 89 | int vdwioffset2; |
| 90 | __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
| 91 | int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D; |
| 92 | __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
| 93 | __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00; |
| 94 | __m128 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10; |
| 95 | __m128 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20; |
| 96 | __m128 velec,felec,velecsum,facel,crf,krf,krf2; |
| 97 | real *charge; |
| 98 | int nvdwtype; |
| 99 | __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6; |
| 100 | int *vdwtype; |
| 101 | real *vdwparam; |
| 102 | __m128 one_sixth = _mm_set1_ps(1.0/6.0); |
| 103 | __m128 one_twelfth = _mm_set1_ps(1.0/12.0); |
| 104 | __m128 rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw; |
| 105 | real rswitch_scalar,d_scalar; |
| 106 | __m128 dummy_mask,cutoff_mask; |
| 107 | __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) ); |
| 108 | __m128 one = _mm_set1_ps(1.0); |
| 109 | __m128 two = _mm_set1_ps(2.0); |
| 110 | x = xx[0]; |
| 111 | f = ff[0]; |
| 112 | |
| 113 | nri = nlist->nri; |
| 114 | iinr = nlist->iinr; |
| 115 | jindex = nlist->jindex; |
| 116 | jjnr = nlist->jjnr; |
| 117 | shiftidx = nlist->shift; |
| 118 | gid = nlist->gid; |
| 119 | shiftvec = fr->shift_vec[0]; |
| 120 | fshift = fr->fshift[0]; |
| 121 | facel = _mm_set1_ps(fr->epsfac); |
| 122 | charge = mdatoms->chargeA; |
| 123 | krf = _mm_set1_ps(fr->ic->k_rf); |
| 124 | krf2 = _mm_set1_ps(fr->ic->k_rf*2.0); |
| 125 | crf = _mm_set1_ps(fr->ic->c_rf); |
| 126 | nvdwtype = fr->ntype; |
| 127 | vdwparam = fr->nbfp; |
| 128 | vdwtype = mdatoms->typeA; |
| 129 | |
| 130 | /* Setup water-specific parameters */ |
| 131 | inr = nlist->iinr[0]; |
| 132 | iq0 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+0])); |
| 133 | iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1])); |
| 134 | iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2])); |
| 135 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
| 136 | |
| 137 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
| 138 | rcutoff_scalar = fr->rcoulomb; |
| 139 | rcutoff = _mm_set1_ps(rcutoff_scalar); |
| 140 | rcutoff2 = _mm_mul_ps(rcutoff,rcutoff); |
| 141 | |
| 142 | rswitch_scalar = fr->rvdw_switch; |
| 143 | rswitch = _mm_set1_ps(rswitch_scalar); |
| 144 | /* Setup switch parameters */ |
| 145 | d_scalar = rcutoff_scalar-rswitch_scalar; |
| 146 | d = _mm_set1_ps(d_scalar); |
| 147 | swV3 = _mm_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar)); |
| 148 | swV4 = _mm_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar)); |
| 149 | swV5 = _mm_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar)); |
| 150 | swF2 = _mm_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar)); |
| 151 | swF3 = _mm_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar)); |
| 152 | swF4 = _mm_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar)); |
| 153 | |
| 154 | /* Avoid stupid compiler warnings */ |
| 155 | jnrA = jnrB = jnrC = jnrD = 0; |
| 156 | j_coord_offsetA = 0; |
Value stored to 'j_coord_offsetA' is never read | |
| 157 | j_coord_offsetB = 0; |
| 158 | j_coord_offsetC = 0; |
| 159 | j_coord_offsetD = 0; |
| 160 | |
| 161 | outeriter = 0; |
| 162 | inneriter = 0; |
| 163 | |
| 164 | for(iidx=0;iidx<4*DIM3;iidx++) |
| 165 | { |
| 166 | scratch[iidx] = 0.0; |
| 167 | } |
| 168 | |
| 169 | /* Start outer loop over neighborlists */ |
| 170 | for(iidx=0; iidx<nri; iidx++) |
| 171 | { |
| 172 | /* Load shift vector for this list */ |
| 173 | i_shift_offset = DIM3*shiftidx[iidx]; |
| 174 | |
| 175 | /* Load limits for loop over neighbors */ |
| 176 | j_index_start = jindex[iidx]; |
| 177 | j_index_end = jindex[iidx+1]; |
| 178 | |
| 179 | /* Get outer coordinate index */ |
| 180 | inr = iinr[iidx]; |
| 181 | i_coord_offset = DIM3*inr; |
| 182 | |
| 183 | /* Load i particle coords and add shift vector */ |
| 184 | gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset, |
| 185 | &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2); |
| 186 | |
| 187 | fix0 = _mm_setzero_ps(); |
| 188 | fiy0 = _mm_setzero_ps(); |
| 189 | fiz0 = _mm_setzero_ps(); |
| 190 | fix1 = _mm_setzero_ps(); |
| 191 | fiy1 = _mm_setzero_ps(); |
| 192 | fiz1 = _mm_setzero_ps(); |
| 193 | fix2 = _mm_setzero_ps(); |
| 194 | fiy2 = _mm_setzero_ps(); |
| 195 | fiz2 = _mm_setzero_ps(); |
| 196 | |
| 197 | /* Reset potential sums */ |
| 198 | velecsum = _mm_setzero_ps(); |
| 199 | vvdwsum = _mm_setzero_ps(); |
| 200 | |
| 201 | /* Start inner kernel loop */ |
| 202 | for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4) |
| 203 | { |
| 204 | |
| 205 | /* Get j neighbor index, and coordinate index */ |
| 206 | jnrA = jjnr[jidx]; |
| 207 | jnrB = jjnr[jidx+1]; |
| 208 | jnrC = jjnr[jidx+2]; |
| 209 | jnrD = jjnr[jidx+3]; |
| 210 | j_coord_offsetA = DIM3*jnrA; |
| 211 | j_coord_offsetB = DIM3*jnrB; |
| 212 | j_coord_offsetC = DIM3*jnrC; |
| 213 | j_coord_offsetD = DIM3*jnrD; |
| 214 | |
| 215 | /* load j atom coordinates */ |
| 216 | gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
| 217 | x+j_coord_offsetC,x+j_coord_offsetD, |
| 218 | &jx0,&jy0,&jz0); |
| 219 | |
| 220 | /* Calculate displacement vector */ |
| 221 | dx00 = _mm_sub_ps(ix0,jx0); |
| 222 | dy00 = _mm_sub_ps(iy0,jy0); |
| 223 | dz00 = _mm_sub_ps(iz0,jz0); |
| 224 | dx10 = _mm_sub_ps(ix1,jx0); |
| 225 | dy10 = _mm_sub_ps(iy1,jy0); |
| 226 | dz10 = _mm_sub_ps(iz1,jz0); |
| 227 | dx20 = _mm_sub_ps(ix2,jx0); |
| 228 | dy20 = _mm_sub_ps(iy2,jy0); |
| 229 | dz20 = _mm_sub_ps(iz2,jz0); |
| 230 | |
| 231 | /* Calculate squared distance and things based on it */ |
| 232 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
| 233 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
| 234 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
| 235 | |
| 236 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
| 237 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
| 238 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
| 239 | |
| 240 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
| 241 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
| 242 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
| 243 | |
| 244 | /* Load parameters for j particles */ |
| 245 | jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0, |
| 246 | charge+jnrC+0,charge+jnrD+0); |
| 247 | vdwjidx0A = 2*vdwtype[jnrA+0]; |
| 248 | vdwjidx0B = 2*vdwtype[jnrB+0]; |
| 249 | vdwjidx0C = 2*vdwtype[jnrC+0]; |
| 250 | vdwjidx0D = 2*vdwtype[jnrD+0]; |
| 251 | |
| 252 | fjx0 = _mm_setzero_ps(); |
| 253 | fjy0 = _mm_setzero_ps(); |
| 254 | fjz0 = _mm_setzero_ps(); |
| 255 | |
| 256 | /************************** |
| 257 | * CALCULATE INTERACTIONS * |
| 258 | **************************/ |
| 259 | |
| 260 | if (gmx_mm_any_lt(rsq00,rcutoff2)) |
| 261 | { |
| 262 | |
| 263 | r00 = _mm_mul_ps(rsq00,rinv00); |
| 264 | |
| 265 | /* Compute parameters for interactions between i and j atoms */ |
| 266 | qq00 = _mm_mul_ps(iq0,jq0); |
| 267 | gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A, |
| 268 | vdwparam+vdwioffset0+vdwjidx0B, |
| 269 | vdwparam+vdwioffset0+vdwjidx0C, |
| 270 | vdwparam+vdwioffset0+vdwjidx0D, |
| 271 | &c6_00,&c12_00); |
| 272 | |
| 273 | /* REACTION-FIELD ELECTROSTATICS */ |
| 274 | velec = _mm_mul_ps(qq00,_mm_sub_ps(_mm_add_ps(rinv00,_mm_mul_ps(krf,rsq00)),crf)); |
| 275 | felec = _mm_mul_ps(qq00,_mm_sub_ps(_mm_mul_ps(rinv00,rinvsq00),krf2)); |
| 276 | |
| 277 | /* LENNARD-JONES DISPERSION/REPULSION */ |
| 278 | |
| 279 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
| 280 | vvdw6 = _mm_mul_ps(c6_00,rinvsix); |
| 281 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
| 282 | vvdw = _mm_sub_ps( _mm_mul_ps(vvdw12,one_twelfth) , _mm_mul_ps(vvdw6,one_sixth) ); |
| 283 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00); |
| 284 | |
| 285 | d = _mm_sub_ps(r00,rswitch); |
| 286 | d = _mm_max_ps(d,_mm_setzero_ps()); |
| 287 | d2 = _mm_mul_ps(d,d); |
| 288 | sw = _mm_add_ps(one,_mm_mul_ps(d2,_mm_mul_ps(d,_mm_add_ps(swV3,_mm_mul_ps(d,_mm_add_ps(swV4,_mm_mul_ps(d,swV5))))))); |
| 289 | |
| 290 | dsw = _mm_mul_ps(d2,_mm_add_ps(swF2,_mm_mul_ps(d,_mm_add_ps(swF3,_mm_mul_ps(d,swF4))))); |
| 291 | |
| 292 | /* Evaluate switch function */ |
| 293 | /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */ |
| 294 | fvdw = _mm_sub_ps( _mm_mul_ps(fvdw,sw) , _mm_mul_ps(rinv00,_mm_mul_ps(vvdw,dsw)) ); |
| 295 | vvdw = _mm_mul_ps(vvdw,sw); |
| 296 | cutoff_mask = _mm_cmplt_ps(rsq00,rcutoff2); |
| 297 | |
| 298 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
| 299 | velec = _mm_and_ps(velec,cutoff_mask); |
| 300 | velecsum = _mm_add_ps(velecsum,velec); |
| 301 | vvdw = _mm_and_ps(vvdw,cutoff_mask); |
| 302 | vvdwsum = _mm_add_ps(vvdwsum,vvdw); |
| 303 | |
| 304 | fscal = _mm_add_ps(felec,fvdw); |
| 305 | |
| 306 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 307 | |
| 308 | /* Calculate temporary vectorial force */ |
| 309 | tx = _mm_mul_ps(fscal,dx00); |
| 310 | ty = _mm_mul_ps(fscal,dy00); |
| 311 | tz = _mm_mul_ps(fscal,dz00); |
| 312 | |
| 313 | /* Update vectorial force */ |
| 314 | fix0 = _mm_add_ps(fix0,tx); |
| 315 | fiy0 = _mm_add_ps(fiy0,ty); |
| 316 | fiz0 = _mm_add_ps(fiz0,tz); |
| 317 | |
| 318 | fjx0 = _mm_add_ps(fjx0,tx); |
| 319 | fjy0 = _mm_add_ps(fjy0,ty); |
| 320 | fjz0 = _mm_add_ps(fjz0,tz); |
| 321 | |
| 322 | } |
| 323 | |
| 324 | /************************** |
| 325 | * CALCULATE INTERACTIONS * |
| 326 | **************************/ |
| 327 | |
| 328 | if (gmx_mm_any_lt(rsq10,rcutoff2)) |
| 329 | { |
| 330 | |
| 331 | /* Compute parameters for interactions between i and j atoms */ |
| 332 | qq10 = _mm_mul_ps(iq1,jq0); |
| 333 | |
| 334 | /* REACTION-FIELD ELECTROSTATICS */ |
| 335 | velec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_add_ps(rinv10,_mm_mul_ps(krf,rsq10)),crf)); |
| 336 | felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2)); |
| 337 | |
| 338 | cutoff_mask = _mm_cmplt_ps(rsq10,rcutoff2); |
| 339 | |
| 340 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
| 341 | velec = _mm_and_ps(velec,cutoff_mask); |
| 342 | velecsum = _mm_add_ps(velecsum,velec); |
| 343 | |
| 344 | fscal = felec; |
| 345 | |
| 346 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 347 | |
| 348 | /* Calculate temporary vectorial force */ |
| 349 | tx = _mm_mul_ps(fscal,dx10); |
| 350 | ty = _mm_mul_ps(fscal,dy10); |
| 351 | tz = _mm_mul_ps(fscal,dz10); |
| 352 | |
| 353 | /* Update vectorial force */ |
| 354 | fix1 = _mm_add_ps(fix1,tx); |
| 355 | fiy1 = _mm_add_ps(fiy1,ty); |
| 356 | fiz1 = _mm_add_ps(fiz1,tz); |
| 357 | |
| 358 | fjx0 = _mm_add_ps(fjx0,tx); |
| 359 | fjy0 = _mm_add_ps(fjy0,ty); |
| 360 | fjz0 = _mm_add_ps(fjz0,tz); |
| 361 | |
| 362 | } |
| 363 | |
| 364 | /************************** |
| 365 | * CALCULATE INTERACTIONS * |
| 366 | **************************/ |
| 367 | |
| 368 | if (gmx_mm_any_lt(rsq20,rcutoff2)) |
| 369 | { |
| 370 | |
| 371 | /* Compute parameters for interactions between i and j atoms */ |
| 372 | qq20 = _mm_mul_ps(iq2,jq0); |
| 373 | |
| 374 | /* REACTION-FIELD ELECTROSTATICS */ |
| 375 | velec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_add_ps(rinv20,_mm_mul_ps(krf,rsq20)),crf)); |
| 376 | felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2)); |
| 377 | |
| 378 | cutoff_mask = _mm_cmplt_ps(rsq20,rcutoff2); |
| 379 | |
| 380 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
| 381 | velec = _mm_and_ps(velec,cutoff_mask); |
| 382 | velecsum = _mm_add_ps(velecsum,velec); |
| 383 | |
| 384 | fscal = felec; |
| 385 | |
| 386 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 387 | |
| 388 | /* Calculate temporary vectorial force */ |
| 389 | tx = _mm_mul_ps(fscal,dx20); |
| 390 | ty = _mm_mul_ps(fscal,dy20); |
| 391 | tz = _mm_mul_ps(fscal,dz20); |
| 392 | |
| 393 | /* Update vectorial force */ |
| 394 | fix2 = _mm_add_ps(fix2,tx); |
| 395 | fiy2 = _mm_add_ps(fiy2,ty); |
| 396 | fiz2 = _mm_add_ps(fiz2,tz); |
| 397 | |
| 398 | fjx0 = _mm_add_ps(fjx0,tx); |
| 399 | fjy0 = _mm_add_ps(fjy0,ty); |
| 400 | fjz0 = _mm_add_ps(fjz0,tz); |
| 401 | |
| 402 | } |
| 403 | |
| 404 | fjptrA = f+j_coord_offsetA; |
| 405 | fjptrB = f+j_coord_offsetB; |
| 406 | fjptrC = f+j_coord_offsetC; |
| 407 | fjptrD = f+j_coord_offsetD; |
| 408 | |
| 409 | gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0); |
| 410 | |
| 411 | /* Inner loop uses 142 flops */ |
| 412 | } |
| 413 | |
| 414 | if(jidx<j_index_end) |
| 415 | { |
| 416 | |
| 417 | /* Get j neighbor index, and coordinate index */ |
| 418 | jnrlistA = jjnr[jidx]; |
| 419 | jnrlistB = jjnr[jidx+1]; |
| 420 | jnrlistC = jjnr[jidx+2]; |
| 421 | jnrlistD = jjnr[jidx+3]; |
| 422 | /* Sign of each element will be negative for non-real atoms. |
| 423 | * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones, |
| 424 | * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries. |
| 425 | */ |
| 426 | dummy_mask = gmx_mm_castsi128_ps_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())); |
| 427 | jnrA = (jnrlistA>=0) ? jnrlistA : 0; |
| 428 | jnrB = (jnrlistB>=0) ? jnrlistB : 0; |
| 429 | jnrC = (jnrlistC>=0) ? jnrlistC : 0; |
| 430 | jnrD = (jnrlistD>=0) ? jnrlistD : 0; |
| 431 | j_coord_offsetA = DIM3*jnrA; |
| 432 | j_coord_offsetB = DIM3*jnrB; |
| 433 | j_coord_offsetC = DIM3*jnrC; |
| 434 | j_coord_offsetD = DIM3*jnrD; |
| 435 | |
| 436 | /* load j atom coordinates */ |
| 437 | gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
| 438 | x+j_coord_offsetC,x+j_coord_offsetD, |
| 439 | &jx0,&jy0,&jz0); |
| 440 | |
| 441 | /* Calculate displacement vector */ |
| 442 | dx00 = _mm_sub_ps(ix0,jx0); |
| 443 | dy00 = _mm_sub_ps(iy0,jy0); |
| 444 | dz00 = _mm_sub_ps(iz0,jz0); |
| 445 | dx10 = _mm_sub_ps(ix1,jx0); |
| 446 | dy10 = _mm_sub_ps(iy1,jy0); |
| 447 | dz10 = _mm_sub_ps(iz1,jz0); |
| 448 | dx20 = _mm_sub_ps(ix2,jx0); |
| 449 | dy20 = _mm_sub_ps(iy2,jy0); |
| 450 | dz20 = _mm_sub_ps(iz2,jz0); |
| 451 | |
| 452 | /* Calculate squared distance and things based on it */ |
| 453 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
| 454 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
| 455 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
| 456 | |
| 457 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
| 458 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
| 459 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
| 460 | |
| 461 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
| 462 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
| 463 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
| 464 | |
| 465 | /* Load parameters for j particles */ |
| 466 | jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0, |
| 467 | charge+jnrC+0,charge+jnrD+0); |
| 468 | vdwjidx0A = 2*vdwtype[jnrA+0]; |
| 469 | vdwjidx0B = 2*vdwtype[jnrB+0]; |
| 470 | vdwjidx0C = 2*vdwtype[jnrC+0]; |
| 471 | vdwjidx0D = 2*vdwtype[jnrD+0]; |
| 472 | |
| 473 | fjx0 = _mm_setzero_ps(); |
| 474 | fjy0 = _mm_setzero_ps(); |
| 475 | fjz0 = _mm_setzero_ps(); |
| 476 | |
| 477 | /************************** |
| 478 | * CALCULATE INTERACTIONS * |
| 479 | **************************/ |
| 480 | |
| 481 | if (gmx_mm_any_lt(rsq00,rcutoff2)) |
| 482 | { |
| 483 | |
| 484 | r00 = _mm_mul_ps(rsq00,rinv00); |
| 485 | r00 = _mm_andnot_ps(dummy_mask,r00); |
| 486 | |
| 487 | /* Compute parameters for interactions between i and j atoms */ |
| 488 | qq00 = _mm_mul_ps(iq0,jq0); |
| 489 | gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A, |
| 490 | vdwparam+vdwioffset0+vdwjidx0B, |
| 491 | vdwparam+vdwioffset0+vdwjidx0C, |
| 492 | vdwparam+vdwioffset0+vdwjidx0D, |
| 493 | &c6_00,&c12_00); |
| 494 | |
| 495 | /* REACTION-FIELD ELECTROSTATICS */ |
| 496 | velec = _mm_mul_ps(qq00,_mm_sub_ps(_mm_add_ps(rinv00,_mm_mul_ps(krf,rsq00)),crf)); |
| 497 | felec = _mm_mul_ps(qq00,_mm_sub_ps(_mm_mul_ps(rinv00,rinvsq00),krf2)); |
| 498 | |
| 499 | /* LENNARD-JONES DISPERSION/REPULSION */ |
| 500 | |
| 501 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
| 502 | vvdw6 = _mm_mul_ps(c6_00,rinvsix); |
| 503 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
| 504 | vvdw = _mm_sub_ps( _mm_mul_ps(vvdw12,one_twelfth) , _mm_mul_ps(vvdw6,one_sixth) ); |
| 505 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00); |
| 506 | |
| 507 | d = _mm_sub_ps(r00,rswitch); |
| 508 | d = _mm_max_ps(d,_mm_setzero_ps()); |
| 509 | d2 = _mm_mul_ps(d,d); |
| 510 | sw = _mm_add_ps(one,_mm_mul_ps(d2,_mm_mul_ps(d,_mm_add_ps(swV3,_mm_mul_ps(d,_mm_add_ps(swV4,_mm_mul_ps(d,swV5))))))); |
| 511 | |
| 512 | dsw = _mm_mul_ps(d2,_mm_add_ps(swF2,_mm_mul_ps(d,_mm_add_ps(swF3,_mm_mul_ps(d,swF4))))); |
| 513 | |
| 514 | /* Evaluate switch function */ |
| 515 | /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */ |
| 516 | fvdw = _mm_sub_ps( _mm_mul_ps(fvdw,sw) , _mm_mul_ps(rinv00,_mm_mul_ps(vvdw,dsw)) ); |
| 517 | vvdw = _mm_mul_ps(vvdw,sw); |
| 518 | cutoff_mask = _mm_cmplt_ps(rsq00,rcutoff2); |
| 519 | |
| 520 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
| 521 | velec = _mm_and_ps(velec,cutoff_mask); |
| 522 | velec = _mm_andnot_ps(dummy_mask,velec); |
| 523 | velecsum = _mm_add_ps(velecsum,velec); |
| 524 | vvdw = _mm_and_ps(vvdw,cutoff_mask); |
| 525 | vvdw = _mm_andnot_ps(dummy_mask,vvdw); |
| 526 | vvdwsum = _mm_add_ps(vvdwsum,vvdw); |
| 527 | |
| 528 | fscal = _mm_add_ps(felec,fvdw); |
| 529 | |
| 530 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 531 | |
| 532 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
| 533 | |
| 534 | /* Calculate temporary vectorial force */ |
| 535 | tx = _mm_mul_ps(fscal,dx00); |
| 536 | ty = _mm_mul_ps(fscal,dy00); |
| 537 | tz = _mm_mul_ps(fscal,dz00); |
| 538 | |
| 539 | /* Update vectorial force */ |
| 540 | fix0 = _mm_add_ps(fix0,tx); |
| 541 | fiy0 = _mm_add_ps(fiy0,ty); |
| 542 | fiz0 = _mm_add_ps(fiz0,tz); |
| 543 | |
| 544 | fjx0 = _mm_add_ps(fjx0,tx); |
| 545 | fjy0 = _mm_add_ps(fjy0,ty); |
| 546 | fjz0 = _mm_add_ps(fjz0,tz); |
| 547 | |
| 548 | } |
| 549 | |
| 550 | /************************** |
| 551 | * CALCULATE INTERACTIONS * |
| 552 | **************************/ |
| 553 | |
| 554 | if (gmx_mm_any_lt(rsq10,rcutoff2)) |
| 555 | { |
| 556 | |
| 557 | /* Compute parameters for interactions between i and j atoms */ |
| 558 | qq10 = _mm_mul_ps(iq1,jq0); |
| 559 | |
| 560 | /* REACTION-FIELD ELECTROSTATICS */ |
| 561 | velec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_add_ps(rinv10,_mm_mul_ps(krf,rsq10)),crf)); |
| 562 | felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2)); |
| 563 | |
| 564 | cutoff_mask = _mm_cmplt_ps(rsq10,rcutoff2); |
| 565 | |
| 566 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
| 567 | velec = _mm_and_ps(velec,cutoff_mask); |
| 568 | velec = _mm_andnot_ps(dummy_mask,velec); |
| 569 | velecsum = _mm_add_ps(velecsum,velec); |
| 570 | |
| 571 | fscal = felec; |
| 572 | |
| 573 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 574 | |
| 575 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
| 576 | |
| 577 | /* Calculate temporary vectorial force */ |
| 578 | tx = _mm_mul_ps(fscal,dx10); |
| 579 | ty = _mm_mul_ps(fscal,dy10); |
| 580 | tz = _mm_mul_ps(fscal,dz10); |
| 581 | |
| 582 | /* Update vectorial force */ |
| 583 | fix1 = _mm_add_ps(fix1,tx); |
| 584 | fiy1 = _mm_add_ps(fiy1,ty); |
| 585 | fiz1 = _mm_add_ps(fiz1,tz); |
| 586 | |
| 587 | fjx0 = _mm_add_ps(fjx0,tx); |
| 588 | fjy0 = _mm_add_ps(fjy0,ty); |
| 589 | fjz0 = _mm_add_ps(fjz0,tz); |
| 590 | |
| 591 | } |
| 592 | |
| 593 | /************************** |
| 594 | * CALCULATE INTERACTIONS * |
| 595 | **************************/ |
| 596 | |
| 597 | if (gmx_mm_any_lt(rsq20,rcutoff2)) |
| 598 | { |
| 599 | |
| 600 | /* Compute parameters for interactions between i and j atoms */ |
| 601 | qq20 = _mm_mul_ps(iq2,jq0); |
| 602 | |
| 603 | /* REACTION-FIELD ELECTROSTATICS */ |
| 604 | velec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_add_ps(rinv20,_mm_mul_ps(krf,rsq20)),crf)); |
| 605 | felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2)); |
| 606 | |
| 607 | cutoff_mask = _mm_cmplt_ps(rsq20,rcutoff2); |
| 608 | |
| 609 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
| 610 | velec = _mm_and_ps(velec,cutoff_mask); |
| 611 | velec = _mm_andnot_ps(dummy_mask,velec); |
| 612 | velecsum = _mm_add_ps(velecsum,velec); |
| 613 | |
| 614 | fscal = felec; |
| 615 | |
| 616 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 617 | |
| 618 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
| 619 | |
| 620 | /* Calculate temporary vectorial force */ |
| 621 | tx = _mm_mul_ps(fscal,dx20); |
| 622 | ty = _mm_mul_ps(fscal,dy20); |
| 623 | tz = _mm_mul_ps(fscal,dz20); |
| 624 | |
| 625 | /* Update vectorial force */ |
| 626 | fix2 = _mm_add_ps(fix2,tx); |
| 627 | fiy2 = _mm_add_ps(fiy2,ty); |
| 628 | fiz2 = _mm_add_ps(fiz2,tz); |
| 629 | |
| 630 | fjx0 = _mm_add_ps(fjx0,tx); |
| 631 | fjy0 = _mm_add_ps(fjy0,ty); |
| 632 | fjz0 = _mm_add_ps(fjz0,tz); |
| 633 | |
| 634 | } |
| 635 | |
| 636 | fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch; |
| 637 | fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch; |
| 638 | fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch; |
| 639 | fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch; |
| 640 | |
| 641 | gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0); |
| 642 | |
| 643 | /* Inner loop uses 143 flops */ |
| 644 | } |
| 645 | |
| 646 | /* End of innermost loop */ |
| 647 | |
| 648 | gmx_mm_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2, |
| 649 | f+i_coord_offset,fshift+i_shift_offset); |
| 650 | |
| 651 | ggid = gid[iidx]; |
| 652 | /* Update potential energies */ |
| 653 | gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid); |
| 654 | gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid); |
| 655 | |
| 656 | /* Increment number of inner iterations */ |
| 657 | inneriter += j_index_end - j_index_start; |
| 658 | |
| 659 | /* Outer loop uses 20 flops */ |
| 660 | } |
| 661 | |
| 662 | /* Increment number of outer iterations */ |
| 663 | outeriter += nri; |
| 664 | |
| 665 | /* Update outer/inner flops */ |
| 666 | |
| 667 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*20 + inneriter*143)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_VF] += outeriter*20 + inneriter *143; |
| 668 | } |
| 669 | /* |
| 670 | * Gromacs nonbonded kernel: nb_kernel_ElecRFCut_VdwLJSw_GeomW3P1_F_sse4_1_single |
| 671 | * Electrostatics interaction: ReactionField |
| 672 | * VdW interaction: LennardJones |
| 673 | * Geometry: Water3-Particle |
| 674 | * Calculate force/pot: Force |
| 675 | */ |
| 676 | void |
| 677 | nb_kernel_ElecRFCut_VdwLJSw_GeomW3P1_F_sse4_1_single |
| 678 | (t_nblist * gmx_restrict nlist, |
| 679 | rvec * gmx_restrict xx, |
| 680 | rvec * gmx_restrict ff, |
| 681 | t_forcerec * gmx_restrict fr, |
| 682 | t_mdatoms * gmx_restrict mdatoms, |
| 683 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict kernel_data, |
| 684 | t_nrnb * gmx_restrict nrnb) |
| 685 | { |
| 686 | /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or |
| 687 | * just 0 for non-waters. |
| 688 | * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different |
| 689 | * jnr indices corresponding to data put in the four positions in the SIMD register. |
| 690 | */ |
| 691 | int i_shift_offset,i_coord_offset,outeriter,inneriter; |
| 692 | int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx; |
| 693 | int jnrA,jnrB,jnrC,jnrD; |
| 694 | int jnrlistA,jnrlistB,jnrlistC,jnrlistD; |
| 695 | int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD; |
| 696 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
| 697 | real rcutoff_scalar; |
| 698 | real *shiftvec,*fshift,*x,*f; |
| 699 | real *fjptrA,*fjptrB,*fjptrC,*fjptrD; |
| 700 | real scratch[4*DIM3]; |
| 701 | __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall; |
| 702 | int vdwioffset0; |
| 703 | __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
| 704 | int vdwioffset1; |
| 705 | __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
| 706 | int vdwioffset2; |
| 707 | __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
| 708 | int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D; |
| 709 | __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
| 710 | __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00; |
| 711 | __m128 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10; |
| 712 | __m128 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20; |
| 713 | __m128 velec,felec,velecsum,facel,crf,krf,krf2; |
| 714 | real *charge; |
| 715 | int nvdwtype; |
| 716 | __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6; |
| 717 | int *vdwtype; |
| 718 | real *vdwparam; |
| 719 | __m128 one_sixth = _mm_set1_ps(1.0/6.0); |
| 720 | __m128 one_twelfth = _mm_set1_ps(1.0/12.0); |
| 721 | __m128 rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw; |
| 722 | real rswitch_scalar,d_scalar; |
| 723 | __m128 dummy_mask,cutoff_mask; |
| 724 | __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) ); |
| 725 | __m128 one = _mm_set1_ps(1.0); |
| 726 | __m128 two = _mm_set1_ps(2.0); |
| 727 | x = xx[0]; |
| 728 | f = ff[0]; |
| 729 | |
| 730 | nri = nlist->nri; |
| 731 | iinr = nlist->iinr; |
| 732 | jindex = nlist->jindex; |
| 733 | jjnr = nlist->jjnr; |
| 734 | shiftidx = nlist->shift; |
| 735 | gid = nlist->gid; |
| 736 | shiftvec = fr->shift_vec[0]; |
| 737 | fshift = fr->fshift[0]; |
| 738 | facel = _mm_set1_ps(fr->epsfac); |
| 739 | charge = mdatoms->chargeA; |
| 740 | krf = _mm_set1_ps(fr->ic->k_rf); |
| 741 | krf2 = _mm_set1_ps(fr->ic->k_rf*2.0); |
| 742 | crf = _mm_set1_ps(fr->ic->c_rf); |
| 743 | nvdwtype = fr->ntype; |
| 744 | vdwparam = fr->nbfp; |
| 745 | vdwtype = mdatoms->typeA; |
| 746 | |
| 747 | /* Setup water-specific parameters */ |
| 748 | inr = nlist->iinr[0]; |
| 749 | iq0 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+0])); |
| 750 | iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1])); |
| 751 | iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2])); |
| 752 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
| 753 | |
| 754 | /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */ |
| 755 | rcutoff_scalar = fr->rcoulomb; |
| 756 | rcutoff = _mm_set1_ps(rcutoff_scalar); |
| 757 | rcutoff2 = _mm_mul_ps(rcutoff,rcutoff); |
| 758 | |
| 759 | rswitch_scalar = fr->rvdw_switch; |
| 760 | rswitch = _mm_set1_ps(rswitch_scalar); |
| 761 | /* Setup switch parameters */ |
| 762 | d_scalar = rcutoff_scalar-rswitch_scalar; |
| 763 | d = _mm_set1_ps(d_scalar); |
| 764 | swV3 = _mm_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar)); |
| 765 | swV4 = _mm_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar)); |
| 766 | swV5 = _mm_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar)); |
| 767 | swF2 = _mm_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar)); |
| 768 | swF3 = _mm_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar)); |
| 769 | swF4 = _mm_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar)); |
| 770 | |
| 771 | /* Avoid stupid compiler warnings */ |
| 772 | jnrA = jnrB = jnrC = jnrD = 0; |
| 773 | j_coord_offsetA = 0; |
| 774 | j_coord_offsetB = 0; |
| 775 | j_coord_offsetC = 0; |
| 776 | j_coord_offsetD = 0; |
| 777 | |
| 778 | outeriter = 0; |
| 779 | inneriter = 0; |
| 780 | |
| 781 | for(iidx=0;iidx<4*DIM3;iidx++) |
| 782 | { |
| 783 | scratch[iidx] = 0.0; |
| 784 | } |
| 785 | |
| 786 | /* Start outer loop over neighborlists */ |
| 787 | for(iidx=0; iidx<nri; iidx++) |
| 788 | { |
| 789 | /* Load shift vector for this list */ |
| 790 | i_shift_offset = DIM3*shiftidx[iidx]; |
| 791 | |
| 792 | /* Load limits for loop over neighbors */ |
| 793 | j_index_start = jindex[iidx]; |
| 794 | j_index_end = jindex[iidx+1]; |
| 795 | |
| 796 | /* Get outer coordinate index */ |
| 797 | inr = iinr[iidx]; |
| 798 | i_coord_offset = DIM3*inr; |
| 799 | |
| 800 | /* Load i particle coords and add shift vector */ |
| 801 | gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset, |
| 802 | &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2); |
| 803 | |
| 804 | fix0 = _mm_setzero_ps(); |
| 805 | fiy0 = _mm_setzero_ps(); |
| 806 | fiz0 = _mm_setzero_ps(); |
| 807 | fix1 = _mm_setzero_ps(); |
| 808 | fiy1 = _mm_setzero_ps(); |
| 809 | fiz1 = _mm_setzero_ps(); |
| 810 | fix2 = _mm_setzero_ps(); |
| 811 | fiy2 = _mm_setzero_ps(); |
| 812 | fiz2 = _mm_setzero_ps(); |
| 813 | |
| 814 | /* Start inner kernel loop */ |
| 815 | for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4) |
| 816 | { |
| 817 | |
| 818 | /* Get j neighbor index, and coordinate index */ |
| 819 | jnrA = jjnr[jidx]; |
| 820 | jnrB = jjnr[jidx+1]; |
| 821 | jnrC = jjnr[jidx+2]; |
| 822 | jnrD = jjnr[jidx+3]; |
| 823 | j_coord_offsetA = DIM3*jnrA; |
| 824 | j_coord_offsetB = DIM3*jnrB; |
| 825 | j_coord_offsetC = DIM3*jnrC; |
| 826 | j_coord_offsetD = DIM3*jnrD; |
| 827 | |
| 828 | /* load j atom coordinates */ |
| 829 | gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
| 830 | x+j_coord_offsetC,x+j_coord_offsetD, |
| 831 | &jx0,&jy0,&jz0); |
| 832 | |
| 833 | /* Calculate displacement vector */ |
| 834 | dx00 = _mm_sub_ps(ix0,jx0); |
| 835 | dy00 = _mm_sub_ps(iy0,jy0); |
| 836 | dz00 = _mm_sub_ps(iz0,jz0); |
| 837 | dx10 = _mm_sub_ps(ix1,jx0); |
| 838 | dy10 = _mm_sub_ps(iy1,jy0); |
| 839 | dz10 = _mm_sub_ps(iz1,jz0); |
| 840 | dx20 = _mm_sub_ps(ix2,jx0); |
| 841 | dy20 = _mm_sub_ps(iy2,jy0); |
| 842 | dz20 = _mm_sub_ps(iz2,jz0); |
| 843 | |
| 844 | /* Calculate squared distance and things based on it */ |
| 845 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
| 846 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
| 847 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
| 848 | |
| 849 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
| 850 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
| 851 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
| 852 | |
| 853 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
| 854 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
| 855 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
| 856 | |
| 857 | /* Load parameters for j particles */ |
| 858 | jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0, |
| 859 | charge+jnrC+0,charge+jnrD+0); |
| 860 | vdwjidx0A = 2*vdwtype[jnrA+0]; |
| 861 | vdwjidx0B = 2*vdwtype[jnrB+0]; |
| 862 | vdwjidx0C = 2*vdwtype[jnrC+0]; |
| 863 | vdwjidx0D = 2*vdwtype[jnrD+0]; |
| 864 | |
| 865 | fjx0 = _mm_setzero_ps(); |
| 866 | fjy0 = _mm_setzero_ps(); |
| 867 | fjz0 = _mm_setzero_ps(); |
| 868 | |
| 869 | /************************** |
| 870 | * CALCULATE INTERACTIONS * |
| 871 | **************************/ |
| 872 | |
| 873 | if (gmx_mm_any_lt(rsq00,rcutoff2)) |
| 874 | { |
| 875 | |
| 876 | r00 = _mm_mul_ps(rsq00,rinv00); |
| 877 | |
| 878 | /* Compute parameters for interactions between i and j atoms */ |
| 879 | qq00 = _mm_mul_ps(iq0,jq0); |
| 880 | gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A, |
| 881 | vdwparam+vdwioffset0+vdwjidx0B, |
| 882 | vdwparam+vdwioffset0+vdwjidx0C, |
| 883 | vdwparam+vdwioffset0+vdwjidx0D, |
| 884 | &c6_00,&c12_00); |
| 885 | |
| 886 | /* REACTION-FIELD ELECTROSTATICS */ |
| 887 | felec = _mm_mul_ps(qq00,_mm_sub_ps(_mm_mul_ps(rinv00,rinvsq00),krf2)); |
| 888 | |
| 889 | /* LENNARD-JONES DISPERSION/REPULSION */ |
| 890 | |
| 891 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
| 892 | vvdw6 = _mm_mul_ps(c6_00,rinvsix); |
| 893 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
| 894 | vvdw = _mm_sub_ps( _mm_mul_ps(vvdw12,one_twelfth) , _mm_mul_ps(vvdw6,one_sixth) ); |
| 895 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00); |
| 896 | |
| 897 | d = _mm_sub_ps(r00,rswitch); |
| 898 | d = _mm_max_ps(d,_mm_setzero_ps()); |
| 899 | d2 = _mm_mul_ps(d,d); |
| 900 | sw = _mm_add_ps(one,_mm_mul_ps(d2,_mm_mul_ps(d,_mm_add_ps(swV3,_mm_mul_ps(d,_mm_add_ps(swV4,_mm_mul_ps(d,swV5))))))); |
| 901 | |
| 902 | dsw = _mm_mul_ps(d2,_mm_add_ps(swF2,_mm_mul_ps(d,_mm_add_ps(swF3,_mm_mul_ps(d,swF4))))); |
| 903 | |
| 904 | /* Evaluate switch function */ |
| 905 | /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */ |
| 906 | fvdw = _mm_sub_ps( _mm_mul_ps(fvdw,sw) , _mm_mul_ps(rinv00,_mm_mul_ps(vvdw,dsw)) ); |
| 907 | cutoff_mask = _mm_cmplt_ps(rsq00,rcutoff2); |
| 908 | |
| 909 | fscal = _mm_add_ps(felec,fvdw); |
| 910 | |
| 911 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 912 | |
| 913 | /* Calculate temporary vectorial force */ |
| 914 | tx = _mm_mul_ps(fscal,dx00); |
| 915 | ty = _mm_mul_ps(fscal,dy00); |
| 916 | tz = _mm_mul_ps(fscal,dz00); |
| 917 | |
| 918 | /* Update vectorial force */ |
| 919 | fix0 = _mm_add_ps(fix0,tx); |
| 920 | fiy0 = _mm_add_ps(fiy0,ty); |
| 921 | fiz0 = _mm_add_ps(fiz0,tz); |
| 922 | |
| 923 | fjx0 = _mm_add_ps(fjx0,tx); |
| 924 | fjy0 = _mm_add_ps(fjy0,ty); |
| 925 | fjz0 = _mm_add_ps(fjz0,tz); |
| 926 | |
| 927 | } |
| 928 | |
| 929 | /************************** |
| 930 | * CALCULATE INTERACTIONS * |
| 931 | **************************/ |
| 932 | |
| 933 | if (gmx_mm_any_lt(rsq10,rcutoff2)) |
| 934 | { |
| 935 | |
| 936 | /* Compute parameters for interactions between i and j atoms */ |
| 937 | qq10 = _mm_mul_ps(iq1,jq0); |
| 938 | |
| 939 | /* REACTION-FIELD ELECTROSTATICS */ |
| 940 | felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2)); |
| 941 | |
| 942 | cutoff_mask = _mm_cmplt_ps(rsq10,rcutoff2); |
| 943 | |
| 944 | fscal = felec; |
| 945 | |
| 946 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 947 | |
| 948 | /* Calculate temporary vectorial force */ |
| 949 | tx = _mm_mul_ps(fscal,dx10); |
| 950 | ty = _mm_mul_ps(fscal,dy10); |
| 951 | tz = _mm_mul_ps(fscal,dz10); |
| 952 | |
| 953 | /* Update vectorial force */ |
| 954 | fix1 = _mm_add_ps(fix1,tx); |
| 955 | fiy1 = _mm_add_ps(fiy1,ty); |
| 956 | fiz1 = _mm_add_ps(fiz1,tz); |
| 957 | |
| 958 | fjx0 = _mm_add_ps(fjx0,tx); |
| 959 | fjy0 = _mm_add_ps(fjy0,ty); |
| 960 | fjz0 = _mm_add_ps(fjz0,tz); |
| 961 | |
| 962 | } |
| 963 | |
| 964 | /************************** |
| 965 | * CALCULATE INTERACTIONS * |
| 966 | **************************/ |
| 967 | |
| 968 | if (gmx_mm_any_lt(rsq20,rcutoff2)) |
| 969 | { |
| 970 | |
| 971 | /* Compute parameters for interactions between i and j atoms */ |
| 972 | qq20 = _mm_mul_ps(iq2,jq0); |
| 973 | |
| 974 | /* REACTION-FIELD ELECTROSTATICS */ |
| 975 | felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2)); |
| 976 | |
| 977 | cutoff_mask = _mm_cmplt_ps(rsq20,rcutoff2); |
| 978 | |
| 979 | fscal = felec; |
| 980 | |
| 981 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 982 | |
| 983 | /* Calculate temporary vectorial force */ |
| 984 | tx = _mm_mul_ps(fscal,dx20); |
| 985 | ty = _mm_mul_ps(fscal,dy20); |
| 986 | tz = _mm_mul_ps(fscal,dz20); |
| 987 | |
| 988 | /* Update vectorial force */ |
| 989 | fix2 = _mm_add_ps(fix2,tx); |
| 990 | fiy2 = _mm_add_ps(fiy2,ty); |
| 991 | fiz2 = _mm_add_ps(fiz2,tz); |
| 992 | |
| 993 | fjx0 = _mm_add_ps(fjx0,tx); |
| 994 | fjy0 = _mm_add_ps(fjy0,ty); |
| 995 | fjz0 = _mm_add_ps(fjz0,tz); |
| 996 | |
| 997 | } |
| 998 | |
| 999 | fjptrA = f+j_coord_offsetA; |
| 1000 | fjptrB = f+j_coord_offsetB; |
| 1001 | fjptrC = f+j_coord_offsetC; |
| 1002 | fjptrD = f+j_coord_offsetD; |
| 1003 | |
| 1004 | gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0); |
| 1005 | |
| 1006 | /* Inner loop uses 121 flops */ |
| 1007 | } |
| 1008 | |
| 1009 | if(jidx<j_index_end) |
| 1010 | { |
| 1011 | |
| 1012 | /* Get j neighbor index, and coordinate index */ |
| 1013 | jnrlistA = jjnr[jidx]; |
| 1014 | jnrlistB = jjnr[jidx+1]; |
| 1015 | jnrlistC = jjnr[jidx+2]; |
| 1016 | jnrlistD = jjnr[jidx+3]; |
| 1017 | /* Sign of each element will be negative for non-real atoms. |
| 1018 | * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones, |
| 1019 | * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries. |
| 1020 | */ |
| 1021 | dummy_mask = gmx_mm_castsi128_ps_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())); |
| 1022 | jnrA = (jnrlistA>=0) ? jnrlistA : 0; |
| 1023 | jnrB = (jnrlistB>=0) ? jnrlistB : 0; |
| 1024 | jnrC = (jnrlistC>=0) ? jnrlistC : 0; |
| 1025 | jnrD = (jnrlistD>=0) ? jnrlistD : 0; |
| 1026 | j_coord_offsetA = DIM3*jnrA; |
| 1027 | j_coord_offsetB = DIM3*jnrB; |
| 1028 | j_coord_offsetC = DIM3*jnrC; |
| 1029 | j_coord_offsetD = DIM3*jnrD; |
| 1030 | |
| 1031 | /* load j atom coordinates */ |
| 1032 | gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
| 1033 | x+j_coord_offsetC,x+j_coord_offsetD, |
| 1034 | &jx0,&jy0,&jz0); |
| 1035 | |
| 1036 | /* Calculate displacement vector */ |
| 1037 | dx00 = _mm_sub_ps(ix0,jx0); |
| 1038 | dy00 = _mm_sub_ps(iy0,jy0); |
| 1039 | dz00 = _mm_sub_ps(iz0,jz0); |
| 1040 | dx10 = _mm_sub_ps(ix1,jx0); |
| 1041 | dy10 = _mm_sub_ps(iy1,jy0); |
| 1042 | dz10 = _mm_sub_ps(iz1,jz0); |
| 1043 | dx20 = _mm_sub_ps(ix2,jx0); |
| 1044 | dy20 = _mm_sub_ps(iy2,jy0); |
| 1045 | dz20 = _mm_sub_ps(iz2,jz0); |
| 1046 | |
| 1047 | /* Calculate squared distance and things based on it */ |
| 1048 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
| 1049 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
| 1050 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
| 1051 | |
| 1052 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
| 1053 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
| 1054 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
| 1055 | |
| 1056 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
| 1057 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
| 1058 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
| 1059 | |
| 1060 | /* Load parameters for j particles */ |
| 1061 | jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0, |
| 1062 | charge+jnrC+0,charge+jnrD+0); |
| 1063 | vdwjidx0A = 2*vdwtype[jnrA+0]; |
| 1064 | vdwjidx0B = 2*vdwtype[jnrB+0]; |
| 1065 | vdwjidx0C = 2*vdwtype[jnrC+0]; |
| 1066 | vdwjidx0D = 2*vdwtype[jnrD+0]; |
| 1067 | |
| 1068 | fjx0 = _mm_setzero_ps(); |
| 1069 | fjy0 = _mm_setzero_ps(); |
| 1070 | fjz0 = _mm_setzero_ps(); |
| 1071 | |
| 1072 | /************************** |
| 1073 | * CALCULATE INTERACTIONS * |
| 1074 | **************************/ |
| 1075 | |
| 1076 | if (gmx_mm_any_lt(rsq00,rcutoff2)) |
| 1077 | { |
| 1078 | |
| 1079 | r00 = _mm_mul_ps(rsq00,rinv00); |
| 1080 | r00 = _mm_andnot_ps(dummy_mask,r00); |
| 1081 | |
| 1082 | /* Compute parameters for interactions between i and j atoms */ |
| 1083 | qq00 = _mm_mul_ps(iq0,jq0); |
| 1084 | gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A, |
| 1085 | vdwparam+vdwioffset0+vdwjidx0B, |
| 1086 | vdwparam+vdwioffset0+vdwjidx0C, |
| 1087 | vdwparam+vdwioffset0+vdwjidx0D, |
| 1088 | &c6_00,&c12_00); |
| 1089 | |
| 1090 | /* REACTION-FIELD ELECTROSTATICS */ |
| 1091 | felec = _mm_mul_ps(qq00,_mm_sub_ps(_mm_mul_ps(rinv00,rinvsq00),krf2)); |
| 1092 | |
| 1093 | /* LENNARD-JONES DISPERSION/REPULSION */ |
| 1094 | |
| 1095 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
| 1096 | vvdw6 = _mm_mul_ps(c6_00,rinvsix); |
| 1097 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
| 1098 | vvdw = _mm_sub_ps( _mm_mul_ps(vvdw12,one_twelfth) , _mm_mul_ps(vvdw6,one_sixth) ); |
| 1099 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00); |
| 1100 | |
| 1101 | d = _mm_sub_ps(r00,rswitch); |
| 1102 | d = _mm_max_ps(d,_mm_setzero_ps()); |
| 1103 | d2 = _mm_mul_ps(d,d); |
| 1104 | sw = _mm_add_ps(one,_mm_mul_ps(d2,_mm_mul_ps(d,_mm_add_ps(swV3,_mm_mul_ps(d,_mm_add_ps(swV4,_mm_mul_ps(d,swV5))))))); |
| 1105 | |
| 1106 | dsw = _mm_mul_ps(d2,_mm_add_ps(swF2,_mm_mul_ps(d,_mm_add_ps(swF3,_mm_mul_ps(d,swF4))))); |
| 1107 | |
| 1108 | /* Evaluate switch function */ |
| 1109 | /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */ |
| 1110 | fvdw = _mm_sub_ps( _mm_mul_ps(fvdw,sw) , _mm_mul_ps(rinv00,_mm_mul_ps(vvdw,dsw)) ); |
| 1111 | cutoff_mask = _mm_cmplt_ps(rsq00,rcutoff2); |
| 1112 | |
| 1113 | fscal = _mm_add_ps(felec,fvdw); |
| 1114 | |
| 1115 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 1116 | |
| 1117 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
| 1118 | |
| 1119 | /* Calculate temporary vectorial force */ |
| 1120 | tx = _mm_mul_ps(fscal,dx00); |
| 1121 | ty = _mm_mul_ps(fscal,dy00); |
| 1122 | tz = _mm_mul_ps(fscal,dz00); |
| 1123 | |
| 1124 | /* Update vectorial force */ |
| 1125 | fix0 = _mm_add_ps(fix0,tx); |
| 1126 | fiy0 = _mm_add_ps(fiy0,ty); |
| 1127 | fiz0 = _mm_add_ps(fiz0,tz); |
| 1128 | |
| 1129 | fjx0 = _mm_add_ps(fjx0,tx); |
| 1130 | fjy0 = _mm_add_ps(fjy0,ty); |
| 1131 | fjz0 = _mm_add_ps(fjz0,tz); |
| 1132 | |
| 1133 | } |
| 1134 | |
| 1135 | /************************** |
| 1136 | * CALCULATE INTERACTIONS * |
| 1137 | **************************/ |
| 1138 | |
| 1139 | if (gmx_mm_any_lt(rsq10,rcutoff2)) |
| 1140 | { |
| 1141 | |
| 1142 | /* Compute parameters for interactions between i and j atoms */ |
| 1143 | qq10 = _mm_mul_ps(iq1,jq0); |
| 1144 | |
| 1145 | /* REACTION-FIELD ELECTROSTATICS */ |
| 1146 | felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2)); |
| 1147 | |
| 1148 | cutoff_mask = _mm_cmplt_ps(rsq10,rcutoff2); |
| 1149 | |
| 1150 | fscal = felec; |
| 1151 | |
| 1152 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 1153 | |
| 1154 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
| 1155 | |
| 1156 | /* Calculate temporary vectorial force */ |
| 1157 | tx = _mm_mul_ps(fscal,dx10); |
| 1158 | ty = _mm_mul_ps(fscal,dy10); |
| 1159 | tz = _mm_mul_ps(fscal,dz10); |
| 1160 | |
| 1161 | /* Update vectorial force */ |
| 1162 | fix1 = _mm_add_ps(fix1,tx); |
| 1163 | fiy1 = _mm_add_ps(fiy1,ty); |
| 1164 | fiz1 = _mm_add_ps(fiz1,tz); |
| 1165 | |
| 1166 | fjx0 = _mm_add_ps(fjx0,tx); |
| 1167 | fjy0 = _mm_add_ps(fjy0,ty); |
| 1168 | fjz0 = _mm_add_ps(fjz0,tz); |
| 1169 | |
| 1170 | } |
| 1171 | |
| 1172 | /************************** |
| 1173 | * CALCULATE INTERACTIONS * |
| 1174 | **************************/ |
| 1175 | |
| 1176 | if (gmx_mm_any_lt(rsq20,rcutoff2)) |
| 1177 | { |
| 1178 | |
| 1179 | /* Compute parameters for interactions between i and j atoms */ |
| 1180 | qq20 = _mm_mul_ps(iq2,jq0); |
| 1181 | |
| 1182 | /* REACTION-FIELD ELECTROSTATICS */ |
| 1183 | felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2)); |
| 1184 | |
| 1185 | cutoff_mask = _mm_cmplt_ps(rsq20,rcutoff2); |
| 1186 | |
| 1187 | fscal = felec; |
| 1188 | |
| 1189 | fscal = _mm_and_ps(fscal,cutoff_mask); |
| 1190 | |
| 1191 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
| 1192 | |
| 1193 | /* Calculate temporary vectorial force */ |
| 1194 | tx = _mm_mul_ps(fscal,dx20); |
| 1195 | ty = _mm_mul_ps(fscal,dy20); |
| 1196 | tz = _mm_mul_ps(fscal,dz20); |
| 1197 | |
| 1198 | /* Update vectorial force */ |
| 1199 | fix2 = _mm_add_ps(fix2,tx); |
| 1200 | fiy2 = _mm_add_ps(fiy2,ty); |
| 1201 | fiz2 = _mm_add_ps(fiz2,tz); |
| 1202 | |
| 1203 | fjx0 = _mm_add_ps(fjx0,tx); |
| 1204 | fjy0 = _mm_add_ps(fjy0,ty); |
| 1205 | fjz0 = _mm_add_ps(fjz0,tz); |
| 1206 | |
| 1207 | } |
| 1208 | |
| 1209 | fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch; |
| 1210 | fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch; |
| 1211 | fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch; |
| 1212 | fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch; |
| 1213 | |
| 1214 | gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0); |
| 1215 | |
| 1216 | /* Inner loop uses 122 flops */ |
| 1217 | } |
| 1218 | |
| 1219 | /* End of innermost loop */ |
| 1220 | |
| 1221 | gmx_mm_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2, |
| 1222 | f+i_coord_offset,fshift+i_shift_offset); |
| 1223 | |
| 1224 | /* Increment number of inner iterations */ |
| 1225 | inneriter += j_index_end - j_index_start; |
| 1226 | |
| 1227 | /* Outer loop uses 18 flops */ |
| 1228 | } |
| 1229 | |
| 1230 | /* Increment number of outer iterations */ |
| 1231 | outeriter += nri; |
| 1232 | |
| 1233 | /* Update outer/inner flops */ |
| 1234 | |
| 1235 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*18 + inneriter*122)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3_F] += outeriter*18 + inneriter *122; |
| 1236 | } |