| File: | gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwLJEw_GeomW3W3_c.c |
| Location: | line 132, column 5 |
| Description: | Value stored to 'sh_lj_ewald' 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 c 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 | /* |
| 50 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJEw_GeomW3W3_VF_c |
| 51 | * Electrostatics interaction: Ewald |
| 52 | * VdW interaction: LJEwald |
| 53 | * Geometry: Water3-Water3 |
| 54 | * Calculate force/pot: PotentialAndForce |
| 55 | */ |
| 56 | void |
| 57 | nb_kernel_ElecEw_VdwLJEw_GeomW3W3_VF_c |
| 58 | (t_nblist * gmx_restrict__restrict nlist, |
| 59 | rvec * gmx_restrict__restrict xx, |
| 60 | rvec * gmx_restrict__restrict ff, |
| 61 | t_forcerec * gmx_restrict__restrict fr, |
| 62 | t_mdatoms * gmx_restrict__restrict mdatoms, |
| 63 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
| 64 | t_nrnb * gmx_restrict__restrict nrnb) |
| 65 | { |
| 66 | int i_shift_offset,i_coord_offset,j_coord_offset; |
| 67 | int j_index_start,j_index_end; |
| 68 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
| 69 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
| 70 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
| 71 | real *shiftvec,*fshift,*x,*f; |
| 72 | int vdwioffset0; |
| 73 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
| 74 | int vdwioffset1; |
| 75 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
| 76 | int vdwioffset2; |
| 77 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
| 78 | int vdwjidx0; |
| 79 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
| 80 | int vdwjidx1; |
| 81 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
| 82 | int vdwjidx2; |
| 83 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
| 84 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
| 85 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
| 86 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
| 87 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
| 88 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
| 89 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
| 90 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
| 91 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
| 92 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
| 93 | real velec,felec,velecsum,facel,crf,krf,krf2; |
| 94 | real *charge; |
| 95 | int nvdwtype; |
| 96 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
| 97 | int *vdwtype; |
| 98 | real *vdwparam; |
| 99 | real c6grid_00; |
| 100 | real c6grid_01; |
| 101 | real c6grid_02; |
| 102 | real c6grid_10; |
| 103 | real c6grid_11; |
| 104 | real c6grid_12; |
| 105 | real c6grid_20; |
| 106 | real c6grid_21; |
| 107 | real c6grid_22; |
| 108 | real ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald; |
| 109 | real *vdwgridparam; |
| 110 | int ewitab; |
| 111 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
| 112 | real *ewtab; |
| 113 | |
| 114 | x = xx[0]; |
| 115 | f = ff[0]; |
| 116 | |
| 117 | nri = nlist->nri; |
| 118 | iinr = nlist->iinr; |
| 119 | jindex = nlist->jindex; |
| 120 | jjnr = nlist->jjnr; |
| 121 | shiftidx = nlist->shift; |
| 122 | gid = nlist->gid; |
| 123 | shiftvec = fr->shift_vec[0]; |
| 124 | fshift = fr->fshift[0]; |
| 125 | facel = fr->epsfac; |
| 126 | charge = mdatoms->chargeA; |
| 127 | nvdwtype = fr->ntype; |
| 128 | vdwparam = fr->nbfp; |
| 129 | vdwtype = mdatoms->typeA; |
| 130 | vdwgridparam = fr->ljpme_c6grid; |
| 131 | ewclj = fr->ewaldcoeff_lj; |
| 132 | sh_lj_ewald = fr->ic->sh_lj_ewald; |
Value stored to 'sh_lj_ewald' is never read | |
| 133 | ewclj2 = ewclj*ewclj; |
| 134 | ewclj6 = ewclj2*ewclj2*ewclj2; |
| 135 | |
| 136 | sh_ewald = fr->ic->sh_ewald; |
| 137 | ewtab = fr->ic->tabq_coul_FDV0; |
| 138 | ewtabscale = fr->ic->tabq_scale; |
| 139 | ewtabhalfspace = 0.5/ewtabscale; |
| 140 | |
| 141 | /* Setup water-specific parameters */ |
| 142 | inr = nlist->iinr[0]; |
| 143 | iq0 = facel*charge[inr+0]; |
| 144 | iq1 = facel*charge[inr+1]; |
| 145 | iq2 = facel*charge[inr+2]; |
| 146 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
| 147 | |
| 148 | jq0 = charge[inr+0]; |
| 149 | jq1 = charge[inr+1]; |
| 150 | jq2 = charge[inr+2]; |
| 151 | vdwjidx0 = 2*vdwtype[inr+0]; |
| 152 | qq00 = iq0*jq0; |
| 153 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
| 154 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
| 155 | c6grid_00 = vdwgridparam[vdwioffset0+vdwjidx0]; |
| 156 | qq01 = iq0*jq1; |
| 157 | qq02 = iq0*jq2; |
| 158 | qq10 = iq1*jq0; |
| 159 | qq11 = iq1*jq1; |
| 160 | qq12 = iq1*jq2; |
| 161 | qq20 = iq2*jq0; |
| 162 | qq21 = iq2*jq1; |
| 163 | qq22 = iq2*jq2; |
| 164 | |
| 165 | outeriter = 0; |
| 166 | inneriter = 0; |
| 167 | |
| 168 | /* Start outer loop over neighborlists */ |
| 169 | for(iidx=0; iidx<nri; iidx++) |
| 170 | { |
| 171 | /* Load shift vector for this list */ |
| 172 | i_shift_offset = DIM3*shiftidx[iidx]; |
| 173 | shX = shiftvec[i_shift_offset+XX0]; |
| 174 | shY = shiftvec[i_shift_offset+YY1]; |
| 175 | shZ = shiftvec[i_shift_offset+ZZ2]; |
| 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 = DIM3*inr; |
| 184 | |
| 185 | /* Load i particle coords and add shift vector */ |
| 186 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
| 187 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
| 188 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
| 189 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
| 190 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
| 191 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
| 192 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
| 193 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
| 194 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
| 195 | |
| 196 | fix0 = 0.0; |
| 197 | fiy0 = 0.0; |
| 198 | fiz0 = 0.0; |
| 199 | fix1 = 0.0; |
| 200 | fiy1 = 0.0; |
| 201 | fiz1 = 0.0; |
| 202 | fix2 = 0.0; |
| 203 | fiy2 = 0.0; |
| 204 | fiz2 = 0.0; |
| 205 | |
| 206 | /* Reset potential sums */ |
| 207 | velecsum = 0.0; |
| 208 | vvdwsum = 0.0; |
| 209 | |
| 210 | /* Start inner kernel loop */ |
| 211 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
| 212 | { |
| 213 | /* Get j neighbor index, and coordinate index */ |
| 214 | jnr = jjnr[jidx]; |
| 215 | j_coord_offset = DIM3*jnr; |
| 216 | |
| 217 | /* load j atom coordinates */ |
| 218 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
| 219 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
| 220 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
| 221 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
| 222 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
| 223 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
| 224 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
| 225 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
| 226 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
| 227 | |
| 228 | /* Calculate displacement vector */ |
| 229 | dx00 = ix0 - jx0; |
| 230 | dy00 = iy0 - jy0; |
| 231 | dz00 = iz0 - jz0; |
| 232 | dx01 = ix0 - jx1; |
| 233 | dy01 = iy0 - jy1; |
| 234 | dz01 = iz0 - jz1; |
| 235 | dx02 = ix0 - jx2; |
| 236 | dy02 = iy0 - jy2; |
| 237 | dz02 = iz0 - jz2; |
| 238 | dx10 = ix1 - jx0; |
| 239 | dy10 = iy1 - jy0; |
| 240 | dz10 = iz1 - jz0; |
| 241 | dx11 = ix1 - jx1; |
| 242 | dy11 = iy1 - jy1; |
| 243 | dz11 = iz1 - jz1; |
| 244 | dx12 = ix1 - jx2; |
| 245 | dy12 = iy1 - jy2; |
| 246 | dz12 = iz1 - jz2; |
| 247 | dx20 = ix2 - jx0; |
| 248 | dy20 = iy2 - jy0; |
| 249 | dz20 = iz2 - jz0; |
| 250 | dx21 = ix2 - jx1; |
| 251 | dy21 = iy2 - jy1; |
| 252 | dz21 = iz2 - jz1; |
| 253 | dx22 = ix2 - jx2; |
| 254 | dy22 = iy2 - jy2; |
| 255 | dz22 = iz2 - jz2; |
| 256 | |
| 257 | /* Calculate squared distance and things based on it */ |
| 258 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
| 259 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
| 260 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
| 261 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
| 262 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
| 263 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
| 264 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
| 265 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
| 266 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
| 267 | |
| 268 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
| 269 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
| 270 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
| 271 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
| 272 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
| 273 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
| 274 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
| 275 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
| 276 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
| 277 | |
| 278 | rinvsq00 = rinv00*rinv00; |
| 279 | rinvsq01 = rinv01*rinv01; |
| 280 | rinvsq02 = rinv02*rinv02; |
| 281 | rinvsq10 = rinv10*rinv10; |
| 282 | rinvsq11 = rinv11*rinv11; |
| 283 | rinvsq12 = rinv12*rinv12; |
| 284 | rinvsq20 = rinv20*rinv20; |
| 285 | rinvsq21 = rinv21*rinv21; |
| 286 | rinvsq22 = rinv22*rinv22; |
| 287 | |
| 288 | /************************** |
| 289 | * CALCULATE INTERACTIONS * |
| 290 | **************************/ |
| 291 | |
| 292 | r00 = rsq00*rinv00; |
| 293 | |
| 294 | /* EWALD ELECTROSTATICS */ |
| 295 | |
| 296 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 297 | ewrt = r00*ewtabscale; |
| 298 | ewitab = ewrt; |
| 299 | eweps = ewrt-ewitab; |
| 300 | ewitab = 4*ewitab; |
| 301 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 302 | velec = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 303 | felec = qq00*rinv00*(rinvsq00-felec); |
| 304 | |
| 305 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
| 306 | ewcljrsq = ewclj2*rsq00; |
| 307 | exponent = exp(-ewcljrsq); |
| 308 | poly = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5); |
| 309 | vvdw6 = (c6_00-c6grid_00*(1.0-poly))*rinvsix; |
| 310 | vvdw12 = c12_00*rinvsix*rinvsix; |
| 311 | vvdw = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0); |
| 312 | fvdw = (vvdw12 - vvdw6 - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00; |
| 313 | |
| 314 | /* Update potential sums from outer loop */ |
| 315 | velecsum += velec; |
| 316 | vvdwsum += vvdw; |
| 317 | |
| 318 | fscal = felec+fvdw; |
| 319 | |
| 320 | /* Calculate temporary vectorial force */ |
| 321 | tx = fscal*dx00; |
| 322 | ty = fscal*dy00; |
| 323 | tz = fscal*dz00; |
| 324 | |
| 325 | /* Update vectorial force */ |
| 326 | fix0 += tx; |
| 327 | fiy0 += ty; |
| 328 | fiz0 += tz; |
| 329 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
| 330 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
| 331 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
| 332 | |
| 333 | /************************** |
| 334 | * CALCULATE INTERACTIONS * |
| 335 | **************************/ |
| 336 | |
| 337 | r01 = rsq01*rinv01; |
| 338 | |
| 339 | /* EWALD ELECTROSTATICS */ |
| 340 | |
| 341 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 342 | ewrt = r01*ewtabscale; |
| 343 | ewitab = ewrt; |
| 344 | eweps = ewrt-ewitab; |
| 345 | ewitab = 4*ewitab; |
| 346 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 347 | velec = qq01*(rinv01-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 348 | felec = qq01*rinv01*(rinvsq01-felec); |
| 349 | |
| 350 | /* Update potential sums from outer loop */ |
| 351 | velecsum += velec; |
| 352 | |
| 353 | fscal = felec; |
| 354 | |
| 355 | /* Calculate temporary vectorial force */ |
| 356 | tx = fscal*dx01; |
| 357 | ty = fscal*dy01; |
| 358 | tz = fscal*dz01; |
| 359 | |
| 360 | /* Update vectorial force */ |
| 361 | fix0 += tx; |
| 362 | fiy0 += ty; |
| 363 | fiz0 += tz; |
| 364 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
| 365 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
| 366 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
| 367 | |
| 368 | /************************** |
| 369 | * CALCULATE INTERACTIONS * |
| 370 | **************************/ |
| 371 | |
| 372 | r02 = rsq02*rinv02; |
| 373 | |
| 374 | /* EWALD ELECTROSTATICS */ |
| 375 | |
| 376 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 377 | ewrt = r02*ewtabscale; |
| 378 | ewitab = ewrt; |
| 379 | eweps = ewrt-ewitab; |
| 380 | ewitab = 4*ewitab; |
| 381 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 382 | velec = qq02*(rinv02-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 383 | felec = qq02*rinv02*(rinvsq02-felec); |
| 384 | |
| 385 | /* Update potential sums from outer loop */ |
| 386 | velecsum += velec; |
| 387 | |
| 388 | fscal = felec; |
| 389 | |
| 390 | /* Calculate temporary vectorial force */ |
| 391 | tx = fscal*dx02; |
| 392 | ty = fscal*dy02; |
| 393 | tz = fscal*dz02; |
| 394 | |
| 395 | /* Update vectorial force */ |
| 396 | fix0 += tx; |
| 397 | fiy0 += ty; |
| 398 | fiz0 += tz; |
| 399 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
| 400 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
| 401 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
| 402 | |
| 403 | /************************** |
| 404 | * CALCULATE INTERACTIONS * |
| 405 | **************************/ |
| 406 | |
| 407 | r10 = rsq10*rinv10; |
| 408 | |
| 409 | /* EWALD ELECTROSTATICS */ |
| 410 | |
| 411 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 412 | ewrt = r10*ewtabscale; |
| 413 | ewitab = ewrt; |
| 414 | eweps = ewrt-ewitab; |
| 415 | ewitab = 4*ewitab; |
| 416 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 417 | velec = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 418 | felec = qq10*rinv10*(rinvsq10-felec); |
| 419 | |
| 420 | /* Update potential sums from outer loop */ |
| 421 | velecsum += velec; |
| 422 | |
| 423 | fscal = felec; |
| 424 | |
| 425 | /* Calculate temporary vectorial force */ |
| 426 | tx = fscal*dx10; |
| 427 | ty = fscal*dy10; |
| 428 | tz = fscal*dz10; |
| 429 | |
| 430 | /* Update vectorial force */ |
| 431 | fix1 += tx; |
| 432 | fiy1 += ty; |
| 433 | fiz1 += tz; |
| 434 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
| 435 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
| 436 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
| 437 | |
| 438 | /************************** |
| 439 | * CALCULATE INTERACTIONS * |
| 440 | **************************/ |
| 441 | |
| 442 | r11 = rsq11*rinv11; |
| 443 | |
| 444 | /* EWALD ELECTROSTATICS */ |
| 445 | |
| 446 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 447 | ewrt = r11*ewtabscale; |
| 448 | ewitab = ewrt; |
| 449 | eweps = ewrt-ewitab; |
| 450 | ewitab = 4*ewitab; |
| 451 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 452 | velec = qq11*(rinv11-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 453 | felec = qq11*rinv11*(rinvsq11-felec); |
| 454 | |
| 455 | /* Update potential sums from outer loop */ |
| 456 | velecsum += velec; |
| 457 | |
| 458 | fscal = felec; |
| 459 | |
| 460 | /* Calculate temporary vectorial force */ |
| 461 | tx = fscal*dx11; |
| 462 | ty = fscal*dy11; |
| 463 | tz = fscal*dz11; |
| 464 | |
| 465 | /* Update vectorial force */ |
| 466 | fix1 += tx; |
| 467 | fiy1 += ty; |
| 468 | fiz1 += tz; |
| 469 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
| 470 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
| 471 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
| 472 | |
| 473 | /************************** |
| 474 | * CALCULATE INTERACTIONS * |
| 475 | **************************/ |
| 476 | |
| 477 | r12 = rsq12*rinv12; |
| 478 | |
| 479 | /* EWALD ELECTROSTATICS */ |
| 480 | |
| 481 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 482 | ewrt = r12*ewtabscale; |
| 483 | ewitab = ewrt; |
| 484 | eweps = ewrt-ewitab; |
| 485 | ewitab = 4*ewitab; |
| 486 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 487 | velec = qq12*(rinv12-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 488 | felec = qq12*rinv12*(rinvsq12-felec); |
| 489 | |
| 490 | /* Update potential sums from outer loop */ |
| 491 | velecsum += velec; |
| 492 | |
| 493 | fscal = felec; |
| 494 | |
| 495 | /* Calculate temporary vectorial force */ |
| 496 | tx = fscal*dx12; |
| 497 | ty = fscal*dy12; |
| 498 | tz = fscal*dz12; |
| 499 | |
| 500 | /* Update vectorial force */ |
| 501 | fix1 += tx; |
| 502 | fiy1 += ty; |
| 503 | fiz1 += tz; |
| 504 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
| 505 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
| 506 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
| 507 | |
| 508 | /************************** |
| 509 | * CALCULATE INTERACTIONS * |
| 510 | **************************/ |
| 511 | |
| 512 | r20 = rsq20*rinv20; |
| 513 | |
| 514 | /* EWALD ELECTROSTATICS */ |
| 515 | |
| 516 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 517 | ewrt = r20*ewtabscale; |
| 518 | ewitab = ewrt; |
| 519 | eweps = ewrt-ewitab; |
| 520 | ewitab = 4*ewitab; |
| 521 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 522 | velec = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 523 | felec = qq20*rinv20*(rinvsq20-felec); |
| 524 | |
| 525 | /* Update potential sums from outer loop */ |
| 526 | velecsum += velec; |
| 527 | |
| 528 | fscal = felec; |
| 529 | |
| 530 | /* Calculate temporary vectorial force */ |
| 531 | tx = fscal*dx20; |
| 532 | ty = fscal*dy20; |
| 533 | tz = fscal*dz20; |
| 534 | |
| 535 | /* Update vectorial force */ |
| 536 | fix2 += tx; |
| 537 | fiy2 += ty; |
| 538 | fiz2 += tz; |
| 539 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
| 540 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
| 541 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
| 542 | |
| 543 | /************************** |
| 544 | * CALCULATE INTERACTIONS * |
| 545 | **************************/ |
| 546 | |
| 547 | r21 = rsq21*rinv21; |
| 548 | |
| 549 | /* EWALD ELECTROSTATICS */ |
| 550 | |
| 551 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 552 | ewrt = r21*ewtabscale; |
| 553 | ewitab = ewrt; |
| 554 | eweps = ewrt-ewitab; |
| 555 | ewitab = 4*ewitab; |
| 556 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 557 | velec = qq21*(rinv21-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 558 | felec = qq21*rinv21*(rinvsq21-felec); |
| 559 | |
| 560 | /* Update potential sums from outer loop */ |
| 561 | velecsum += velec; |
| 562 | |
| 563 | fscal = felec; |
| 564 | |
| 565 | /* Calculate temporary vectorial force */ |
| 566 | tx = fscal*dx21; |
| 567 | ty = fscal*dy21; |
| 568 | tz = fscal*dz21; |
| 569 | |
| 570 | /* Update vectorial force */ |
| 571 | fix2 += tx; |
| 572 | fiy2 += ty; |
| 573 | fiz2 += tz; |
| 574 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
| 575 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
| 576 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
| 577 | |
| 578 | /************************** |
| 579 | * CALCULATE INTERACTIONS * |
| 580 | **************************/ |
| 581 | |
| 582 | r22 = rsq22*rinv22; |
| 583 | |
| 584 | /* EWALD ELECTROSTATICS */ |
| 585 | |
| 586 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 587 | ewrt = r22*ewtabscale; |
| 588 | ewitab = ewrt; |
| 589 | eweps = ewrt-ewitab; |
| 590 | ewitab = 4*ewitab; |
| 591 | felec = ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 592 | velec = qq22*(rinv22-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec))); |
| 593 | felec = qq22*rinv22*(rinvsq22-felec); |
| 594 | |
| 595 | /* Update potential sums from outer loop */ |
| 596 | velecsum += velec; |
| 597 | |
| 598 | fscal = felec; |
| 599 | |
| 600 | /* Calculate temporary vectorial force */ |
| 601 | tx = fscal*dx22; |
| 602 | ty = fscal*dy22; |
| 603 | tz = fscal*dz22; |
| 604 | |
| 605 | /* Update vectorial force */ |
| 606 | fix2 += tx; |
| 607 | fiy2 += ty; |
| 608 | fiz2 += tz; |
| 609 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
| 610 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
| 611 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
| 612 | |
| 613 | /* Inner loop uses 386 flops */ |
| 614 | } |
| 615 | /* End of innermost loop */ |
| 616 | |
| 617 | tx = ty = tz = 0; |
| 618 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
| 619 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
| 620 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
| 621 | tx += fix0; |
| 622 | ty += fiy0; |
| 623 | tz += fiz0; |
| 624 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
| 625 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
| 626 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
| 627 | tx += fix1; |
| 628 | ty += fiy1; |
| 629 | tz += fiz1; |
| 630 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
| 631 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
| 632 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
| 633 | tx += fix2; |
| 634 | ty += fiy2; |
| 635 | tz += fiz2; |
| 636 | fshift[i_shift_offset+XX0] += tx; |
| 637 | fshift[i_shift_offset+YY1] += ty; |
| 638 | fshift[i_shift_offset+ZZ2] += tz; |
| 639 | |
| 640 | ggid = gid[iidx]; |
| 641 | /* Update potential energies */ |
| 642 | kernel_data->energygrp_elec[ggid] += velecsum; |
| 643 | kernel_data->energygrp_vdw[ggid] += vvdwsum; |
| 644 | |
| 645 | /* Increment number of inner iterations */ |
| 646 | inneriter += j_index_end - j_index_start; |
| 647 | |
| 648 | /* Outer loop uses 32 flops */ |
| 649 | } |
| 650 | |
| 651 | /* Increment number of outer iterations */ |
| 652 | outeriter += nri; |
| 653 | |
| 654 | /* Update outer/inner flops */ |
| 655 | |
| 656 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*32 + inneriter*386)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_VF] += outeriter*32 + inneriter*386; |
| 657 | } |
| 658 | /* |
| 659 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJEw_GeomW3W3_F_c |
| 660 | * Electrostatics interaction: Ewald |
| 661 | * VdW interaction: LJEwald |
| 662 | * Geometry: Water3-Water3 |
| 663 | * Calculate force/pot: Force |
| 664 | */ |
| 665 | void |
| 666 | nb_kernel_ElecEw_VdwLJEw_GeomW3W3_F_c |
| 667 | (t_nblist * gmx_restrict__restrict nlist, |
| 668 | rvec * gmx_restrict__restrict xx, |
| 669 | rvec * gmx_restrict__restrict ff, |
| 670 | t_forcerec * gmx_restrict__restrict fr, |
| 671 | t_mdatoms * gmx_restrict__restrict mdatoms, |
| 672 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict__restrict kernel_data, |
| 673 | t_nrnb * gmx_restrict__restrict nrnb) |
| 674 | { |
| 675 | int i_shift_offset,i_coord_offset,j_coord_offset; |
| 676 | int j_index_start,j_index_end; |
| 677 | int nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter; |
| 678 | real shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2; |
| 679 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
| 680 | real *shiftvec,*fshift,*x,*f; |
| 681 | int vdwioffset0; |
| 682 | real ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
| 683 | int vdwioffset1; |
| 684 | real ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
| 685 | int vdwioffset2; |
| 686 | real ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
| 687 | int vdwjidx0; |
| 688 | real jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
| 689 | int vdwjidx1; |
| 690 | real jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
| 691 | int vdwjidx2; |
| 692 | real jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
| 693 | real dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00; |
| 694 | real dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01; |
| 695 | real dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02; |
| 696 | real dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10; |
| 697 | real dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11; |
| 698 | real dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12; |
| 699 | real dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20; |
| 700 | real dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21; |
| 701 | real dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22; |
| 702 | real velec,felec,velecsum,facel,crf,krf,krf2; |
| 703 | real *charge; |
| 704 | int nvdwtype; |
| 705 | real rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6; |
| 706 | int *vdwtype; |
| 707 | real *vdwparam; |
| 708 | real c6grid_00; |
| 709 | real c6grid_01; |
| 710 | real c6grid_02; |
| 711 | real c6grid_10; |
| 712 | real c6grid_11; |
| 713 | real c6grid_12; |
| 714 | real c6grid_20; |
| 715 | real c6grid_21; |
| 716 | real c6grid_22; |
| 717 | real ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald; |
| 718 | real *vdwgridparam; |
| 719 | int ewitab; |
| 720 | real ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace; |
| 721 | real *ewtab; |
| 722 | |
| 723 | x = xx[0]; |
| 724 | f = ff[0]; |
| 725 | |
| 726 | nri = nlist->nri; |
| 727 | iinr = nlist->iinr; |
| 728 | jindex = nlist->jindex; |
| 729 | jjnr = nlist->jjnr; |
| 730 | shiftidx = nlist->shift; |
| 731 | gid = nlist->gid; |
| 732 | shiftvec = fr->shift_vec[0]; |
| 733 | fshift = fr->fshift[0]; |
| 734 | facel = fr->epsfac; |
| 735 | charge = mdatoms->chargeA; |
| 736 | nvdwtype = fr->ntype; |
| 737 | vdwparam = fr->nbfp; |
| 738 | vdwtype = mdatoms->typeA; |
| 739 | vdwgridparam = fr->ljpme_c6grid; |
| 740 | ewclj = fr->ewaldcoeff_lj; |
| 741 | sh_lj_ewald = fr->ic->sh_lj_ewald; |
| 742 | ewclj2 = ewclj*ewclj; |
| 743 | ewclj6 = ewclj2*ewclj2*ewclj2; |
| 744 | |
| 745 | sh_ewald = fr->ic->sh_ewald; |
| 746 | ewtab = fr->ic->tabq_coul_F; |
| 747 | ewtabscale = fr->ic->tabq_scale; |
| 748 | ewtabhalfspace = 0.5/ewtabscale; |
| 749 | |
| 750 | /* Setup water-specific parameters */ |
| 751 | inr = nlist->iinr[0]; |
| 752 | iq0 = facel*charge[inr+0]; |
| 753 | iq1 = facel*charge[inr+1]; |
| 754 | iq2 = facel*charge[inr+2]; |
| 755 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
| 756 | |
| 757 | jq0 = charge[inr+0]; |
| 758 | jq1 = charge[inr+1]; |
| 759 | jq2 = charge[inr+2]; |
| 760 | vdwjidx0 = 2*vdwtype[inr+0]; |
| 761 | qq00 = iq0*jq0; |
| 762 | c6_00 = vdwparam[vdwioffset0+vdwjidx0]; |
| 763 | c12_00 = vdwparam[vdwioffset0+vdwjidx0+1]; |
| 764 | c6grid_00 = vdwgridparam[vdwioffset0+vdwjidx0]; |
| 765 | qq01 = iq0*jq1; |
| 766 | qq02 = iq0*jq2; |
| 767 | qq10 = iq1*jq0; |
| 768 | qq11 = iq1*jq1; |
| 769 | qq12 = iq1*jq2; |
| 770 | qq20 = iq2*jq0; |
| 771 | qq21 = iq2*jq1; |
| 772 | qq22 = iq2*jq2; |
| 773 | |
| 774 | outeriter = 0; |
| 775 | inneriter = 0; |
| 776 | |
| 777 | /* Start outer loop over neighborlists */ |
| 778 | for(iidx=0; iidx<nri; iidx++) |
| 779 | { |
| 780 | /* Load shift vector for this list */ |
| 781 | i_shift_offset = DIM3*shiftidx[iidx]; |
| 782 | shX = shiftvec[i_shift_offset+XX0]; |
| 783 | shY = shiftvec[i_shift_offset+YY1]; |
| 784 | shZ = shiftvec[i_shift_offset+ZZ2]; |
| 785 | |
| 786 | /* Load limits for loop over neighbors */ |
| 787 | j_index_start = jindex[iidx]; |
| 788 | j_index_end = jindex[iidx+1]; |
| 789 | |
| 790 | /* Get outer coordinate index */ |
| 791 | inr = iinr[iidx]; |
| 792 | i_coord_offset = DIM3*inr; |
| 793 | |
| 794 | /* Load i particle coords and add shift vector */ |
| 795 | ix0 = shX + x[i_coord_offset+DIM3*0+XX0]; |
| 796 | iy0 = shY + x[i_coord_offset+DIM3*0+YY1]; |
| 797 | iz0 = shZ + x[i_coord_offset+DIM3*0+ZZ2]; |
| 798 | ix1 = shX + x[i_coord_offset+DIM3*1+XX0]; |
| 799 | iy1 = shY + x[i_coord_offset+DIM3*1+YY1]; |
| 800 | iz1 = shZ + x[i_coord_offset+DIM3*1+ZZ2]; |
| 801 | ix2 = shX + x[i_coord_offset+DIM3*2+XX0]; |
| 802 | iy2 = shY + x[i_coord_offset+DIM3*2+YY1]; |
| 803 | iz2 = shZ + x[i_coord_offset+DIM3*2+ZZ2]; |
| 804 | |
| 805 | fix0 = 0.0; |
| 806 | fiy0 = 0.0; |
| 807 | fiz0 = 0.0; |
| 808 | fix1 = 0.0; |
| 809 | fiy1 = 0.0; |
| 810 | fiz1 = 0.0; |
| 811 | fix2 = 0.0; |
| 812 | fiy2 = 0.0; |
| 813 | fiz2 = 0.0; |
| 814 | |
| 815 | /* Start inner kernel loop */ |
| 816 | for(jidx=j_index_start; jidx<j_index_end; jidx++) |
| 817 | { |
| 818 | /* Get j neighbor index, and coordinate index */ |
| 819 | jnr = jjnr[jidx]; |
| 820 | j_coord_offset = DIM3*jnr; |
| 821 | |
| 822 | /* load j atom coordinates */ |
| 823 | jx0 = x[j_coord_offset+DIM3*0+XX0]; |
| 824 | jy0 = x[j_coord_offset+DIM3*0+YY1]; |
| 825 | jz0 = x[j_coord_offset+DIM3*0+ZZ2]; |
| 826 | jx1 = x[j_coord_offset+DIM3*1+XX0]; |
| 827 | jy1 = x[j_coord_offset+DIM3*1+YY1]; |
| 828 | jz1 = x[j_coord_offset+DIM3*1+ZZ2]; |
| 829 | jx2 = x[j_coord_offset+DIM3*2+XX0]; |
| 830 | jy2 = x[j_coord_offset+DIM3*2+YY1]; |
| 831 | jz2 = x[j_coord_offset+DIM3*2+ZZ2]; |
| 832 | |
| 833 | /* Calculate displacement vector */ |
| 834 | dx00 = ix0 - jx0; |
| 835 | dy00 = iy0 - jy0; |
| 836 | dz00 = iz0 - jz0; |
| 837 | dx01 = ix0 - jx1; |
| 838 | dy01 = iy0 - jy1; |
| 839 | dz01 = iz0 - jz1; |
| 840 | dx02 = ix0 - jx2; |
| 841 | dy02 = iy0 - jy2; |
| 842 | dz02 = iz0 - jz2; |
| 843 | dx10 = ix1 - jx0; |
| 844 | dy10 = iy1 - jy0; |
| 845 | dz10 = iz1 - jz0; |
| 846 | dx11 = ix1 - jx1; |
| 847 | dy11 = iy1 - jy1; |
| 848 | dz11 = iz1 - jz1; |
| 849 | dx12 = ix1 - jx2; |
| 850 | dy12 = iy1 - jy2; |
| 851 | dz12 = iz1 - jz2; |
| 852 | dx20 = ix2 - jx0; |
| 853 | dy20 = iy2 - jy0; |
| 854 | dz20 = iz2 - jz0; |
| 855 | dx21 = ix2 - jx1; |
| 856 | dy21 = iy2 - jy1; |
| 857 | dz21 = iz2 - jz1; |
| 858 | dx22 = ix2 - jx2; |
| 859 | dy22 = iy2 - jy2; |
| 860 | dz22 = iz2 - jz2; |
| 861 | |
| 862 | /* Calculate squared distance and things based on it */ |
| 863 | rsq00 = dx00*dx00+dy00*dy00+dz00*dz00; |
| 864 | rsq01 = dx01*dx01+dy01*dy01+dz01*dz01; |
| 865 | rsq02 = dx02*dx02+dy02*dy02+dz02*dz02; |
| 866 | rsq10 = dx10*dx10+dy10*dy10+dz10*dz10; |
| 867 | rsq11 = dx11*dx11+dy11*dy11+dz11*dz11; |
| 868 | rsq12 = dx12*dx12+dy12*dy12+dz12*dz12; |
| 869 | rsq20 = dx20*dx20+dy20*dy20+dz20*dz20; |
| 870 | rsq21 = dx21*dx21+dy21*dy21+dz21*dz21; |
| 871 | rsq22 = dx22*dx22+dy22*dy22+dz22*dz22; |
| 872 | |
| 873 | rinv00 = gmx_invsqrt(rsq00)gmx_software_invsqrt(rsq00); |
| 874 | rinv01 = gmx_invsqrt(rsq01)gmx_software_invsqrt(rsq01); |
| 875 | rinv02 = gmx_invsqrt(rsq02)gmx_software_invsqrt(rsq02); |
| 876 | rinv10 = gmx_invsqrt(rsq10)gmx_software_invsqrt(rsq10); |
| 877 | rinv11 = gmx_invsqrt(rsq11)gmx_software_invsqrt(rsq11); |
| 878 | rinv12 = gmx_invsqrt(rsq12)gmx_software_invsqrt(rsq12); |
| 879 | rinv20 = gmx_invsqrt(rsq20)gmx_software_invsqrt(rsq20); |
| 880 | rinv21 = gmx_invsqrt(rsq21)gmx_software_invsqrt(rsq21); |
| 881 | rinv22 = gmx_invsqrt(rsq22)gmx_software_invsqrt(rsq22); |
| 882 | |
| 883 | rinvsq00 = rinv00*rinv00; |
| 884 | rinvsq01 = rinv01*rinv01; |
| 885 | rinvsq02 = rinv02*rinv02; |
| 886 | rinvsq10 = rinv10*rinv10; |
| 887 | rinvsq11 = rinv11*rinv11; |
| 888 | rinvsq12 = rinv12*rinv12; |
| 889 | rinvsq20 = rinv20*rinv20; |
| 890 | rinvsq21 = rinv21*rinv21; |
| 891 | rinvsq22 = rinv22*rinv22; |
| 892 | |
| 893 | /************************** |
| 894 | * CALCULATE INTERACTIONS * |
| 895 | **************************/ |
| 896 | |
| 897 | r00 = rsq00*rinv00; |
| 898 | |
| 899 | /* EWALD ELECTROSTATICS */ |
| 900 | |
| 901 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 902 | ewrt = r00*ewtabscale; |
| 903 | ewitab = ewrt; |
| 904 | eweps = ewrt-ewitab; |
| 905 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 906 | felec = qq00*rinv00*(rinvsq00-felec); |
| 907 | |
| 908 | rinvsix = rinvsq00*rinvsq00*rinvsq00; |
| 909 | ewcljrsq = ewclj2*rsq00; |
| 910 | exponent = exp(-ewcljrsq); |
| 911 | poly = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5); |
| 912 | fvdw = (((c12_00*rinvsix - c6_00 + c6grid_00*(1.0-poly))*rinvsix) - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00; |
| 913 | |
| 914 | fscal = felec+fvdw; |
| 915 | |
| 916 | /* Calculate temporary vectorial force */ |
| 917 | tx = fscal*dx00; |
| 918 | ty = fscal*dy00; |
| 919 | tz = fscal*dz00; |
| 920 | |
| 921 | /* Update vectorial force */ |
| 922 | fix0 += tx; |
| 923 | fiy0 += ty; |
| 924 | fiz0 += tz; |
| 925 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
| 926 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
| 927 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
| 928 | |
| 929 | /************************** |
| 930 | * CALCULATE INTERACTIONS * |
| 931 | **************************/ |
| 932 | |
| 933 | r01 = rsq01*rinv01; |
| 934 | |
| 935 | /* EWALD ELECTROSTATICS */ |
| 936 | |
| 937 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 938 | ewrt = r01*ewtabscale; |
| 939 | ewitab = ewrt; |
| 940 | eweps = ewrt-ewitab; |
| 941 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 942 | felec = qq01*rinv01*(rinvsq01-felec); |
| 943 | |
| 944 | fscal = felec; |
| 945 | |
| 946 | /* Calculate temporary vectorial force */ |
| 947 | tx = fscal*dx01; |
| 948 | ty = fscal*dy01; |
| 949 | tz = fscal*dz01; |
| 950 | |
| 951 | /* Update vectorial force */ |
| 952 | fix0 += tx; |
| 953 | fiy0 += ty; |
| 954 | fiz0 += tz; |
| 955 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
| 956 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
| 957 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
| 958 | |
| 959 | /************************** |
| 960 | * CALCULATE INTERACTIONS * |
| 961 | **************************/ |
| 962 | |
| 963 | r02 = rsq02*rinv02; |
| 964 | |
| 965 | /* EWALD ELECTROSTATICS */ |
| 966 | |
| 967 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 968 | ewrt = r02*ewtabscale; |
| 969 | ewitab = ewrt; |
| 970 | eweps = ewrt-ewitab; |
| 971 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 972 | felec = qq02*rinv02*(rinvsq02-felec); |
| 973 | |
| 974 | fscal = felec; |
| 975 | |
| 976 | /* Calculate temporary vectorial force */ |
| 977 | tx = fscal*dx02; |
| 978 | ty = fscal*dy02; |
| 979 | tz = fscal*dz02; |
| 980 | |
| 981 | /* Update vectorial force */ |
| 982 | fix0 += tx; |
| 983 | fiy0 += ty; |
| 984 | fiz0 += tz; |
| 985 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
| 986 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
| 987 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
| 988 | |
| 989 | /************************** |
| 990 | * CALCULATE INTERACTIONS * |
| 991 | **************************/ |
| 992 | |
| 993 | r10 = rsq10*rinv10; |
| 994 | |
| 995 | /* EWALD ELECTROSTATICS */ |
| 996 | |
| 997 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 998 | ewrt = r10*ewtabscale; |
| 999 | ewitab = ewrt; |
| 1000 | eweps = ewrt-ewitab; |
| 1001 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 1002 | felec = qq10*rinv10*(rinvsq10-felec); |
| 1003 | |
| 1004 | fscal = felec; |
| 1005 | |
| 1006 | /* Calculate temporary vectorial force */ |
| 1007 | tx = fscal*dx10; |
| 1008 | ty = fscal*dy10; |
| 1009 | tz = fscal*dz10; |
| 1010 | |
| 1011 | /* Update vectorial force */ |
| 1012 | fix1 += tx; |
| 1013 | fiy1 += ty; |
| 1014 | fiz1 += tz; |
| 1015 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
| 1016 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
| 1017 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
| 1018 | |
| 1019 | /************************** |
| 1020 | * CALCULATE INTERACTIONS * |
| 1021 | **************************/ |
| 1022 | |
| 1023 | r11 = rsq11*rinv11; |
| 1024 | |
| 1025 | /* EWALD ELECTROSTATICS */ |
| 1026 | |
| 1027 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 1028 | ewrt = r11*ewtabscale; |
| 1029 | ewitab = ewrt; |
| 1030 | eweps = ewrt-ewitab; |
| 1031 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 1032 | felec = qq11*rinv11*(rinvsq11-felec); |
| 1033 | |
| 1034 | fscal = felec; |
| 1035 | |
| 1036 | /* Calculate temporary vectorial force */ |
| 1037 | tx = fscal*dx11; |
| 1038 | ty = fscal*dy11; |
| 1039 | tz = fscal*dz11; |
| 1040 | |
| 1041 | /* Update vectorial force */ |
| 1042 | fix1 += tx; |
| 1043 | fiy1 += ty; |
| 1044 | fiz1 += tz; |
| 1045 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
| 1046 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
| 1047 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
| 1048 | |
| 1049 | /************************** |
| 1050 | * CALCULATE INTERACTIONS * |
| 1051 | **************************/ |
| 1052 | |
| 1053 | r12 = rsq12*rinv12; |
| 1054 | |
| 1055 | /* EWALD ELECTROSTATICS */ |
| 1056 | |
| 1057 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 1058 | ewrt = r12*ewtabscale; |
| 1059 | ewitab = ewrt; |
| 1060 | eweps = ewrt-ewitab; |
| 1061 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 1062 | felec = qq12*rinv12*(rinvsq12-felec); |
| 1063 | |
| 1064 | fscal = felec; |
| 1065 | |
| 1066 | /* Calculate temporary vectorial force */ |
| 1067 | tx = fscal*dx12; |
| 1068 | ty = fscal*dy12; |
| 1069 | tz = fscal*dz12; |
| 1070 | |
| 1071 | /* Update vectorial force */ |
| 1072 | fix1 += tx; |
| 1073 | fiy1 += ty; |
| 1074 | fiz1 += tz; |
| 1075 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
| 1076 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
| 1077 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
| 1078 | |
| 1079 | /************************** |
| 1080 | * CALCULATE INTERACTIONS * |
| 1081 | **************************/ |
| 1082 | |
| 1083 | r20 = rsq20*rinv20; |
| 1084 | |
| 1085 | /* EWALD ELECTROSTATICS */ |
| 1086 | |
| 1087 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 1088 | ewrt = r20*ewtabscale; |
| 1089 | ewitab = ewrt; |
| 1090 | eweps = ewrt-ewitab; |
| 1091 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 1092 | felec = qq20*rinv20*(rinvsq20-felec); |
| 1093 | |
| 1094 | fscal = felec; |
| 1095 | |
| 1096 | /* Calculate temporary vectorial force */ |
| 1097 | tx = fscal*dx20; |
| 1098 | ty = fscal*dy20; |
| 1099 | tz = fscal*dz20; |
| 1100 | |
| 1101 | /* Update vectorial force */ |
| 1102 | fix2 += tx; |
| 1103 | fiy2 += ty; |
| 1104 | fiz2 += tz; |
| 1105 | f[j_coord_offset+DIM3*0+XX0] -= tx; |
| 1106 | f[j_coord_offset+DIM3*0+YY1] -= ty; |
| 1107 | f[j_coord_offset+DIM3*0+ZZ2] -= tz; |
| 1108 | |
| 1109 | /************************** |
| 1110 | * CALCULATE INTERACTIONS * |
| 1111 | **************************/ |
| 1112 | |
| 1113 | r21 = rsq21*rinv21; |
| 1114 | |
| 1115 | /* EWALD ELECTROSTATICS */ |
| 1116 | |
| 1117 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 1118 | ewrt = r21*ewtabscale; |
| 1119 | ewitab = ewrt; |
| 1120 | eweps = ewrt-ewitab; |
| 1121 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 1122 | felec = qq21*rinv21*(rinvsq21-felec); |
| 1123 | |
| 1124 | fscal = felec; |
| 1125 | |
| 1126 | /* Calculate temporary vectorial force */ |
| 1127 | tx = fscal*dx21; |
| 1128 | ty = fscal*dy21; |
| 1129 | tz = fscal*dz21; |
| 1130 | |
| 1131 | /* Update vectorial force */ |
| 1132 | fix2 += tx; |
| 1133 | fiy2 += ty; |
| 1134 | fiz2 += tz; |
| 1135 | f[j_coord_offset+DIM3*1+XX0] -= tx; |
| 1136 | f[j_coord_offset+DIM3*1+YY1] -= ty; |
| 1137 | f[j_coord_offset+DIM3*1+ZZ2] -= tz; |
| 1138 | |
| 1139 | /************************** |
| 1140 | * CALCULATE INTERACTIONS * |
| 1141 | **************************/ |
| 1142 | |
| 1143 | r22 = rsq22*rinv22; |
| 1144 | |
| 1145 | /* EWALD ELECTROSTATICS */ |
| 1146 | |
| 1147 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
| 1148 | ewrt = r22*ewtabscale; |
| 1149 | ewitab = ewrt; |
| 1150 | eweps = ewrt-ewitab; |
| 1151 | felec = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1]; |
| 1152 | felec = qq22*rinv22*(rinvsq22-felec); |
| 1153 | |
| 1154 | fscal = felec; |
| 1155 | |
| 1156 | /* Calculate temporary vectorial force */ |
| 1157 | tx = fscal*dx22; |
| 1158 | ty = fscal*dy22; |
| 1159 | tz = fscal*dz22; |
| 1160 | |
| 1161 | /* Update vectorial force */ |
| 1162 | fix2 += tx; |
| 1163 | fiy2 += ty; |
| 1164 | fiz2 += tz; |
| 1165 | f[j_coord_offset+DIM3*2+XX0] -= tx; |
| 1166 | f[j_coord_offset+DIM3*2+YY1] -= ty; |
| 1167 | f[j_coord_offset+DIM3*2+ZZ2] -= tz; |
| 1168 | |
| 1169 | /* Inner loop uses 318 flops */ |
| 1170 | } |
| 1171 | /* End of innermost loop */ |
| 1172 | |
| 1173 | tx = ty = tz = 0; |
| 1174 | f[i_coord_offset+DIM3*0+XX0] += fix0; |
| 1175 | f[i_coord_offset+DIM3*0+YY1] += fiy0; |
| 1176 | f[i_coord_offset+DIM3*0+ZZ2] += fiz0; |
| 1177 | tx += fix0; |
| 1178 | ty += fiy0; |
| 1179 | tz += fiz0; |
| 1180 | f[i_coord_offset+DIM3*1+XX0] += fix1; |
| 1181 | f[i_coord_offset+DIM3*1+YY1] += fiy1; |
| 1182 | f[i_coord_offset+DIM3*1+ZZ2] += fiz1; |
| 1183 | tx += fix1; |
| 1184 | ty += fiy1; |
| 1185 | tz += fiz1; |
| 1186 | f[i_coord_offset+DIM3*2+XX0] += fix2; |
| 1187 | f[i_coord_offset+DIM3*2+YY1] += fiy2; |
| 1188 | f[i_coord_offset+DIM3*2+ZZ2] += fiz2; |
| 1189 | tx += fix2; |
| 1190 | ty += fiy2; |
| 1191 | tz += fiz2; |
| 1192 | fshift[i_shift_offset+XX0] += tx; |
| 1193 | fshift[i_shift_offset+YY1] += ty; |
| 1194 | fshift[i_shift_offset+ZZ2] += tz; |
| 1195 | |
| 1196 | /* Increment number of inner iterations */ |
| 1197 | inneriter += j_index_end - j_index_start; |
| 1198 | |
| 1199 | /* Outer loop uses 30 flops */ |
| 1200 | } |
| 1201 | |
| 1202 | /* Increment number of outer iterations */ |
| 1203 | outeriter += nri; |
| 1204 | |
| 1205 | /* Update outer/inner flops */ |
| 1206 | |
| 1207 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*30 + inneriter*318)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_F] += outeriter*30 + inneriter *318; |
| 1208 | } |