2 * This file is part of the GROMACS molecular simulation package.
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.
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.
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.
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.
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.
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.
36 * Note: this file was generated by the GROMACS sse2_single kernel generator.
42 #include "../nb_kernel.h"
43 #include "types/simple.h"
44 #include "gromacs/math/vec.h"
47 #include "gromacs/simd/math_x86_sse2_single.h"
48 #include "kernelutil_x86_sse2_single.h"
51 * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwNone_GeomW4W4_VF_sse2_single
52 * Electrostatics interaction: Ewald
53 * VdW interaction: None
54 * Geometry: Water4-Water4
55 * Calculate force/pot: PotentialAndForce
58 nb_kernel_ElecEw_VdwNone_GeomW4W4_VF_sse2_single
59 (t_nblist * gmx_restrict nlist,
60 rvec * gmx_restrict xx,
61 rvec * gmx_restrict ff,
62 t_forcerec * gmx_restrict fr,
63 t_mdatoms * gmx_restrict mdatoms,
64 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
65 t_nrnb * gmx_restrict nrnb)
67 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
68 * just 0 for non-waters.
69 * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
70 * jnr indices corresponding to data put in the four positions in the SIMD register.
72 int i_shift_offset,i_coord_offset,outeriter,inneriter;
73 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
74 int jnrA,jnrB,jnrC,jnrD;
75 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
76 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
77 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
79 real *shiftvec,*fshift,*x,*f;
80 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
82 __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
84 __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
86 __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
88 __m128 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
89 int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
90 __m128 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
91 int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
92 __m128 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
93 int vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
94 __m128 jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
95 __m128 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
96 __m128 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
97 __m128 dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
98 __m128 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
99 __m128 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
100 __m128 dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
101 __m128 dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
102 __m128 dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
103 __m128 dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
104 __m128 velec,felec,velecsum,facel,crf,krf,krf2;
107 __m128 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
109 __m128 dummy_mask,cutoff_mask;
110 __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
111 __m128 one = _mm_set1_ps(1.0);
112 __m128 two = _mm_set1_ps(2.0);
118 jindex = nlist->jindex;
120 shiftidx = nlist->shift;
122 shiftvec = fr->shift_vec[0];
123 fshift = fr->fshift[0];
124 facel = _mm_set1_ps(fr->epsfac);
125 charge = mdatoms->chargeA;
127 sh_ewald = _mm_set1_ps(fr->ic->sh_ewald);
128 ewtab = fr->ic->tabq_coul_FDV0;
129 ewtabscale = _mm_set1_ps(fr->ic->tabq_scale);
130 ewtabhalfspace = _mm_set1_ps(0.5/fr->ic->tabq_scale);
132 /* Setup water-specific parameters */
133 inr = nlist->iinr[0];
134 iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
135 iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
136 iq3 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
138 jq1 = _mm_set1_ps(charge[inr+1]);
139 jq2 = _mm_set1_ps(charge[inr+2]);
140 jq3 = _mm_set1_ps(charge[inr+3]);
141 qq11 = _mm_mul_ps(iq1,jq1);
142 qq12 = _mm_mul_ps(iq1,jq2);
143 qq13 = _mm_mul_ps(iq1,jq3);
144 qq21 = _mm_mul_ps(iq2,jq1);
145 qq22 = _mm_mul_ps(iq2,jq2);
146 qq23 = _mm_mul_ps(iq2,jq3);
147 qq31 = _mm_mul_ps(iq3,jq1);
148 qq32 = _mm_mul_ps(iq3,jq2);
149 qq33 = _mm_mul_ps(iq3,jq3);
151 /* Avoid stupid compiler warnings */
152 jnrA = jnrB = jnrC = jnrD = 0;
161 for(iidx=0;iidx<4*DIM;iidx++)
166 /* Start outer loop over neighborlists */
167 for(iidx=0; iidx<nri; iidx++)
169 /* Load shift vector for this list */
170 i_shift_offset = DIM*shiftidx[iidx];
172 /* Load limits for loop over neighbors */
173 j_index_start = jindex[iidx];
174 j_index_end = jindex[iidx+1];
176 /* Get outer coordinate index */
178 i_coord_offset = DIM*inr;
180 /* Load i particle coords and add shift vector */
181 gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
182 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
184 fix1 = _mm_setzero_ps();
185 fiy1 = _mm_setzero_ps();
186 fiz1 = _mm_setzero_ps();
187 fix2 = _mm_setzero_ps();
188 fiy2 = _mm_setzero_ps();
189 fiz2 = _mm_setzero_ps();
190 fix3 = _mm_setzero_ps();
191 fiy3 = _mm_setzero_ps();
192 fiz3 = _mm_setzero_ps();
194 /* Reset potential sums */
195 velecsum = _mm_setzero_ps();
197 /* Start inner kernel loop */
198 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
201 /* Get j neighbor index, and coordinate index */
206 j_coord_offsetA = DIM*jnrA;
207 j_coord_offsetB = DIM*jnrB;
208 j_coord_offsetC = DIM*jnrC;
209 j_coord_offsetD = DIM*jnrD;
211 /* load j atom coordinates */
212 gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
213 x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
214 &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
216 /* Calculate displacement vector */
217 dx11 = _mm_sub_ps(ix1,jx1);
218 dy11 = _mm_sub_ps(iy1,jy1);
219 dz11 = _mm_sub_ps(iz1,jz1);
220 dx12 = _mm_sub_ps(ix1,jx2);
221 dy12 = _mm_sub_ps(iy1,jy2);
222 dz12 = _mm_sub_ps(iz1,jz2);
223 dx13 = _mm_sub_ps(ix1,jx3);
224 dy13 = _mm_sub_ps(iy1,jy3);
225 dz13 = _mm_sub_ps(iz1,jz3);
226 dx21 = _mm_sub_ps(ix2,jx1);
227 dy21 = _mm_sub_ps(iy2,jy1);
228 dz21 = _mm_sub_ps(iz2,jz1);
229 dx22 = _mm_sub_ps(ix2,jx2);
230 dy22 = _mm_sub_ps(iy2,jy2);
231 dz22 = _mm_sub_ps(iz2,jz2);
232 dx23 = _mm_sub_ps(ix2,jx3);
233 dy23 = _mm_sub_ps(iy2,jy3);
234 dz23 = _mm_sub_ps(iz2,jz3);
235 dx31 = _mm_sub_ps(ix3,jx1);
236 dy31 = _mm_sub_ps(iy3,jy1);
237 dz31 = _mm_sub_ps(iz3,jz1);
238 dx32 = _mm_sub_ps(ix3,jx2);
239 dy32 = _mm_sub_ps(iy3,jy2);
240 dz32 = _mm_sub_ps(iz3,jz2);
241 dx33 = _mm_sub_ps(ix3,jx3);
242 dy33 = _mm_sub_ps(iy3,jy3);
243 dz33 = _mm_sub_ps(iz3,jz3);
245 /* Calculate squared distance and things based on it */
246 rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
247 rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
248 rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
249 rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
250 rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
251 rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
252 rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
253 rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
254 rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
256 rinv11 = gmx_mm_invsqrt_ps(rsq11);
257 rinv12 = gmx_mm_invsqrt_ps(rsq12);
258 rinv13 = gmx_mm_invsqrt_ps(rsq13);
259 rinv21 = gmx_mm_invsqrt_ps(rsq21);
260 rinv22 = gmx_mm_invsqrt_ps(rsq22);
261 rinv23 = gmx_mm_invsqrt_ps(rsq23);
262 rinv31 = gmx_mm_invsqrt_ps(rsq31);
263 rinv32 = gmx_mm_invsqrt_ps(rsq32);
264 rinv33 = gmx_mm_invsqrt_ps(rsq33);
266 rinvsq11 = _mm_mul_ps(rinv11,rinv11);
267 rinvsq12 = _mm_mul_ps(rinv12,rinv12);
268 rinvsq13 = _mm_mul_ps(rinv13,rinv13);
269 rinvsq21 = _mm_mul_ps(rinv21,rinv21);
270 rinvsq22 = _mm_mul_ps(rinv22,rinv22);
271 rinvsq23 = _mm_mul_ps(rinv23,rinv23);
272 rinvsq31 = _mm_mul_ps(rinv31,rinv31);
273 rinvsq32 = _mm_mul_ps(rinv32,rinv32);
274 rinvsq33 = _mm_mul_ps(rinv33,rinv33);
276 fjx1 = _mm_setzero_ps();
277 fjy1 = _mm_setzero_ps();
278 fjz1 = _mm_setzero_ps();
279 fjx2 = _mm_setzero_ps();
280 fjy2 = _mm_setzero_ps();
281 fjz2 = _mm_setzero_ps();
282 fjx3 = _mm_setzero_ps();
283 fjy3 = _mm_setzero_ps();
284 fjz3 = _mm_setzero_ps();
286 /**************************
287 * CALCULATE INTERACTIONS *
288 **************************/
290 r11 = _mm_mul_ps(rsq11,rinv11);
292 /* EWALD ELECTROSTATICS */
294 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
295 ewrt = _mm_mul_ps(r11,ewtabscale);
296 ewitab = _mm_cvttps_epi32(ewrt);
297 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
298 ewitab = _mm_slli_epi32(ewitab,2);
299 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
300 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
301 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
302 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
303 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
304 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
305 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
306 velec = _mm_mul_ps(qq11,_mm_sub_ps(rinv11,velec));
307 felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec));
309 /* Update potential sum for this i atom from the interaction with this j atom. */
310 velecsum = _mm_add_ps(velecsum,velec);
314 /* Calculate temporary vectorial force */
315 tx = _mm_mul_ps(fscal,dx11);
316 ty = _mm_mul_ps(fscal,dy11);
317 tz = _mm_mul_ps(fscal,dz11);
319 /* Update vectorial force */
320 fix1 = _mm_add_ps(fix1,tx);
321 fiy1 = _mm_add_ps(fiy1,ty);
322 fiz1 = _mm_add_ps(fiz1,tz);
324 fjx1 = _mm_add_ps(fjx1,tx);
325 fjy1 = _mm_add_ps(fjy1,ty);
326 fjz1 = _mm_add_ps(fjz1,tz);
328 /**************************
329 * CALCULATE INTERACTIONS *
330 **************************/
332 r12 = _mm_mul_ps(rsq12,rinv12);
334 /* EWALD ELECTROSTATICS */
336 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
337 ewrt = _mm_mul_ps(r12,ewtabscale);
338 ewitab = _mm_cvttps_epi32(ewrt);
339 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
340 ewitab = _mm_slli_epi32(ewitab,2);
341 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
342 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
343 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
344 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
345 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
346 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
347 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
348 velec = _mm_mul_ps(qq12,_mm_sub_ps(rinv12,velec));
349 felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec));
351 /* Update potential sum for this i atom from the interaction with this j atom. */
352 velecsum = _mm_add_ps(velecsum,velec);
356 /* Calculate temporary vectorial force */
357 tx = _mm_mul_ps(fscal,dx12);
358 ty = _mm_mul_ps(fscal,dy12);
359 tz = _mm_mul_ps(fscal,dz12);
361 /* Update vectorial force */
362 fix1 = _mm_add_ps(fix1,tx);
363 fiy1 = _mm_add_ps(fiy1,ty);
364 fiz1 = _mm_add_ps(fiz1,tz);
366 fjx2 = _mm_add_ps(fjx2,tx);
367 fjy2 = _mm_add_ps(fjy2,ty);
368 fjz2 = _mm_add_ps(fjz2,tz);
370 /**************************
371 * CALCULATE INTERACTIONS *
372 **************************/
374 r13 = _mm_mul_ps(rsq13,rinv13);
376 /* EWALD ELECTROSTATICS */
378 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
379 ewrt = _mm_mul_ps(r13,ewtabscale);
380 ewitab = _mm_cvttps_epi32(ewrt);
381 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
382 ewitab = _mm_slli_epi32(ewitab,2);
383 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
384 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
385 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
386 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
387 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
388 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
389 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
390 velec = _mm_mul_ps(qq13,_mm_sub_ps(rinv13,velec));
391 felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec));
393 /* Update potential sum for this i atom from the interaction with this j atom. */
394 velecsum = _mm_add_ps(velecsum,velec);
398 /* Calculate temporary vectorial force */
399 tx = _mm_mul_ps(fscal,dx13);
400 ty = _mm_mul_ps(fscal,dy13);
401 tz = _mm_mul_ps(fscal,dz13);
403 /* Update vectorial force */
404 fix1 = _mm_add_ps(fix1,tx);
405 fiy1 = _mm_add_ps(fiy1,ty);
406 fiz1 = _mm_add_ps(fiz1,tz);
408 fjx3 = _mm_add_ps(fjx3,tx);
409 fjy3 = _mm_add_ps(fjy3,ty);
410 fjz3 = _mm_add_ps(fjz3,tz);
412 /**************************
413 * CALCULATE INTERACTIONS *
414 **************************/
416 r21 = _mm_mul_ps(rsq21,rinv21);
418 /* EWALD ELECTROSTATICS */
420 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
421 ewrt = _mm_mul_ps(r21,ewtabscale);
422 ewitab = _mm_cvttps_epi32(ewrt);
423 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
424 ewitab = _mm_slli_epi32(ewitab,2);
425 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
426 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
427 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
428 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
429 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
430 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
431 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
432 velec = _mm_mul_ps(qq21,_mm_sub_ps(rinv21,velec));
433 felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec));
435 /* Update potential sum for this i atom from the interaction with this j atom. */
436 velecsum = _mm_add_ps(velecsum,velec);
440 /* Calculate temporary vectorial force */
441 tx = _mm_mul_ps(fscal,dx21);
442 ty = _mm_mul_ps(fscal,dy21);
443 tz = _mm_mul_ps(fscal,dz21);
445 /* Update vectorial force */
446 fix2 = _mm_add_ps(fix2,tx);
447 fiy2 = _mm_add_ps(fiy2,ty);
448 fiz2 = _mm_add_ps(fiz2,tz);
450 fjx1 = _mm_add_ps(fjx1,tx);
451 fjy1 = _mm_add_ps(fjy1,ty);
452 fjz1 = _mm_add_ps(fjz1,tz);
454 /**************************
455 * CALCULATE INTERACTIONS *
456 **************************/
458 r22 = _mm_mul_ps(rsq22,rinv22);
460 /* EWALD ELECTROSTATICS */
462 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
463 ewrt = _mm_mul_ps(r22,ewtabscale);
464 ewitab = _mm_cvttps_epi32(ewrt);
465 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
466 ewitab = _mm_slli_epi32(ewitab,2);
467 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
468 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
469 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
470 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
471 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
472 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
473 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
474 velec = _mm_mul_ps(qq22,_mm_sub_ps(rinv22,velec));
475 felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec));
477 /* Update potential sum for this i atom from the interaction with this j atom. */
478 velecsum = _mm_add_ps(velecsum,velec);
482 /* Calculate temporary vectorial force */
483 tx = _mm_mul_ps(fscal,dx22);
484 ty = _mm_mul_ps(fscal,dy22);
485 tz = _mm_mul_ps(fscal,dz22);
487 /* Update vectorial force */
488 fix2 = _mm_add_ps(fix2,tx);
489 fiy2 = _mm_add_ps(fiy2,ty);
490 fiz2 = _mm_add_ps(fiz2,tz);
492 fjx2 = _mm_add_ps(fjx2,tx);
493 fjy2 = _mm_add_ps(fjy2,ty);
494 fjz2 = _mm_add_ps(fjz2,tz);
496 /**************************
497 * CALCULATE INTERACTIONS *
498 **************************/
500 r23 = _mm_mul_ps(rsq23,rinv23);
502 /* EWALD ELECTROSTATICS */
504 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
505 ewrt = _mm_mul_ps(r23,ewtabscale);
506 ewitab = _mm_cvttps_epi32(ewrt);
507 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
508 ewitab = _mm_slli_epi32(ewitab,2);
509 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
510 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
511 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
512 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
513 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
514 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
515 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
516 velec = _mm_mul_ps(qq23,_mm_sub_ps(rinv23,velec));
517 felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec));
519 /* Update potential sum for this i atom from the interaction with this j atom. */
520 velecsum = _mm_add_ps(velecsum,velec);
524 /* Calculate temporary vectorial force */
525 tx = _mm_mul_ps(fscal,dx23);
526 ty = _mm_mul_ps(fscal,dy23);
527 tz = _mm_mul_ps(fscal,dz23);
529 /* Update vectorial force */
530 fix2 = _mm_add_ps(fix2,tx);
531 fiy2 = _mm_add_ps(fiy2,ty);
532 fiz2 = _mm_add_ps(fiz2,tz);
534 fjx3 = _mm_add_ps(fjx3,tx);
535 fjy3 = _mm_add_ps(fjy3,ty);
536 fjz3 = _mm_add_ps(fjz3,tz);
538 /**************************
539 * CALCULATE INTERACTIONS *
540 **************************/
542 r31 = _mm_mul_ps(rsq31,rinv31);
544 /* EWALD ELECTROSTATICS */
546 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
547 ewrt = _mm_mul_ps(r31,ewtabscale);
548 ewitab = _mm_cvttps_epi32(ewrt);
549 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
550 ewitab = _mm_slli_epi32(ewitab,2);
551 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
552 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
553 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
554 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
555 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
556 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
557 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
558 velec = _mm_mul_ps(qq31,_mm_sub_ps(rinv31,velec));
559 felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec));
561 /* Update potential sum for this i atom from the interaction with this j atom. */
562 velecsum = _mm_add_ps(velecsum,velec);
566 /* Calculate temporary vectorial force */
567 tx = _mm_mul_ps(fscal,dx31);
568 ty = _mm_mul_ps(fscal,dy31);
569 tz = _mm_mul_ps(fscal,dz31);
571 /* Update vectorial force */
572 fix3 = _mm_add_ps(fix3,tx);
573 fiy3 = _mm_add_ps(fiy3,ty);
574 fiz3 = _mm_add_ps(fiz3,tz);
576 fjx1 = _mm_add_ps(fjx1,tx);
577 fjy1 = _mm_add_ps(fjy1,ty);
578 fjz1 = _mm_add_ps(fjz1,tz);
580 /**************************
581 * CALCULATE INTERACTIONS *
582 **************************/
584 r32 = _mm_mul_ps(rsq32,rinv32);
586 /* EWALD ELECTROSTATICS */
588 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
589 ewrt = _mm_mul_ps(r32,ewtabscale);
590 ewitab = _mm_cvttps_epi32(ewrt);
591 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
592 ewitab = _mm_slli_epi32(ewitab,2);
593 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
594 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
595 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
596 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
597 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
598 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
599 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
600 velec = _mm_mul_ps(qq32,_mm_sub_ps(rinv32,velec));
601 felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec));
603 /* Update potential sum for this i atom from the interaction with this j atom. */
604 velecsum = _mm_add_ps(velecsum,velec);
608 /* Calculate temporary vectorial force */
609 tx = _mm_mul_ps(fscal,dx32);
610 ty = _mm_mul_ps(fscal,dy32);
611 tz = _mm_mul_ps(fscal,dz32);
613 /* Update vectorial force */
614 fix3 = _mm_add_ps(fix3,tx);
615 fiy3 = _mm_add_ps(fiy3,ty);
616 fiz3 = _mm_add_ps(fiz3,tz);
618 fjx2 = _mm_add_ps(fjx2,tx);
619 fjy2 = _mm_add_ps(fjy2,ty);
620 fjz2 = _mm_add_ps(fjz2,tz);
622 /**************************
623 * CALCULATE INTERACTIONS *
624 **************************/
626 r33 = _mm_mul_ps(rsq33,rinv33);
628 /* EWALD ELECTROSTATICS */
630 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
631 ewrt = _mm_mul_ps(r33,ewtabscale);
632 ewitab = _mm_cvttps_epi32(ewrt);
633 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
634 ewitab = _mm_slli_epi32(ewitab,2);
635 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
636 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
637 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
638 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
639 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
640 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
641 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
642 velec = _mm_mul_ps(qq33,_mm_sub_ps(rinv33,velec));
643 felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec));
645 /* Update potential sum for this i atom from the interaction with this j atom. */
646 velecsum = _mm_add_ps(velecsum,velec);
650 /* Calculate temporary vectorial force */
651 tx = _mm_mul_ps(fscal,dx33);
652 ty = _mm_mul_ps(fscal,dy33);
653 tz = _mm_mul_ps(fscal,dz33);
655 /* Update vectorial force */
656 fix3 = _mm_add_ps(fix3,tx);
657 fiy3 = _mm_add_ps(fiy3,ty);
658 fiz3 = _mm_add_ps(fiz3,tz);
660 fjx3 = _mm_add_ps(fjx3,tx);
661 fjy3 = _mm_add_ps(fjy3,ty);
662 fjz3 = _mm_add_ps(fjz3,tz);
664 fjptrA = f+j_coord_offsetA;
665 fjptrB = f+j_coord_offsetB;
666 fjptrC = f+j_coord_offsetC;
667 fjptrD = f+j_coord_offsetD;
669 gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
670 fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
672 /* Inner loop uses 369 flops */
678 /* Get j neighbor index, and coordinate index */
679 jnrlistA = jjnr[jidx];
680 jnrlistB = jjnr[jidx+1];
681 jnrlistC = jjnr[jidx+2];
682 jnrlistD = jjnr[jidx+3];
683 /* Sign of each element will be negative for non-real atoms.
684 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
685 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
687 dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
688 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
689 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
690 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
691 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
692 j_coord_offsetA = DIM*jnrA;
693 j_coord_offsetB = DIM*jnrB;
694 j_coord_offsetC = DIM*jnrC;
695 j_coord_offsetD = DIM*jnrD;
697 /* load j atom coordinates */
698 gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
699 x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
700 &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
702 /* Calculate displacement vector */
703 dx11 = _mm_sub_ps(ix1,jx1);
704 dy11 = _mm_sub_ps(iy1,jy1);
705 dz11 = _mm_sub_ps(iz1,jz1);
706 dx12 = _mm_sub_ps(ix1,jx2);
707 dy12 = _mm_sub_ps(iy1,jy2);
708 dz12 = _mm_sub_ps(iz1,jz2);
709 dx13 = _mm_sub_ps(ix1,jx3);
710 dy13 = _mm_sub_ps(iy1,jy3);
711 dz13 = _mm_sub_ps(iz1,jz3);
712 dx21 = _mm_sub_ps(ix2,jx1);
713 dy21 = _mm_sub_ps(iy2,jy1);
714 dz21 = _mm_sub_ps(iz2,jz1);
715 dx22 = _mm_sub_ps(ix2,jx2);
716 dy22 = _mm_sub_ps(iy2,jy2);
717 dz22 = _mm_sub_ps(iz2,jz2);
718 dx23 = _mm_sub_ps(ix2,jx3);
719 dy23 = _mm_sub_ps(iy2,jy3);
720 dz23 = _mm_sub_ps(iz2,jz3);
721 dx31 = _mm_sub_ps(ix3,jx1);
722 dy31 = _mm_sub_ps(iy3,jy1);
723 dz31 = _mm_sub_ps(iz3,jz1);
724 dx32 = _mm_sub_ps(ix3,jx2);
725 dy32 = _mm_sub_ps(iy3,jy2);
726 dz32 = _mm_sub_ps(iz3,jz2);
727 dx33 = _mm_sub_ps(ix3,jx3);
728 dy33 = _mm_sub_ps(iy3,jy3);
729 dz33 = _mm_sub_ps(iz3,jz3);
731 /* Calculate squared distance and things based on it */
732 rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
733 rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
734 rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
735 rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
736 rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
737 rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
738 rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
739 rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
740 rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
742 rinv11 = gmx_mm_invsqrt_ps(rsq11);
743 rinv12 = gmx_mm_invsqrt_ps(rsq12);
744 rinv13 = gmx_mm_invsqrt_ps(rsq13);
745 rinv21 = gmx_mm_invsqrt_ps(rsq21);
746 rinv22 = gmx_mm_invsqrt_ps(rsq22);
747 rinv23 = gmx_mm_invsqrt_ps(rsq23);
748 rinv31 = gmx_mm_invsqrt_ps(rsq31);
749 rinv32 = gmx_mm_invsqrt_ps(rsq32);
750 rinv33 = gmx_mm_invsqrt_ps(rsq33);
752 rinvsq11 = _mm_mul_ps(rinv11,rinv11);
753 rinvsq12 = _mm_mul_ps(rinv12,rinv12);
754 rinvsq13 = _mm_mul_ps(rinv13,rinv13);
755 rinvsq21 = _mm_mul_ps(rinv21,rinv21);
756 rinvsq22 = _mm_mul_ps(rinv22,rinv22);
757 rinvsq23 = _mm_mul_ps(rinv23,rinv23);
758 rinvsq31 = _mm_mul_ps(rinv31,rinv31);
759 rinvsq32 = _mm_mul_ps(rinv32,rinv32);
760 rinvsq33 = _mm_mul_ps(rinv33,rinv33);
762 fjx1 = _mm_setzero_ps();
763 fjy1 = _mm_setzero_ps();
764 fjz1 = _mm_setzero_ps();
765 fjx2 = _mm_setzero_ps();
766 fjy2 = _mm_setzero_ps();
767 fjz2 = _mm_setzero_ps();
768 fjx3 = _mm_setzero_ps();
769 fjy3 = _mm_setzero_ps();
770 fjz3 = _mm_setzero_ps();
772 /**************************
773 * CALCULATE INTERACTIONS *
774 **************************/
776 r11 = _mm_mul_ps(rsq11,rinv11);
777 r11 = _mm_andnot_ps(dummy_mask,r11);
779 /* EWALD ELECTROSTATICS */
781 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
782 ewrt = _mm_mul_ps(r11,ewtabscale);
783 ewitab = _mm_cvttps_epi32(ewrt);
784 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
785 ewitab = _mm_slli_epi32(ewitab,2);
786 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
787 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
788 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
789 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
790 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
791 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
792 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
793 velec = _mm_mul_ps(qq11,_mm_sub_ps(rinv11,velec));
794 felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec));
796 /* Update potential sum for this i atom from the interaction with this j atom. */
797 velec = _mm_andnot_ps(dummy_mask,velec);
798 velecsum = _mm_add_ps(velecsum,velec);
802 fscal = _mm_andnot_ps(dummy_mask,fscal);
804 /* Calculate temporary vectorial force */
805 tx = _mm_mul_ps(fscal,dx11);
806 ty = _mm_mul_ps(fscal,dy11);
807 tz = _mm_mul_ps(fscal,dz11);
809 /* Update vectorial force */
810 fix1 = _mm_add_ps(fix1,tx);
811 fiy1 = _mm_add_ps(fiy1,ty);
812 fiz1 = _mm_add_ps(fiz1,tz);
814 fjx1 = _mm_add_ps(fjx1,tx);
815 fjy1 = _mm_add_ps(fjy1,ty);
816 fjz1 = _mm_add_ps(fjz1,tz);
818 /**************************
819 * CALCULATE INTERACTIONS *
820 **************************/
822 r12 = _mm_mul_ps(rsq12,rinv12);
823 r12 = _mm_andnot_ps(dummy_mask,r12);
825 /* EWALD ELECTROSTATICS */
827 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
828 ewrt = _mm_mul_ps(r12,ewtabscale);
829 ewitab = _mm_cvttps_epi32(ewrt);
830 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
831 ewitab = _mm_slli_epi32(ewitab,2);
832 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
833 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
834 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
835 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
836 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
837 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
838 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
839 velec = _mm_mul_ps(qq12,_mm_sub_ps(rinv12,velec));
840 felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec));
842 /* Update potential sum for this i atom from the interaction with this j atom. */
843 velec = _mm_andnot_ps(dummy_mask,velec);
844 velecsum = _mm_add_ps(velecsum,velec);
848 fscal = _mm_andnot_ps(dummy_mask,fscal);
850 /* Calculate temporary vectorial force */
851 tx = _mm_mul_ps(fscal,dx12);
852 ty = _mm_mul_ps(fscal,dy12);
853 tz = _mm_mul_ps(fscal,dz12);
855 /* Update vectorial force */
856 fix1 = _mm_add_ps(fix1,tx);
857 fiy1 = _mm_add_ps(fiy1,ty);
858 fiz1 = _mm_add_ps(fiz1,tz);
860 fjx2 = _mm_add_ps(fjx2,tx);
861 fjy2 = _mm_add_ps(fjy2,ty);
862 fjz2 = _mm_add_ps(fjz2,tz);
864 /**************************
865 * CALCULATE INTERACTIONS *
866 **************************/
868 r13 = _mm_mul_ps(rsq13,rinv13);
869 r13 = _mm_andnot_ps(dummy_mask,r13);
871 /* EWALD ELECTROSTATICS */
873 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
874 ewrt = _mm_mul_ps(r13,ewtabscale);
875 ewitab = _mm_cvttps_epi32(ewrt);
876 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
877 ewitab = _mm_slli_epi32(ewitab,2);
878 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
879 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
880 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
881 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
882 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
883 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
884 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
885 velec = _mm_mul_ps(qq13,_mm_sub_ps(rinv13,velec));
886 felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec));
888 /* Update potential sum for this i atom from the interaction with this j atom. */
889 velec = _mm_andnot_ps(dummy_mask,velec);
890 velecsum = _mm_add_ps(velecsum,velec);
894 fscal = _mm_andnot_ps(dummy_mask,fscal);
896 /* Calculate temporary vectorial force */
897 tx = _mm_mul_ps(fscal,dx13);
898 ty = _mm_mul_ps(fscal,dy13);
899 tz = _mm_mul_ps(fscal,dz13);
901 /* Update vectorial force */
902 fix1 = _mm_add_ps(fix1,tx);
903 fiy1 = _mm_add_ps(fiy1,ty);
904 fiz1 = _mm_add_ps(fiz1,tz);
906 fjx3 = _mm_add_ps(fjx3,tx);
907 fjy3 = _mm_add_ps(fjy3,ty);
908 fjz3 = _mm_add_ps(fjz3,tz);
910 /**************************
911 * CALCULATE INTERACTIONS *
912 **************************/
914 r21 = _mm_mul_ps(rsq21,rinv21);
915 r21 = _mm_andnot_ps(dummy_mask,r21);
917 /* EWALD ELECTROSTATICS */
919 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
920 ewrt = _mm_mul_ps(r21,ewtabscale);
921 ewitab = _mm_cvttps_epi32(ewrt);
922 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
923 ewitab = _mm_slli_epi32(ewitab,2);
924 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
925 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
926 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
927 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
928 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
929 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
930 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
931 velec = _mm_mul_ps(qq21,_mm_sub_ps(rinv21,velec));
932 felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec));
934 /* Update potential sum for this i atom from the interaction with this j atom. */
935 velec = _mm_andnot_ps(dummy_mask,velec);
936 velecsum = _mm_add_ps(velecsum,velec);
940 fscal = _mm_andnot_ps(dummy_mask,fscal);
942 /* Calculate temporary vectorial force */
943 tx = _mm_mul_ps(fscal,dx21);
944 ty = _mm_mul_ps(fscal,dy21);
945 tz = _mm_mul_ps(fscal,dz21);
947 /* Update vectorial force */
948 fix2 = _mm_add_ps(fix2,tx);
949 fiy2 = _mm_add_ps(fiy2,ty);
950 fiz2 = _mm_add_ps(fiz2,tz);
952 fjx1 = _mm_add_ps(fjx1,tx);
953 fjy1 = _mm_add_ps(fjy1,ty);
954 fjz1 = _mm_add_ps(fjz1,tz);
956 /**************************
957 * CALCULATE INTERACTIONS *
958 **************************/
960 r22 = _mm_mul_ps(rsq22,rinv22);
961 r22 = _mm_andnot_ps(dummy_mask,r22);
963 /* EWALD ELECTROSTATICS */
965 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
966 ewrt = _mm_mul_ps(r22,ewtabscale);
967 ewitab = _mm_cvttps_epi32(ewrt);
968 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
969 ewitab = _mm_slli_epi32(ewitab,2);
970 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
971 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
972 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
973 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
974 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
975 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
976 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
977 velec = _mm_mul_ps(qq22,_mm_sub_ps(rinv22,velec));
978 felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec));
980 /* Update potential sum for this i atom from the interaction with this j atom. */
981 velec = _mm_andnot_ps(dummy_mask,velec);
982 velecsum = _mm_add_ps(velecsum,velec);
986 fscal = _mm_andnot_ps(dummy_mask,fscal);
988 /* Calculate temporary vectorial force */
989 tx = _mm_mul_ps(fscal,dx22);
990 ty = _mm_mul_ps(fscal,dy22);
991 tz = _mm_mul_ps(fscal,dz22);
993 /* Update vectorial force */
994 fix2 = _mm_add_ps(fix2,tx);
995 fiy2 = _mm_add_ps(fiy2,ty);
996 fiz2 = _mm_add_ps(fiz2,tz);
998 fjx2 = _mm_add_ps(fjx2,tx);
999 fjy2 = _mm_add_ps(fjy2,ty);
1000 fjz2 = _mm_add_ps(fjz2,tz);
1002 /**************************
1003 * CALCULATE INTERACTIONS *
1004 **************************/
1006 r23 = _mm_mul_ps(rsq23,rinv23);
1007 r23 = _mm_andnot_ps(dummy_mask,r23);
1009 /* EWALD ELECTROSTATICS */
1011 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1012 ewrt = _mm_mul_ps(r23,ewtabscale);
1013 ewitab = _mm_cvttps_epi32(ewrt);
1014 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1015 ewitab = _mm_slli_epi32(ewitab,2);
1016 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1017 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
1018 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
1019 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
1020 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
1021 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
1022 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
1023 velec = _mm_mul_ps(qq23,_mm_sub_ps(rinv23,velec));
1024 felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec));
1026 /* Update potential sum for this i atom from the interaction with this j atom. */
1027 velec = _mm_andnot_ps(dummy_mask,velec);
1028 velecsum = _mm_add_ps(velecsum,velec);
1032 fscal = _mm_andnot_ps(dummy_mask,fscal);
1034 /* Calculate temporary vectorial force */
1035 tx = _mm_mul_ps(fscal,dx23);
1036 ty = _mm_mul_ps(fscal,dy23);
1037 tz = _mm_mul_ps(fscal,dz23);
1039 /* Update vectorial force */
1040 fix2 = _mm_add_ps(fix2,tx);
1041 fiy2 = _mm_add_ps(fiy2,ty);
1042 fiz2 = _mm_add_ps(fiz2,tz);
1044 fjx3 = _mm_add_ps(fjx3,tx);
1045 fjy3 = _mm_add_ps(fjy3,ty);
1046 fjz3 = _mm_add_ps(fjz3,tz);
1048 /**************************
1049 * CALCULATE INTERACTIONS *
1050 **************************/
1052 r31 = _mm_mul_ps(rsq31,rinv31);
1053 r31 = _mm_andnot_ps(dummy_mask,r31);
1055 /* EWALD ELECTROSTATICS */
1057 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1058 ewrt = _mm_mul_ps(r31,ewtabscale);
1059 ewitab = _mm_cvttps_epi32(ewrt);
1060 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1061 ewitab = _mm_slli_epi32(ewitab,2);
1062 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1063 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
1064 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
1065 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
1066 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
1067 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
1068 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
1069 velec = _mm_mul_ps(qq31,_mm_sub_ps(rinv31,velec));
1070 felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec));
1072 /* Update potential sum for this i atom from the interaction with this j atom. */
1073 velec = _mm_andnot_ps(dummy_mask,velec);
1074 velecsum = _mm_add_ps(velecsum,velec);
1078 fscal = _mm_andnot_ps(dummy_mask,fscal);
1080 /* Calculate temporary vectorial force */
1081 tx = _mm_mul_ps(fscal,dx31);
1082 ty = _mm_mul_ps(fscal,dy31);
1083 tz = _mm_mul_ps(fscal,dz31);
1085 /* Update vectorial force */
1086 fix3 = _mm_add_ps(fix3,tx);
1087 fiy3 = _mm_add_ps(fiy3,ty);
1088 fiz3 = _mm_add_ps(fiz3,tz);
1090 fjx1 = _mm_add_ps(fjx1,tx);
1091 fjy1 = _mm_add_ps(fjy1,ty);
1092 fjz1 = _mm_add_ps(fjz1,tz);
1094 /**************************
1095 * CALCULATE INTERACTIONS *
1096 **************************/
1098 r32 = _mm_mul_ps(rsq32,rinv32);
1099 r32 = _mm_andnot_ps(dummy_mask,r32);
1101 /* EWALD ELECTROSTATICS */
1103 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1104 ewrt = _mm_mul_ps(r32,ewtabscale);
1105 ewitab = _mm_cvttps_epi32(ewrt);
1106 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1107 ewitab = _mm_slli_epi32(ewitab,2);
1108 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1109 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
1110 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
1111 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
1112 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
1113 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
1114 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
1115 velec = _mm_mul_ps(qq32,_mm_sub_ps(rinv32,velec));
1116 felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec));
1118 /* Update potential sum for this i atom from the interaction with this j atom. */
1119 velec = _mm_andnot_ps(dummy_mask,velec);
1120 velecsum = _mm_add_ps(velecsum,velec);
1124 fscal = _mm_andnot_ps(dummy_mask,fscal);
1126 /* Calculate temporary vectorial force */
1127 tx = _mm_mul_ps(fscal,dx32);
1128 ty = _mm_mul_ps(fscal,dy32);
1129 tz = _mm_mul_ps(fscal,dz32);
1131 /* Update vectorial force */
1132 fix3 = _mm_add_ps(fix3,tx);
1133 fiy3 = _mm_add_ps(fiy3,ty);
1134 fiz3 = _mm_add_ps(fiz3,tz);
1136 fjx2 = _mm_add_ps(fjx2,tx);
1137 fjy2 = _mm_add_ps(fjy2,ty);
1138 fjz2 = _mm_add_ps(fjz2,tz);
1140 /**************************
1141 * CALCULATE INTERACTIONS *
1142 **************************/
1144 r33 = _mm_mul_ps(rsq33,rinv33);
1145 r33 = _mm_andnot_ps(dummy_mask,r33);
1147 /* EWALD ELECTROSTATICS */
1149 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1150 ewrt = _mm_mul_ps(r33,ewtabscale);
1151 ewitab = _mm_cvttps_epi32(ewrt);
1152 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1153 ewitab = _mm_slli_epi32(ewitab,2);
1154 ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
1155 ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
1156 ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
1157 ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
1158 _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
1159 felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
1160 velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
1161 velec = _mm_mul_ps(qq33,_mm_sub_ps(rinv33,velec));
1162 felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec));
1164 /* Update potential sum for this i atom from the interaction with this j atom. */
1165 velec = _mm_andnot_ps(dummy_mask,velec);
1166 velecsum = _mm_add_ps(velecsum,velec);
1170 fscal = _mm_andnot_ps(dummy_mask,fscal);
1172 /* Calculate temporary vectorial force */
1173 tx = _mm_mul_ps(fscal,dx33);
1174 ty = _mm_mul_ps(fscal,dy33);
1175 tz = _mm_mul_ps(fscal,dz33);
1177 /* Update vectorial force */
1178 fix3 = _mm_add_ps(fix3,tx);
1179 fiy3 = _mm_add_ps(fiy3,ty);
1180 fiz3 = _mm_add_ps(fiz3,tz);
1182 fjx3 = _mm_add_ps(fjx3,tx);
1183 fjy3 = _mm_add_ps(fjy3,ty);
1184 fjz3 = _mm_add_ps(fjz3,tz);
1186 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1187 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1188 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1189 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1191 gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1192 fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1194 /* Inner loop uses 378 flops */
1197 /* End of innermost loop */
1199 gmx_mm_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1200 f+i_coord_offset+DIM,fshift+i_shift_offset);
1203 /* Update potential energies */
1204 gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1206 /* Increment number of inner iterations */
1207 inneriter += j_index_end - j_index_start;
1209 /* Outer loop uses 19 flops */
1212 /* Increment number of outer iterations */
1215 /* Update outer/inner flops */
1217 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_VF,outeriter*19 + inneriter*378);
1220 * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwNone_GeomW4W4_F_sse2_single
1221 * Electrostatics interaction: Ewald
1222 * VdW interaction: None
1223 * Geometry: Water4-Water4
1224 * Calculate force/pot: Force
1227 nb_kernel_ElecEw_VdwNone_GeomW4W4_F_sse2_single
1228 (t_nblist * gmx_restrict nlist,
1229 rvec * gmx_restrict xx,
1230 rvec * gmx_restrict ff,
1231 t_forcerec * gmx_restrict fr,
1232 t_mdatoms * gmx_restrict mdatoms,
1233 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1234 t_nrnb * gmx_restrict nrnb)
1236 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1237 * just 0 for non-waters.
1238 * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
1239 * jnr indices corresponding to data put in the four positions in the SIMD register.
1241 int i_shift_offset,i_coord_offset,outeriter,inneriter;
1242 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1243 int jnrA,jnrB,jnrC,jnrD;
1244 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1245 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1246 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
1247 real rcutoff_scalar;
1248 real *shiftvec,*fshift,*x,*f;
1249 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
1250 real scratch[4*DIM];
1251 __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1253 __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1255 __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1257 __m128 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1258 int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
1259 __m128 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1260 int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
1261 __m128 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1262 int vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
1263 __m128 jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
1264 __m128 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1265 __m128 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1266 __m128 dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
1267 __m128 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1268 __m128 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1269 __m128 dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
1270 __m128 dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
1271 __m128 dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
1272 __m128 dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
1273 __m128 velec,felec,velecsum,facel,crf,krf,krf2;
1276 __m128 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1278 __m128 dummy_mask,cutoff_mask;
1279 __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
1280 __m128 one = _mm_set1_ps(1.0);
1281 __m128 two = _mm_set1_ps(2.0);
1287 jindex = nlist->jindex;
1289 shiftidx = nlist->shift;
1291 shiftvec = fr->shift_vec[0];
1292 fshift = fr->fshift[0];
1293 facel = _mm_set1_ps(fr->epsfac);
1294 charge = mdatoms->chargeA;
1296 sh_ewald = _mm_set1_ps(fr->ic->sh_ewald);
1297 ewtab = fr->ic->tabq_coul_F;
1298 ewtabscale = _mm_set1_ps(fr->ic->tabq_scale);
1299 ewtabhalfspace = _mm_set1_ps(0.5/fr->ic->tabq_scale);
1301 /* Setup water-specific parameters */
1302 inr = nlist->iinr[0];
1303 iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
1304 iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
1305 iq3 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
1307 jq1 = _mm_set1_ps(charge[inr+1]);
1308 jq2 = _mm_set1_ps(charge[inr+2]);
1309 jq3 = _mm_set1_ps(charge[inr+3]);
1310 qq11 = _mm_mul_ps(iq1,jq1);
1311 qq12 = _mm_mul_ps(iq1,jq2);
1312 qq13 = _mm_mul_ps(iq1,jq3);
1313 qq21 = _mm_mul_ps(iq2,jq1);
1314 qq22 = _mm_mul_ps(iq2,jq2);
1315 qq23 = _mm_mul_ps(iq2,jq3);
1316 qq31 = _mm_mul_ps(iq3,jq1);
1317 qq32 = _mm_mul_ps(iq3,jq2);
1318 qq33 = _mm_mul_ps(iq3,jq3);
1320 /* Avoid stupid compiler warnings */
1321 jnrA = jnrB = jnrC = jnrD = 0;
1322 j_coord_offsetA = 0;
1323 j_coord_offsetB = 0;
1324 j_coord_offsetC = 0;
1325 j_coord_offsetD = 0;
1330 for(iidx=0;iidx<4*DIM;iidx++)
1332 scratch[iidx] = 0.0;
1335 /* Start outer loop over neighborlists */
1336 for(iidx=0; iidx<nri; iidx++)
1338 /* Load shift vector for this list */
1339 i_shift_offset = DIM*shiftidx[iidx];
1341 /* Load limits for loop over neighbors */
1342 j_index_start = jindex[iidx];
1343 j_index_end = jindex[iidx+1];
1345 /* Get outer coordinate index */
1347 i_coord_offset = DIM*inr;
1349 /* Load i particle coords and add shift vector */
1350 gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
1351 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1353 fix1 = _mm_setzero_ps();
1354 fiy1 = _mm_setzero_ps();
1355 fiz1 = _mm_setzero_ps();
1356 fix2 = _mm_setzero_ps();
1357 fiy2 = _mm_setzero_ps();
1358 fiz2 = _mm_setzero_ps();
1359 fix3 = _mm_setzero_ps();
1360 fiy3 = _mm_setzero_ps();
1361 fiz3 = _mm_setzero_ps();
1363 /* Start inner kernel loop */
1364 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1367 /* Get j neighbor index, and coordinate index */
1369 jnrB = jjnr[jidx+1];
1370 jnrC = jjnr[jidx+2];
1371 jnrD = jjnr[jidx+3];
1372 j_coord_offsetA = DIM*jnrA;
1373 j_coord_offsetB = DIM*jnrB;
1374 j_coord_offsetC = DIM*jnrC;
1375 j_coord_offsetD = DIM*jnrD;
1377 /* load j atom coordinates */
1378 gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
1379 x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
1380 &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
1382 /* Calculate displacement vector */
1383 dx11 = _mm_sub_ps(ix1,jx1);
1384 dy11 = _mm_sub_ps(iy1,jy1);
1385 dz11 = _mm_sub_ps(iz1,jz1);
1386 dx12 = _mm_sub_ps(ix1,jx2);
1387 dy12 = _mm_sub_ps(iy1,jy2);
1388 dz12 = _mm_sub_ps(iz1,jz2);
1389 dx13 = _mm_sub_ps(ix1,jx3);
1390 dy13 = _mm_sub_ps(iy1,jy3);
1391 dz13 = _mm_sub_ps(iz1,jz3);
1392 dx21 = _mm_sub_ps(ix2,jx1);
1393 dy21 = _mm_sub_ps(iy2,jy1);
1394 dz21 = _mm_sub_ps(iz2,jz1);
1395 dx22 = _mm_sub_ps(ix2,jx2);
1396 dy22 = _mm_sub_ps(iy2,jy2);
1397 dz22 = _mm_sub_ps(iz2,jz2);
1398 dx23 = _mm_sub_ps(ix2,jx3);
1399 dy23 = _mm_sub_ps(iy2,jy3);
1400 dz23 = _mm_sub_ps(iz2,jz3);
1401 dx31 = _mm_sub_ps(ix3,jx1);
1402 dy31 = _mm_sub_ps(iy3,jy1);
1403 dz31 = _mm_sub_ps(iz3,jz1);
1404 dx32 = _mm_sub_ps(ix3,jx2);
1405 dy32 = _mm_sub_ps(iy3,jy2);
1406 dz32 = _mm_sub_ps(iz3,jz2);
1407 dx33 = _mm_sub_ps(ix3,jx3);
1408 dy33 = _mm_sub_ps(iy3,jy3);
1409 dz33 = _mm_sub_ps(iz3,jz3);
1411 /* Calculate squared distance and things based on it */
1412 rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
1413 rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
1414 rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
1415 rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
1416 rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
1417 rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
1418 rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
1419 rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
1420 rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
1422 rinv11 = gmx_mm_invsqrt_ps(rsq11);
1423 rinv12 = gmx_mm_invsqrt_ps(rsq12);
1424 rinv13 = gmx_mm_invsqrt_ps(rsq13);
1425 rinv21 = gmx_mm_invsqrt_ps(rsq21);
1426 rinv22 = gmx_mm_invsqrt_ps(rsq22);
1427 rinv23 = gmx_mm_invsqrt_ps(rsq23);
1428 rinv31 = gmx_mm_invsqrt_ps(rsq31);
1429 rinv32 = gmx_mm_invsqrt_ps(rsq32);
1430 rinv33 = gmx_mm_invsqrt_ps(rsq33);
1432 rinvsq11 = _mm_mul_ps(rinv11,rinv11);
1433 rinvsq12 = _mm_mul_ps(rinv12,rinv12);
1434 rinvsq13 = _mm_mul_ps(rinv13,rinv13);
1435 rinvsq21 = _mm_mul_ps(rinv21,rinv21);
1436 rinvsq22 = _mm_mul_ps(rinv22,rinv22);
1437 rinvsq23 = _mm_mul_ps(rinv23,rinv23);
1438 rinvsq31 = _mm_mul_ps(rinv31,rinv31);
1439 rinvsq32 = _mm_mul_ps(rinv32,rinv32);
1440 rinvsq33 = _mm_mul_ps(rinv33,rinv33);
1442 fjx1 = _mm_setzero_ps();
1443 fjy1 = _mm_setzero_ps();
1444 fjz1 = _mm_setzero_ps();
1445 fjx2 = _mm_setzero_ps();
1446 fjy2 = _mm_setzero_ps();
1447 fjz2 = _mm_setzero_ps();
1448 fjx3 = _mm_setzero_ps();
1449 fjy3 = _mm_setzero_ps();
1450 fjz3 = _mm_setzero_ps();
1452 /**************************
1453 * CALCULATE INTERACTIONS *
1454 **************************/
1456 r11 = _mm_mul_ps(rsq11,rinv11);
1458 /* EWALD ELECTROSTATICS */
1460 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1461 ewrt = _mm_mul_ps(r11,ewtabscale);
1462 ewitab = _mm_cvttps_epi32(ewrt);
1463 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1464 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1465 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1467 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1468 felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec));
1472 /* Calculate temporary vectorial force */
1473 tx = _mm_mul_ps(fscal,dx11);
1474 ty = _mm_mul_ps(fscal,dy11);
1475 tz = _mm_mul_ps(fscal,dz11);
1477 /* Update vectorial force */
1478 fix1 = _mm_add_ps(fix1,tx);
1479 fiy1 = _mm_add_ps(fiy1,ty);
1480 fiz1 = _mm_add_ps(fiz1,tz);
1482 fjx1 = _mm_add_ps(fjx1,tx);
1483 fjy1 = _mm_add_ps(fjy1,ty);
1484 fjz1 = _mm_add_ps(fjz1,tz);
1486 /**************************
1487 * CALCULATE INTERACTIONS *
1488 **************************/
1490 r12 = _mm_mul_ps(rsq12,rinv12);
1492 /* EWALD ELECTROSTATICS */
1494 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1495 ewrt = _mm_mul_ps(r12,ewtabscale);
1496 ewitab = _mm_cvttps_epi32(ewrt);
1497 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1498 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1499 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1501 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1502 felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec));
1506 /* Calculate temporary vectorial force */
1507 tx = _mm_mul_ps(fscal,dx12);
1508 ty = _mm_mul_ps(fscal,dy12);
1509 tz = _mm_mul_ps(fscal,dz12);
1511 /* Update vectorial force */
1512 fix1 = _mm_add_ps(fix1,tx);
1513 fiy1 = _mm_add_ps(fiy1,ty);
1514 fiz1 = _mm_add_ps(fiz1,tz);
1516 fjx2 = _mm_add_ps(fjx2,tx);
1517 fjy2 = _mm_add_ps(fjy2,ty);
1518 fjz2 = _mm_add_ps(fjz2,tz);
1520 /**************************
1521 * CALCULATE INTERACTIONS *
1522 **************************/
1524 r13 = _mm_mul_ps(rsq13,rinv13);
1526 /* EWALD ELECTROSTATICS */
1528 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1529 ewrt = _mm_mul_ps(r13,ewtabscale);
1530 ewitab = _mm_cvttps_epi32(ewrt);
1531 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1532 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1533 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1535 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1536 felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec));
1540 /* Calculate temporary vectorial force */
1541 tx = _mm_mul_ps(fscal,dx13);
1542 ty = _mm_mul_ps(fscal,dy13);
1543 tz = _mm_mul_ps(fscal,dz13);
1545 /* Update vectorial force */
1546 fix1 = _mm_add_ps(fix1,tx);
1547 fiy1 = _mm_add_ps(fiy1,ty);
1548 fiz1 = _mm_add_ps(fiz1,tz);
1550 fjx3 = _mm_add_ps(fjx3,tx);
1551 fjy3 = _mm_add_ps(fjy3,ty);
1552 fjz3 = _mm_add_ps(fjz3,tz);
1554 /**************************
1555 * CALCULATE INTERACTIONS *
1556 **************************/
1558 r21 = _mm_mul_ps(rsq21,rinv21);
1560 /* EWALD ELECTROSTATICS */
1562 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1563 ewrt = _mm_mul_ps(r21,ewtabscale);
1564 ewitab = _mm_cvttps_epi32(ewrt);
1565 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1566 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1567 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1569 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1570 felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec));
1574 /* Calculate temporary vectorial force */
1575 tx = _mm_mul_ps(fscal,dx21);
1576 ty = _mm_mul_ps(fscal,dy21);
1577 tz = _mm_mul_ps(fscal,dz21);
1579 /* Update vectorial force */
1580 fix2 = _mm_add_ps(fix2,tx);
1581 fiy2 = _mm_add_ps(fiy2,ty);
1582 fiz2 = _mm_add_ps(fiz2,tz);
1584 fjx1 = _mm_add_ps(fjx1,tx);
1585 fjy1 = _mm_add_ps(fjy1,ty);
1586 fjz1 = _mm_add_ps(fjz1,tz);
1588 /**************************
1589 * CALCULATE INTERACTIONS *
1590 **************************/
1592 r22 = _mm_mul_ps(rsq22,rinv22);
1594 /* EWALD ELECTROSTATICS */
1596 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1597 ewrt = _mm_mul_ps(r22,ewtabscale);
1598 ewitab = _mm_cvttps_epi32(ewrt);
1599 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1600 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1601 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1603 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1604 felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec));
1608 /* Calculate temporary vectorial force */
1609 tx = _mm_mul_ps(fscal,dx22);
1610 ty = _mm_mul_ps(fscal,dy22);
1611 tz = _mm_mul_ps(fscal,dz22);
1613 /* Update vectorial force */
1614 fix2 = _mm_add_ps(fix2,tx);
1615 fiy2 = _mm_add_ps(fiy2,ty);
1616 fiz2 = _mm_add_ps(fiz2,tz);
1618 fjx2 = _mm_add_ps(fjx2,tx);
1619 fjy2 = _mm_add_ps(fjy2,ty);
1620 fjz2 = _mm_add_ps(fjz2,tz);
1622 /**************************
1623 * CALCULATE INTERACTIONS *
1624 **************************/
1626 r23 = _mm_mul_ps(rsq23,rinv23);
1628 /* EWALD ELECTROSTATICS */
1630 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1631 ewrt = _mm_mul_ps(r23,ewtabscale);
1632 ewitab = _mm_cvttps_epi32(ewrt);
1633 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1634 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1635 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1637 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1638 felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec));
1642 /* Calculate temporary vectorial force */
1643 tx = _mm_mul_ps(fscal,dx23);
1644 ty = _mm_mul_ps(fscal,dy23);
1645 tz = _mm_mul_ps(fscal,dz23);
1647 /* Update vectorial force */
1648 fix2 = _mm_add_ps(fix2,tx);
1649 fiy2 = _mm_add_ps(fiy2,ty);
1650 fiz2 = _mm_add_ps(fiz2,tz);
1652 fjx3 = _mm_add_ps(fjx3,tx);
1653 fjy3 = _mm_add_ps(fjy3,ty);
1654 fjz3 = _mm_add_ps(fjz3,tz);
1656 /**************************
1657 * CALCULATE INTERACTIONS *
1658 **************************/
1660 r31 = _mm_mul_ps(rsq31,rinv31);
1662 /* EWALD ELECTROSTATICS */
1664 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1665 ewrt = _mm_mul_ps(r31,ewtabscale);
1666 ewitab = _mm_cvttps_epi32(ewrt);
1667 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1668 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1669 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1671 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1672 felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec));
1676 /* Calculate temporary vectorial force */
1677 tx = _mm_mul_ps(fscal,dx31);
1678 ty = _mm_mul_ps(fscal,dy31);
1679 tz = _mm_mul_ps(fscal,dz31);
1681 /* Update vectorial force */
1682 fix3 = _mm_add_ps(fix3,tx);
1683 fiy3 = _mm_add_ps(fiy3,ty);
1684 fiz3 = _mm_add_ps(fiz3,tz);
1686 fjx1 = _mm_add_ps(fjx1,tx);
1687 fjy1 = _mm_add_ps(fjy1,ty);
1688 fjz1 = _mm_add_ps(fjz1,tz);
1690 /**************************
1691 * CALCULATE INTERACTIONS *
1692 **************************/
1694 r32 = _mm_mul_ps(rsq32,rinv32);
1696 /* EWALD ELECTROSTATICS */
1698 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1699 ewrt = _mm_mul_ps(r32,ewtabscale);
1700 ewitab = _mm_cvttps_epi32(ewrt);
1701 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1702 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1703 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1705 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1706 felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec));
1710 /* Calculate temporary vectorial force */
1711 tx = _mm_mul_ps(fscal,dx32);
1712 ty = _mm_mul_ps(fscal,dy32);
1713 tz = _mm_mul_ps(fscal,dz32);
1715 /* Update vectorial force */
1716 fix3 = _mm_add_ps(fix3,tx);
1717 fiy3 = _mm_add_ps(fiy3,ty);
1718 fiz3 = _mm_add_ps(fiz3,tz);
1720 fjx2 = _mm_add_ps(fjx2,tx);
1721 fjy2 = _mm_add_ps(fjy2,ty);
1722 fjz2 = _mm_add_ps(fjz2,tz);
1724 /**************************
1725 * CALCULATE INTERACTIONS *
1726 **************************/
1728 r33 = _mm_mul_ps(rsq33,rinv33);
1730 /* EWALD ELECTROSTATICS */
1732 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1733 ewrt = _mm_mul_ps(r33,ewtabscale);
1734 ewitab = _mm_cvttps_epi32(ewrt);
1735 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1736 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1737 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1739 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1740 felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec));
1744 /* Calculate temporary vectorial force */
1745 tx = _mm_mul_ps(fscal,dx33);
1746 ty = _mm_mul_ps(fscal,dy33);
1747 tz = _mm_mul_ps(fscal,dz33);
1749 /* Update vectorial force */
1750 fix3 = _mm_add_ps(fix3,tx);
1751 fiy3 = _mm_add_ps(fiy3,ty);
1752 fiz3 = _mm_add_ps(fiz3,tz);
1754 fjx3 = _mm_add_ps(fjx3,tx);
1755 fjy3 = _mm_add_ps(fjy3,ty);
1756 fjz3 = _mm_add_ps(fjz3,tz);
1758 fjptrA = f+j_coord_offsetA;
1759 fjptrB = f+j_coord_offsetB;
1760 fjptrC = f+j_coord_offsetC;
1761 fjptrD = f+j_coord_offsetD;
1763 gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1764 fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1766 /* Inner loop uses 324 flops */
1769 if(jidx<j_index_end)
1772 /* Get j neighbor index, and coordinate index */
1773 jnrlistA = jjnr[jidx];
1774 jnrlistB = jjnr[jidx+1];
1775 jnrlistC = jjnr[jidx+2];
1776 jnrlistD = jjnr[jidx+3];
1777 /* Sign of each element will be negative for non-real atoms.
1778 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1779 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1781 dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1782 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
1783 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
1784 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
1785 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
1786 j_coord_offsetA = DIM*jnrA;
1787 j_coord_offsetB = DIM*jnrB;
1788 j_coord_offsetC = DIM*jnrC;
1789 j_coord_offsetD = DIM*jnrD;
1791 /* load j atom coordinates */
1792 gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
1793 x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
1794 &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
1796 /* Calculate displacement vector */
1797 dx11 = _mm_sub_ps(ix1,jx1);
1798 dy11 = _mm_sub_ps(iy1,jy1);
1799 dz11 = _mm_sub_ps(iz1,jz1);
1800 dx12 = _mm_sub_ps(ix1,jx2);
1801 dy12 = _mm_sub_ps(iy1,jy2);
1802 dz12 = _mm_sub_ps(iz1,jz2);
1803 dx13 = _mm_sub_ps(ix1,jx3);
1804 dy13 = _mm_sub_ps(iy1,jy3);
1805 dz13 = _mm_sub_ps(iz1,jz3);
1806 dx21 = _mm_sub_ps(ix2,jx1);
1807 dy21 = _mm_sub_ps(iy2,jy1);
1808 dz21 = _mm_sub_ps(iz2,jz1);
1809 dx22 = _mm_sub_ps(ix2,jx2);
1810 dy22 = _mm_sub_ps(iy2,jy2);
1811 dz22 = _mm_sub_ps(iz2,jz2);
1812 dx23 = _mm_sub_ps(ix2,jx3);
1813 dy23 = _mm_sub_ps(iy2,jy3);
1814 dz23 = _mm_sub_ps(iz2,jz3);
1815 dx31 = _mm_sub_ps(ix3,jx1);
1816 dy31 = _mm_sub_ps(iy3,jy1);
1817 dz31 = _mm_sub_ps(iz3,jz1);
1818 dx32 = _mm_sub_ps(ix3,jx2);
1819 dy32 = _mm_sub_ps(iy3,jy2);
1820 dz32 = _mm_sub_ps(iz3,jz2);
1821 dx33 = _mm_sub_ps(ix3,jx3);
1822 dy33 = _mm_sub_ps(iy3,jy3);
1823 dz33 = _mm_sub_ps(iz3,jz3);
1825 /* Calculate squared distance and things based on it */
1826 rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11);
1827 rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12);
1828 rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13);
1829 rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21);
1830 rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22);
1831 rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23);
1832 rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31);
1833 rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32);
1834 rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33);
1836 rinv11 = gmx_mm_invsqrt_ps(rsq11);
1837 rinv12 = gmx_mm_invsqrt_ps(rsq12);
1838 rinv13 = gmx_mm_invsqrt_ps(rsq13);
1839 rinv21 = gmx_mm_invsqrt_ps(rsq21);
1840 rinv22 = gmx_mm_invsqrt_ps(rsq22);
1841 rinv23 = gmx_mm_invsqrt_ps(rsq23);
1842 rinv31 = gmx_mm_invsqrt_ps(rsq31);
1843 rinv32 = gmx_mm_invsqrt_ps(rsq32);
1844 rinv33 = gmx_mm_invsqrt_ps(rsq33);
1846 rinvsq11 = _mm_mul_ps(rinv11,rinv11);
1847 rinvsq12 = _mm_mul_ps(rinv12,rinv12);
1848 rinvsq13 = _mm_mul_ps(rinv13,rinv13);
1849 rinvsq21 = _mm_mul_ps(rinv21,rinv21);
1850 rinvsq22 = _mm_mul_ps(rinv22,rinv22);
1851 rinvsq23 = _mm_mul_ps(rinv23,rinv23);
1852 rinvsq31 = _mm_mul_ps(rinv31,rinv31);
1853 rinvsq32 = _mm_mul_ps(rinv32,rinv32);
1854 rinvsq33 = _mm_mul_ps(rinv33,rinv33);
1856 fjx1 = _mm_setzero_ps();
1857 fjy1 = _mm_setzero_ps();
1858 fjz1 = _mm_setzero_ps();
1859 fjx2 = _mm_setzero_ps();
1860 fjy2 = _mm_setzero_ps();
1861 fjz2 = _mm_setzero_ps();
1862 fjx3 = _mm_setzero_ps();
1863 fjy3 = _mm_setzero_ps();
1864 fjz3 = _mm_setzero_ps();
1866 /**************************
1867 * CALCULATE INTERACTIONS *
1868 **************************/
1870 r11 = _mm_mul_ps(rsq11,rinv11);
1871 r11 = _mm_andnot_ps(dummy_mask,r11);
1873 /* EWALD ELECTROSTATICS */
1875 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1876 ewrt = _mm_mul_ps(r11,ewtabscale);
1877 ewitab = _mm_cvttps_epi32(ewrt);
1878 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1879 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1880 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1882 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1883 felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec));
1887 fscal = _mm_andnot_ps(dummy_mask,fscal);
1889 /* Calculate temporary vectorial force */
1890 tx = _mm_mul_ps(fscal,dx11);
1891 ty = _mm_mul_ps(fscal,dy11);
1892 tz = _mm_mul_ps(fscal,dz11);
1894 /* Update vectorial force */
1895 fix1 = _mm_add_ps(fix1,tx);
1896 fiy1 = _mm_add_ps(fiy1,ty);
1897 fiz1 = _mm_add_ps(fiz1,tz);
1899 fjx1 = _mm_add_ps(fjx1,tx);
1900 fjy1 = _mm_add_ps(fjy1,ty);
1901 fjz1 = _mm_add_ps(fjz1,tz);
1903 /**************************
1904 * CALCULATE INTERACTIONS *
1905 **************************/
1907 r12 = _mm_mul_ps(rsq12,rinv12);
1908 r12 = _mm_andnot_ps(dummy_mask,r12);
1910 /* EWALD ELECTROSTATICS */
1912 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1913 ewrt = _mm_mul_ps(r12,ewtabscale);
1914 ewitab = _mm_cvttps_epi32(ewrt);
1915 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1916 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1917 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1919 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1920 felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec));
1924 fscal = _mm_andnot_ps(dummy_mask,fscal);
1926 /* Calculate temporary vectorial force */
1927 tx = _mm_mul_ps(fscal,dx12);
1928 ty = _mm_mul_ps(fscal,dy12);
1929 tz = _mm_mul_ps(fscal,dz12);
1931 /* Update vectorial force */
1932 fix1 = _mm_add_ps(fix1,tx);
1933 fiy1 = _mm_add_ps(fiy1,ty);
1934 fiz1 = _mm_add_ps(fiz1,tz);
1936 fjx2 = _mm_add_ps(fjx2,tx);
1937 fjy2 = _mm_add_ps(fjy2,ty);
1938 fjz2 = _mm_add_ps(fjz2,tz);
1940 /**************************
1941 * CALCULATE INTERACTIONS *
1942 **************************/
1944 r13 = _mm_mul_ps(rsq13,rinv13);
1945 r13 = _mm_andnot_ps(dummy_mask,r13);
1947 /* EWALD ELECTROSTATICS */
1949 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1950 ewrt = _mm_mul_ps(r13,ewtabscale);
1951 ewitab = _mm_cvttps_epi32(ewrt);
1952 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1953 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1954 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1956 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1957 felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec));
1961 fscal = _mm_andnot_ps(dummy_mask,fscal);
1963 /* Calculate temporary vectorial force */
1964 tx = _mm_mul_ps(fscal,dx13);
1965 ty = _mm_mul_ps(fscal,dy13);
1966 tz = _mm_mul_ps(fscal,dz13);
1968 /* Update vectorial force */
1969 fix1 = _mm_add_ps(fix1,tx);
1970 fiy1 = _mm_add_ps(fiy1,ty);
1971 fiz1 = _mm_add_ps(fiz1,tz);
1973 fjx3 = _mm_add_ps(fjx3,tx);
1974 fjy3 = _mm_add_ps(fjy3,ty);
1975 fjz3 = _mm_add_ps(fjz3,tz);
1977 /**************************
1978 * CALCULATE INTERACTIONS *
1979 **************************/
1981 r21 = _mm_mul_ps(rsq21,rinv21);
1982 r21 = _mm_andnot_ps(dummy_mask,r21);
1984 /* EWALD ELECTROSTATICS */
1986 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1987 ewrt = _mm_mul_ps(r21,ewtabscale);
1988 ewitab = _mm_cvttps_epi32(ewrt);
1989 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
1990 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
1991 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
1993 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
1994 felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec));
1998 fscal = _mm_andnot_ps(dummy_mask,fscal);
2000 /* Calculate temporary vectorial force */
2001 tx = _mm_mul_ps(fscal,dx21);
2002 ty = _mm_mul_ps(fscal,dy21);
2003 tz = _mm_mul_ps(fscal,dz21);
2005 /* Update vectorial force */
2006 fix2 = _mm_add_ps(fix2,tx);
2007 fiy2 = _mm_add_ps(fiy2,ty);
2008 fiz2 = _mm_add_ps(fiz2,tz);
2010 fjx1 = _mm_add_ps(fjx1,tx);
2011 fjy1 = _mm_add_ps(fjy1,ty);
2012 fjz1 = _mm_add_ps(fjz1,tz);
2014 /**************************
2015 * CALCULATE INTERACTIONS *
2016 **************************/
2018 r22 = _mm_mul_ps(rsq22,rinv22);
2019 r22 = _mm_andnot_ps(dummy_mask,r22);
2021 /* EWALD ELECTROSTATICS */
2023 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2024 ewrt = _mm_mul_ps(r22,ewtabscale);
2025 ewitab = _mm_cvttps_epi32(ewrt);
2026 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
2027 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
2028 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
2030 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
2031 felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec));
2035 fscal = _mm_andnot_ps(dummy_mask,fscal);
2037 /* Calculate temporary vectorial force */
2038 tx = _mm_mul_ps(fscal,dx22);
2039 ty = _mm_mul_ps(fscal,dy22);
2040 tz = _mm_mul_ps(fscal,dz22);
2042 /* Update vectorial force */
2043 fix2 = _mm_add_ps(fix2,tx);
2044 fiy2 = _mm_add_ps(fiy2,ty);
2045 fiz2 = _mm_add_ps(fiz2,tz);
2047 fjx2 = _mm_add_ps(fjx2,tx);
2048 fjy2 = _mm_add_ps(fjy2,ty);
2049 fjz2 = _mm_add_ps(fjz2,tz);
2051 /**************************
2052 * CALCULATE INTERACTIONS *
2053 **************************/
2055 r23 = _mm_mul_ps(rsq23,rinv23);
2056 r23 = _mm_andnot_ps(dummy_mask,r23);
2058 /* EWALD ELECTROSTATICS */
2060 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2061 ewrt = _mm_mul_ps(r23,ewtabscale);
2062 ewitab = _mm_cvttps_epi32(ewrt);
2063 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
2064 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
2065 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
2067 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
2068 felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec));
2072 fscal = _mm_andnot_ps(dummy_mask,fscal);
2074 /* Calculate temporary vectorial force */
2075 tx = _mm_mul_ps(fscal,dx23);
2076 ty = _mm_mul_ps(fscal,dy23);
2077 tz = _mm_mul_ps(fscal,dz23);
2079 /* Update vectorial force */
2080 fix2 = _mm_add_ps(fix2,tx);
2081 fiy2 = _mm_add_ps(fiy2,ty);
2082 fiz2 = _mm_add_ps(fiz2,tz);
2084 fjx3 = _mm_add_ps(fjx3,tx);
2085 fjy3 = _mm_add_ps(fjy3,ty);
2086 fjz3 = _mm_add_ps(fjz3,tz);
2088 /**************************
2089 * CALCULATE INTERACTIONS *
2090 **************************/
2092 r31 = _mm_mul_ps(rsq31,rinv31);
2093 r31 = _mm_andnot_ps(dummy_mask,r31);
2095 /* EWALD ELECTROSTATICS */
2097 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2098 ewrt = _mm_mul_ps(r31,ewtabscale);
2099 ewitab = _mm_cvttps_epi32(ewrt);
2100 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
2101 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
2102 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
2104 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
2105 felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec));
2109 fscal = _mm_andnot_ps(dummy_mask,fscal);
2111 /* Calculate temporary vectorial force */
2112 tx = _mm_mul_ps(fscal,dx31);
2113 ty = _mm_mul_ps(fscal,dy31);
2114 tz = _mm_mul_ps(fscal,dz31);
2116 /* Update vectorial force */
2117 fix3 = _mm_add_ps(fix3,tx);
2118 fiy3 = _mm_add_ps(fiy3,ty);
2119 fiz3 = _mm_add_ps(fiz3,tz);
2121 fjx1 = _mm_add_ps(fjx1,tx);
2122 fjy1 = _mm_add_ps(fjy1,ty);
2123 fjz1 = _mm_add_ps(fjz1,tz);
2125 /**************************
2126 * CALCULATE INTERACTIONS *
2127 **************************/
2129 r32 = _mm_mul_ps(rsq32,rinv32);
2130 r32 = _mm_andnot_ps(dummy_mask,r32);
2132 /* EWALD ELECTROSTATICS */
2134 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2135 ewrt = _mm_mul_ps(r32,ewtabscale);
2136 ewitab = _mm_cvttps_epi32(ewrt);
2137 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
2138 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
2139 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
2141 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
2142 felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec));
2146 fscal = _mm_andnot_ps(dummy_mask,fscal);
2148 /* Calculate temporary vectorial force */
2149 tx = _mm_mul_ps(fscal,dx32);
2150 ty = _mm_mul_ps(fscal,dy32);
2151 tz = _mm_mul_ps(fscal,dz32);
2153 /* Update vectorial force */
2154 fix3 = _mm_add_ps(fix3,tx);
2155 fiy3 = _mm_add_ps(fiy3,ty);
2156 fiz3 = _mm_add_ps(fiz3,tz);
2158 fjx2 = _mm_add_ps(fjx2,tx);
2159 fjy2 = _mm_add_ps(fjy2,ty);
2160 fjz2 = _mm_add_ps(fjz2,tz);
2162 /**************************
2163 * CALCULATE INTERACTIONS *
2164 **************************/
2166 r33 = _mm_mul_ps(rsq33,rinv33);
2167 r33 = _mm_andnot_ps(dummy_mask,r33);
2169 /* EWALD ELECTROSTATICS */
2171 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2172 ewrt = _mm_mul_ps(r33,ewtabscale);
2173 ewitab = _mm_cvttps_epi32(ewrt);
2174 eweps = _mm_sub_ps(ewrt,_mm_cvtepi32_ps(ewitab));
2175 gmx_mm_load_4pair_swizzle_ps(ewtab+gmx_mm_extract_epi32(ewitab,0),ewtab+gmx_mm_extract_epi32(ewitab,1),
2176 ewtab+gmx_mm_extract_epi32(ewitab,2),ewtab+gmx_mm_extract_epi32(ewitab,3),
2178 felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
2179 felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec));
2183 fscal = _mm_andnot_ps(dummy_mask,fscal);
2185 /* Calculate temporary vectorial force */
2186 tx = _mm_mul_ps(fscal,dx33);
2187 ty = _mm_mul_ps(fscal,dy33);
2188 tz = _mm_mul_ps(fscal,dz33);
2190 /* Update vectorial force */
2191 fix3 = _mm_add_ps(fix3,tx);
2192 fiy3 = _mm_add_ps(fiy3,ty);
2193 fiz3 = _mm_add_ps(fiz3,tz);
2195 fjx3 = _mm_add_ps(fjx3,tx);
2196 fjy3 = _mm_add_ps(fjy3,ty);
2197 fjz3 = _mm_add_ps(fjz3,tz);
2199 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2200 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2201 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2202 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2204 gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
2205 fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
2207 /* Inner loop uses 333 flops */
2210 /* End of innermost loop */
2212 gmx_mm_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
2213 f+i_coord_offset+DIM,fshift+i_shift_offset);
2215 /* Increment number of inner iterations */
2216 inneriter += j_index_end - j_index_start;
2218 /* Outer loop uses 18 flops */
2221 /* Increment number of outer iterations */
2224 /* Update outer/inner flops */
2226 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_F,outeriter*18 + inneriter*333);