File: | gromacs/gmxlib/nonbonded/nb_kernel_sse4_1_single/nb_kernel_ElecEw_VdwLJ_GeomW3W3_sse4_1_single.c |
Location: | line 168, column 5 |
Description: | Value stored to 'j_coord_offsetA' is never read |
1 | /* |
2 | * This file is part of the GROMACS molecular simulation package. |
3 | * |
4 | * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by |
5 | * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl, |
6 | * and including many others, as listed in the AUTHORS file in the |
7 | * top-level source directory and at http://www.gromacs.org. |
8 | * |
9 | * GROMACS is free software; you can redistribute it and/or |
10 | * modify it under the terms of the GNU Lesser General Public License |
11 | * as published by the Free Software Foundation; either version 2.1 |
12 | * of the License, or (at your option) any later version. |
13 | * |
14 | * GROMACS is distributed in the hope that it will be useful, |
15 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
16 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
17 | * Lesser General Public License for more details. |
18 | * |
19 | * You should have received a copy of the GNU Lesser General Public |
20 | * License along with GROMACS; if not, see |
21 | * http://www.gnu.org/licenses, or write to the Free Software Foundation, |
22 | * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. |
23 | * |
24 | * If you want to redistribute modifications to GROMACS, please |
25 | * consider that scientific software is very special. Version |
26 | * control is crucial - bugs must be traceable. We will be happy to |
27 | * consider code for inclusion in the official distribution, but |
28 | * derived work must not be called official GROMACS. Details are found |
29 | * in the README & COPYING files - if they are missing, get the |
30 | * official version at http://www.gromacs.org. |
31 | * |
32 | * To help us fund GROMACS development, we humbly ask that you cite |
33 | * the research papers on the package. Check out http://www.gromacs.org. |
34 | */ |
35 | /* |
36 | * Note: this file was generated by the GROMACS sse4_1_single kernel generator. |
37 | */ |
38 | #ifdef HAVE_CONFIG_H1 |
39 | #include <config.h> |
40 | #endif |
41 | |
42 | #include <math.h> |
43 | |
44 | #include "../nb_kernel.h" |
45 | #include "types/simple.h" |
46 | #include "gromacs/math/vec.h" |
47 | #include "nrnb.h" |
48 | |
49 | #include "gromacs/simd/math_x86_sse4_1_single.h" |
50 | #include "kernelutil_x86_sse4_1_single.h" |
51 | |
52 | /* |
53 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJ_GeomW3W3_VF_sse4_1_single |
54 | * Electrostatics interaction: Ewald |
55 | * VdW interaction: LennardJones |
56 | * Geometry: Water3-Water3 |
57 | * Calculate force/pot: PotentialAndForce |
58 | */ |
59 | void |
60 | nb_kernel_ElecEw_VdwLJ_GeomW3W3_VF_sse4_1_single |
61 | (t_nblist * gmx_restrict nlist, |
62 | rvec * gmx_restrict xx, |
63 | rvec * gmx_restrict ff, |
64 | t_forcerec * gmx_restrict fr, |
65 | t_mdatoms * gmx_restrict mdatoms, |
66 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict kernel_data, |
67 | t_nrnb * gmx_restrict nrnb) |
68 | { |
69 | /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or |
70 | * just 0 for non-waters. |
71 | * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different |
72 | * jnr indices corresponding to data put in the four positions in the SIMD register. |
73 | */ |
74 | int i_shift_offset,i_coord_offset,outeriter,inneriter; |
75 | int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx; |
76 | int jnrA,jnrB,jnrC,jnrD; |
77 | int jnrlistA,jnrlistB,jnrlistC,jnrlistD; |
78 | int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD; |
79 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
80 | real rcutoff_scalar; |
81 | real *shiftvec,*fshift,*x,*f; |
82 | real *fjptrA,*fjptrB,*fjptrC,*fjptrD; |
83 | real scratch[4*DIM3]; |
84 | __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall; |
85 | int vdwioffset0; |
86 | __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
87 | int vdwioffset1; |
88 | __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
89 | int vdwioffset2; |
90 | __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
91 | int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D; |
92 | __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
93 | int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D; |
94 | __m128 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
95 | int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D; |
96 | __m128 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
97 | __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00; |
98 | __m128 dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01; |
99 | __m128 dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02; |
100 | __m128 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10; |
101 | __m128 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11; |
102 | __m128 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12; |
103 | __m128 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20; |
104 | __m128 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21; |
105 | __m128 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22; |
106 | __m128 velec,felec,velecsum,facel,crf,krf,krf2; |
107 | real *charge; |
108 | int nvdwtype; |
109 | __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6; |
110 | int *vdwtype; |
111 | real *vdwparam; |
112 | __m128 one_sixth = _mm_set1_ps(1.0/6.0); |
113 | __m128 one_twelfth = _mm_set1_ps(1.0/12.0); |
114 | __m128i ewitab; |
115 | __m128 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV; |
116 | real *ewtab; |
117 | __m128 dummy_mask,cutoff_mask; |
118 | __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) ); |
119 | __m128 one = _mm_set1_ps(1.0); |
120 | __m128 two = _mm_set1_ps(2.0); |
121 | x = xx[0]; |
122 | f = ff[0]; |
123 | |
124 | nri = nlist->nri; |
125 | iinr = nlist->iinr; |
126 | jindex = nlist->jindex; |
127 | jjnr = nlist->jjnr; |
128 | shiftidx = nlist->shift; |
129 | gid = nlist->gid; |
130 | shiftvec = fr->shift_vec[0]; |
131 | fshift = fr->fshift[0]; |
132 | facel = _mm_set1_ps(fr->epsfac); |
133 | charge = mdatoms->chargeA; |
134 | nvdwtype = fr->ntype; |
135 | vdwparam = fr->nbfp; |
136 | vdwtype = mdatoms->typeA; |
137 | |
138 | sh_ewald = _mm_set1_ps(fr->ic->sh_ewald); |
139 | ewtab = fr->ic->tabq_coul_FDV0; |
140 | ewtabscale = _mm_set1_ps(fr->ic->tabq_scale); |
141 | ewtabhalfspace = _mm_set1_ps(0.5/fr->ic->tabq_scale); |
142 | |
143 | /* Setup water-specific parameters */ |
144 | inr = nlist->iinr[0]; |
145 | iq0 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+0])); |
146 | iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1])); |
147 | iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2])); |
148 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
149 | |
150 | jq0 = _mm_set1_ps(charge[inr+0]); |
151 | jq1 = _mm_set1_ps(charge[inr+1]); |
152 | jq2 = _mm_set1_ps(charge[inr+2]); |
153 | vdwjidx0A = 2*vdwtype[inr+0]; |
154 | qq00 = _mm_mul_ps(iq0,jq0); |
155 | c6_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A]); |
156 | c12_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A+1]); |
157 | qq01 = _mm_mul_ps(iq0,jq1); |
158 | qq02 = _mm_mul_ps(iq0,jq2); |
159 | qq10 = _mm_mul_ps(iq1,jq0); |
160 | qq11 = _mm_mul_ps(iq1,jq1); |
161 | qq12 = _mm_mul_ps(iq1,jq2); |
162 | qq20 = _mm_mul_ps(iq2,jq0); |
163 | qq21 = _mm_mul_ps(iq2,jq1); |
164 | qq22 = _mm_mul_ps(iq2,jq2); |
165 | |
166 | /* Avoid stupid compiler warnings */ |
167 | jnrA = jnrB = jnrC = jnrD = 0; |
168 | j_coord_offsetA = 0; |
Value stored to 'j_coord_offsetA' is never read | |
169 | j_coord_offsetB = 0; |
170 | j_coord_offsetC = 0; |
171 | j_coord_offsetD = 0; |
172 | |
173 | outeriter = 0; |
174 | inneriter = 0; |
175 | |
176 | for(iidx=0;iidx<4*DIM3;iidx++) |
177 | { |
178 | scratch[iidx] = 0.0; |
179 | } |
180 | |
181 | /* Start outer loop over neighborlists */ |
182 | for(iidx=0; iidx<nri; iidx++) |
183 | { |
184 | /* Load shift vector for this list */ |
185 | i_shift_offset = DIM3*shiftidx[iidx]; |
186 | |
187 | /* Load limits for loop over neighbors */ |
188 | j_index_start = jindex[iidx]; |
189 | j_index_end = jindex[iidx+1]; |
190 | |
191 | /* Get outer coordinate index */ |
192 | inr = iinr[iidx]; |
193 | i_coord_offset = DIM3*inr; |
194 | |
195 | /* Load i particle coords and add shift vector */ |
196 | gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset, |
197 | &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2); |
198 | |
199 | fix0 = _mm_setzero_ps(); |
200 | fiy0 = _mm_setzero_ps(); |
201 | fiz0 = _mm_setzero_ps(); |
202 | fix1 = _mm_setzero_ps(); |
203 | fiy1 = _mm_setzero_ps(); |
204 | fiz1 = _mm_setzero_ps(); |
205 | fix2 = _mm_setzero_ps(); |
206 | fiy2 = _mm_setzero_ps(); |
207 | fiz2 = _mm_setzero_ps(); |
208 | |
209 | /* Reset potential sums */ |
210 | velecsum = _mm_setzero_ps(); |
211 | vvdwsum = _mm_setzero_ps(); |
212 | |
213 | /* Start inner kernel loop */ |
214 | for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4) |
215 | { |
216 | |
217 | /* Get j neighbor index, and coordinate index */ |
218 | jnrA = jjnr[jidx]; |
219 | jnrB = jjnr[jidx+1]; |
220 | jnrC = jjnr[jidx+2]; |
221 | jnrD = jjnr[jidx+3]; |
222 | j_coord_offsetA = DIM3*jnrA; |
223 | j_coord_offsetB = DIM3*jnrB; |
224 | j_coord_offsetC = DIM3*jnrC; |
225 | j_coord_offsetD = DIM3*jnrD; |
226 | |
227 | /* load j atom coordinates */ |
228 | gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
229 | x+j_coord_offsetC,x+j_coord_offsetD, |
230 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2); |
231 | |
232 | /* Calculate displacement vector */ |
233 | dx00 = _mm_sub_ps(ix0,jx0); |
234 | dy00 = _mm_sub_ps(iy0,jy0); |
235 | dz00 = _mm_sub_ps(iz0,jz0); |
236 | dx01 = _mm_sub_ps(ix0,jx1); |
237 | dy01 = _mm_sub_ps(iy0,jy1); |
238 | dz01 = _mm_sub_ps(iz0,jz1); |
239 | dx02 = _mm_sub_ps(ix0,jx2); |
240 | dy02 = _mm_sub_ps(iy0,jy2); |
241 | dz02 = _mm_sub_ps(iz0,jz2); |
242 | dx10 = _mm_sub_ps(ix1,jx0); |
243 | dy10 = _mm_sub_ps(iy1,jy0); |
244 | dz10 = _mm_sub_ps(iz1,jz0); |
245 | dx11 = _mm_sub_ps(ix1,jx1); |
246 | dy11 = _mm_sub_ps(iy1,jy1); |
247 | dz11 = _mm_sub_ps(iz1,jz1); |
248 | dx12 = _mm_sub_ps(ix1,jx2); |
249 | dy12 = _mm_sub_ps(iy1,jy2); |
250 | dz12 = _mm_sub_ps(iz1,jz2); |
251 | dx20 = _mm_sub_ps(ix2,jx0); |
252 | dy20 = _mm_sub_ps(iy2,jy0); |
253 | dz20 = _mm_sub_ps(iz2,jz0); |
254 | dx21 = _mm_sub_ps(ix2,jx1); |
255 | dy21 = _mm_sub_ps(iy2,jy1); |
256 | dz21 = _mm_sub_ps(iz2,jz1); |
257 | dx22 = _mm_sub_ps(ix2,jx2); |
258 | dy22 = _mm_sub_ps(iy2,jy2); |
259 | dz22 = _mm_sub_ps(iz2,jz2); |
260 | |
261 | /* Calculate squared distance and things based on it */ |
262 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
263 | rsq01 = gmx_mm_calc_rsq_ps(dx01,dy01,dz01); |
264 | rsq02 = gmx_mm_calc_rsq_ps(dx02,dy02,dz02); |
265 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
266 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
267 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
268 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
269 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
270 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
271 | |
272 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
273 | rinv01 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq01); |
274 | rinv02 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq02); |
275 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
276 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
277 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
278 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
279 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
280 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
281 | |
282 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
283 | rinvsq01 = _mm_mul_ps(rinv01,rinv01); |
284 | rinvsq02 = _mm_mul_ps(rinv02,rinv02); |
285 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
286 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
287 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
288 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
289 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
290 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
291 | |
292 | fjx0 = _mm_setzero_ps(); |
293 | fjy0 = _mm_setzero_ps(); |
294 | fjz0 = _mm_setzero_ps(); |
295 | fjx1 = _mm_setzero_ps(); |
296 | fjy1 = _mm_setzero_ps(); |
297 | fjz1 = _mm_setzero_ps(); |
298 | fjx2 = _mm_setzero_ps(); |
299 | fjy2 = _mm_setzero_ps(); |
300 | fjz2 = _mm_setzero_ps(); |
301 | |
302 | /************************** |
303 | * CALCULATE INTERACTIONS * |
304 | **************************/ |
305 | |
306 | r00 = _mm_mul_ps(rsq00,rinv00); |
307 | |
308 | /* EWALD ELECTROSTATICS */ |
309 | |
310 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
311 | ewrt = _mm_mul_ps(r00,ewtabscale); |
312 | ewitab = _mm_cvttps_epi32(ewrt); |
313 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
314 | ewitab = _mm_slli_epi32(ewitab,2); |
315 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
316 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
317 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
318 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
319 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
320 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
321 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
322 | velec = _mm_mul_ps(qq00,_mm_sub_ps(rinv00,velec)); |
323 | felec = _mm_mul_ps(_mm_mul_ps(qq00,rinv00),_mm_sub_ps(rinvsq00,felec)); |
324 | |
325 | /* LENNARD-JONES DISPERSION/REPULSION */ |
326 | |
327 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
328 | vvdw6 = _mm_mul_ps(c6_00,rinvsix); |
329 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
330 | vvdw = _mm_sub_ps( _mm_mul_ps(vvdw12,one_twelfth) , _mm_mul_ps(vvdw6,one_sixth) ); |
331 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00); |
332 | |
333 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
334 | velecsum = _mm_add_ps(velecsum,velec); |
335 | vvdwsum = _mm_add_ps(vvdwsum,vvdw); |
336 | |
337 | fscal = _mm_add_ps(felec,fvdw); |
338 | |
339 | /* Calculate temporary vectorial force */ |
340 | tx = _mm_mul_ps(fscal,dx00); |
341 | ty = _mm_mul_ps(fscal,dy00); |
342 | tz = _mm_mul_ps(fscal,dz00); |
343 | |
344 | /* Update vectorial force */ |
345 | fix0 = _mm_add_ps(fix0,tx); |
346 | fiy0 = _mm_add_ps(fiy0,ty); |
347 | fiz0 = _mm_add_ps(fiz0,tz); |
348 | |
349 | fjx0 = _mm_add_ps(fjx0,tx); |
350 | fjy0 = _mm_add_ps(fjy0,ty); |
351 | fjz0 = _mm_add_ps(fjz0,tz); |
352 | |
353 | /************************** |
354 | * CALCULATE INTERACTIONS * |
355 | **************************/ |
356 | |
357 | r01 = _mm_mul_ps(rsq01,rinv01); |
358 | |
359 | /* EWALD ELECTROSTATICS */ |
360 | |
361 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
362 | ewrt = _mm_mul_ps(r01,ewtabscale); |
363 | ewitab = _mm_cvttps_epi32(ewrt); |
364 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
365 | ewitab = _mm_slli_epi32(ewitab,2); |
366 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
367 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
368 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
369 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
370 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
371 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
372 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
373 | velec = _mm_mul_ps(qq01,_mm_sub_ps(rinv01,velec)); |
374 | felec = _mm_mul_ps(_mm_mul_ps(qq01,rinv01),_mm_sub_ps(rinvsq01,felec)); |
375 | |
376 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
377 | velecsum = _mm_add_ps(velecsum,velec); |
378 | |
379 | fscal = felec; |
380 | |
381 | /* Calculate temporary vectorial force */ |
382 | tx = _mm_mul_ps(fscal,dx01); |
383 | ty = _mm_mul_ps(fscal,dy01); |
384 | tz = _mm_mul_ps(fscal,dz01); |
385 | |
386 | /* Update vectorial force */ |
387 | fix0 = _mm_add_ps(fix0,tx); |
388 | fiy0 = _mm_add_ps(fiy0,ty); |
389 | fiz0 = _mm_add_ps(fiz0,tz); |
390 | |
391 | fjx1 = _mm_add_ps(fjx1,tx); |
392 | fjy1 = _mm_add_ps(fjy1,ty); |
393 | fjz1 = _mm_add_ps(fjz1,tz); |
394 | |
395 | /************************** |
396 | * CALCULATE INTERACTIONS * |
397 | **************************/ |
398 | |
399 | r02 = _mm_mul_ps(rsq02,rinv02); |
400 | |
401 | /* EWALD ELECTROSTATICS */ |
402 | |
403 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
404 | ewrt = _mm_mul_ps(r02,ewtabscale); |
405 | ewitab = _mm_cvttps_epi32(ewrt); |
406 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
407 | ewitab = _mm_slli_epi32(ewitab,2); |
408 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
409 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
410 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
411 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
412 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
413 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
414 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
415 | velec = _mm_mul_ps(qq02,_mm_sub_ps(rinv02,velec)); |
416 | felec = _mm_mul_ps(_mm_mul_ps(qq02,rinv02),_mm_sub_ps(rinvsq02,felec)); |
417 | |
418 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
419 | velecsum = _mm_add_ps(velecsum,velec); |
420 | |
421 | fscal = felec; |
422 | |
423 | /* Calculate temporary vectorial force */ |
424 | tx = _mm_mul_ps(fscal,dx02); |
425 | ty = _mm_mul_ps(fscal,dy02); |
426 | tz = _mm_mul_ps(fscal,dz02); |
427 | |
428 | /* Update vectorial force */ |
429 | fix0 = _mm_add_ps(fix0,tx); |
430 | fiy0 = _mm_add_ps(fiy0,ty); |
431 | fiz0 = _mm_add_ps(fiz0,tz); |
432 | |
433 | fjx2 = _mm_add_ps(fjx2,tx); |
434 | fjy2 = _mm_add_ps(fjy2,ty); |
435 | fjz2 = _mm_add_ps(fjz2,tz); |
436 | |
437 | /************************** |
438 | * CALCULATE INTERACTIONS * |
439 | **************************/ |
440 | |
441 | r10 = _mm_mul_ps(rsq10,rinv10); |
442 | |
443 | /* EWALD ELECTROSTATICS */ |
444 | |
445 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
446 | ewrt = _mm_mul_ps(r10,ewtabscale); |
447 | ewitab = _mm_cvttps_epi32(ewrt); |
448 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
449 | ewitab = _mm_slli_epi32(ewitab,2); |
450 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
451 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
452 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
453 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
454 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
455 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
456 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
457 | velec = _mm_mul_ps(qq10,_mm_sub_ps(rinv10,velec)); |
458 | felec = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec)); |
459 | |
460 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
461 | velecsum = _mm_add_ps(velecsum,velec); |
462 | |
463 | fscal = felec; |
464 | |
465 | /* Calculate temporary vectorial force */ |
466 | tx = _mm_mul_ps(fscal,dx10); |
467 | ty = _mm_mul_ps(fscal,dy10); |
468 | tz = _mm_mul_ps(fscal,dz10); |
469 | |
470 | /* Update vectorial force */ |
471 | fix1 = _mm_add_ps(fix1,tx); |
472 | fiy1 = _mm_add_ps(fiy1,ty); |
473 | fiz1 = _mm_add_ps(fiz1,tz); |
474 | |
475 | fjx0 = _mm_add_ps(fjx0,tx); |
476 | fjy0 = _mm_add_ps(fjy0,ty); |
477 | fjz0 = _mm_add_ps(fjz0,tz); |
478 | |
479 | /************************** |
480 | * CALCULATE INTERACTIONS * |
481 | **************************/ |
482 | |
483 | r11 = _mm_mul_ps(rsq11,rinv11); |
484 | |
485 | /* EWALD ELECTROSTATICS */ |
486 | |
487 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
488 | ewrt = _mm_mul_ps(r11,ewtabscale); |
489 | ewitab = _mm_cvttps_epi32(ewrt); |
490 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
491 | ewitab = _mm_slli_epi32(ewitab,2); |
492 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
493 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
494 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
495 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
496 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
497 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
498 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
499 | velec = _mm_mul_ps(qq11,_mm_sub_ps(rinv11,velec)); |
500 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
501 | |
502 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
503 | velecsum = _mm_add_ps(velecsum,velec); |
504 | |
505 | fscal = felec; |
506 | |
507 | /* Calculate temporary vectorial force */ |
508 | tx = _mm_mul_ps(fscal,dx11); |
509 | ty = _mm_mul_ps(fscal,dy11); |
510 | tz = _mm_mul_ps(fscal,dz11); |
511 | |
512 | /* Update vectorial force */ |
513 | fix1 = _mm_add_ps(fix1,tx); |
514 | fiy1 = _mm_add_ps(fiy1,ty); |
515 | fiz1 = _mm_add_ps(fiz1,tz); |
516 | |
517 | fjx1 = _mm_add_ps(fjx1,tx); |
518 | fjy1 = _mm_add_ps(fjy1,ty); |
519 | fjz1 = _mm_add_ps(fjz1,tz); |
520 | |
521 | /************************** |
522 | * CALCULATE INTERACTIONS * |
523 | **************************/ |
524 | |
525 | r12 = _mm_mul_ps(rsq12,rinv12); |
526 | |
527 | /* EWALD ELECTROSTATICS */ |
528 | |
529 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
530 | ewrt = _mm_mul_ps(r12,ewtabscale); |
531 | ewitab = _mm_cvttps_epi32(ewrt); |
532 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
533 | ewitab = _mm_slli_epi32(ewitab,2); |
534 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
535 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
536 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
537 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
538 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
539 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
540 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
541 | velec = _mm_mul_ps(qq12,_mm_sub_ps(rinv12,velec)); |
542 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
543 | |
544 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
545 | velecsum = _mm_add_ps(velecsum,velec); |
546 | |
547 | fscal = felec; |
548 | |
549 | /* Calculate temporary vectorial force */ |
550 | tx = _mm_mul_ps(fscal,dx12); |
551 | ty = _mm_mul_ps(fscal,dy12); |
552 | tz = _mm_mul_ps(fscal,dz12); |
553 | |
554 | /* Update vectorial force */ |
555 | fix1 = _mm_add_ps(fix1,tx); |
556 | fiy1 = _mm_add_ps(fiy1,ty); |
557 | fiz1 = _mm_add_ps(fiz1,tz); |
558 | |
559 | fjx2 = _mm_add_ps(fjx2,tx); |
560 | fjy2 = _mm_add_ps(fjy2,ty); |
561 | fjz2 = _mm_add_ps(fjz2,tz); |
562 | |
563 | /************************** |
564 | * CALCULATE INTERACTIONS * |
565 | **************************/ |
566 | |
567 | r20 = _mm_mul_ps(rsq20,rinv20); |
568 | |
569 | /* EWALD ELECTROSTATICS */ |
570 | |
571 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
572 | ewrt = _mm_mul_ps(r20,ewtabscale); |
573 | ewitab = _mm_cvttps_epi32(ewrt); |
574 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
575 | ewitab = _mm_slli_epi32(ewitab,2); |
576 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
577 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
578 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
579 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
580 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
581 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
582 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
583 | velec = _mm_mul_ps(qq20,_mm_sub_ps(rinv20,velec)); |
584 | felec = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec)); |
585 | |
586 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
587 | velecsum = _mm_add_ps(velecsum,velec); |
588 | |
589 | fscal = felec; |
590 | |
591 | /* Calculate temporary vectorial force */ |
592 | tx = _mm_mul_ps(fscal,dx20); |
593 | ty = _mm_mul_ps(fscal,dy20); |
594 | tz = _mm_mul_ps(fscal,dz20); |
595 | |
596 | /* Update vectorial force */ |
597 | fix2 = _mm_add_ps(fix2,tx); |
598 | fiy2 = _mm_add_ps(fiy2,ty); |
599 | fiz2 = _mm_add_ps(fiz2,tz); |
600 | |
601 | fjx0 = _mm_add_ps(fjx0,tx); |
602 | fjy0 = _mm_add_ps(fjy0,ty); |
603 | fjz0 = _mm_add_ps(fjz0,tz); |
604 | |
605 | /************************** |
606 | * CALCULATE INTERACTIONS * |
607 | **************************/ |
608 | |
609 | r21 = _mm_mul_ps(rsq21,rinv21); |
610 | |
611 | /* EWALD ELECTROSTATICS */ |
612 | |
613 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
614 | ewrt = _mm_mul_ps(r21,ewtabscale); |
615 | ewitab = _mm_cvttps_epi32(ewrt); |
616 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
617 | ewitab = _mm_slli_epi32(ewitab,2); |
618 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
619 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
620 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
621 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
622 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
623 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
624 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
625 | velec = _mm_mul_ps(qq21,_mm_sub_ps(rinv21,velec)); |
626 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
627 | |
628 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
629 | velecsum = _mm_add_ps(velecsum,velec); |
630 | |
631 | fscal = felec; |
632 | |
633 | /* Calculate temporary vectorial force */ |
634 | tx = _mm_mul_ps(fscal,dx21); |
635 | ty = _mm_mul_ps(fscal,dy21); |
636 | tz = _mm_mul_ps(fscal,dz21); |
637 | |
638 | /* Update vectorial force */ |
639 | fix2 = _mm_add_ps(fix2,tx); |
640 | fiy2 = _mm_add_ps(fiy2,ty); |
641 | fiz2 = _mm_add_ps(fiz2,tz); |
642 | |
643 | fjx1 = _mm_add_ps(fjx1,tx); |
644 | fjy1 = _mm_add_ps(fjy1,ty); |
645 | fjz1 = _mm_add_ps(fjz1,tz); |
646 | |
647 | /************************** |
648 | * CALCULATE INTERACTIONS * |
649 | **************************/ |
650 | |
651 | r22 = _mm_mul_ps(rsq22,rinv22); |
652 | |
653 | /* EWALD ELECTROSTATICS */ |
654 | |
655 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
656 | ewrt = _mm_mul_ps(r22,ewtabscale); |
657 | ewitab = _mm_cvttps_epi32(ewrt); |
658 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
659 | ewitab = _mm_slli_epi32(ewitab,2); |
660 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
661 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
662 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
663 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
664 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
665 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
666 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
667 | velec = _mm_mul_ps(qq22,_mm_sub_ps(rinv22,velec)); |
668 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
669 | |
670 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
671 | velecsum = _mm_add_ps(velecsum,velec); |
672 | |
673 | fscal = felec; |
674 | |
675 | /* Calculate temporary vectorial force */ |
676 | tx = _mm_mul_ps(fscal,dx22); |
677 | ty = _mm_mul_ps(fscal,dy22); |
678 | tz = _mm_mul_ps(fscal,dz22); |
679 | |
680 | /* Update vectorial force */ |
681 | fix2 = _mm_add_ps(fix2,tx); |
682 | fiy2 = _mm_add_ps(fiy2,ty); |
683 | fiz2 = _mm_add_ps(fiz2,tz); |
684 | |
685 | fjx2 = _mm_add_ps(fjx2,tx); |
686 | fjy2 = _mm_add_ps(fjy2,ty); |
687 | fjz2 = _mm_add_ps(fjz2,tz); |
688 | |
689 | fjptrA = f+j_coord_offsetA; |
690 | fjptrB = f+j_coord_offsetB; |
691 | fjptrC = f+j_coord_offsetC; |
692 | fjptrD = f+j_coord_offsetD; |
693 | |
694 | gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
695 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2); |
696 | |
697 | /* Inner loop uses 381 flops */ |
698 | } |
699 | |
700 | if(jidx<j_index_end) |
701 | { |
702 | |
703 | /* Get j neighbor index, and coordinate index */ |
704 | jnrlistA = jjnr[jidx]; |
705 | jnrlistB = jjnr[jidx+1]; |
706 | jnrlistC = jjnr[jidx+2]; |
707 | jnrlistD = jjnr[jidx+3]; |
708 | /* Sign of each element will be negative for non-real atoms. |
709 | * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones, |
710 | * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries. |
711 | */ |
712 | dummy_mask = gmx_mm_castsi128_ps_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())); |
713 | jnrA = (jnrlistA>=0) ? jnrlistA : 0; |
714 | jnrB = (jnrlistB>=0) ? jnrlistB : 0; |
715 | jnrC = (jnrlistC>=0) ? jnrlistC : 0; |
716 | jnrD = (jnrlistD>=0) ? jnrlistD : 0; |
717 | j_coord_offsetA = DIM3*jnrA; |
718 | j_coord_offsetB = DIM3*jnrB; |
719 | j_coord_offsetC = DIM3*jnrC; |
720 | j_coord_offsetD = DIM3*jnrD; |
721 | |
722 | /* load j atom coordinates */ |
723 | gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
724 | x+j_coord_offsetC,x+j_coord_offsetD, |
725 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2); |
726 | |
727 | /* Calculate displacement vector */ |
728 | dx00 = _mm_sub_ps(ix0,jx0); |
729 | dy00 = _mm_sub_ps(iy0,jy0); |
730 | dz00 = _mm_sub_ps(iz0,jz0); |
731 | dx01 = _mm_sub_ps(ix0,jx1); |
732 | dy01 = _mm_sub_ps(iy0,jy1); |
733 | dz01 = _mm_sub_ps(iz0,jz1); |
734 | dx02 = _mm_sub_ps(ix0,jx2); |
735 | dy02 = _mm_sub_ps(iy0,jy2); |
736 | dz02 = _mm_sub_ps(iz0,jz2); |
737 | dx10 = _mm_sub_ps(ix1,jx0); |
738 | dy10 = _mm_sub_ps(iy1,jy0); |
739 | dz10 = _mm_sub_ps(iz1,jz0); |
740 | dx11 = _mm_sub_ps(ix1,jx1); |
741 | dy11 = _mm_sub_ps(iy1,jy1); |
742 | dz11 = _mm_sub_ps(iz1,jz1); |
743 | dx12 = _mm_sub_ps(ix1,jx2); |
744 | dy12 = _mm_sub_ps(iy1,jy2); |
745 | dz12 = _mm_sub_ps(iz1,jz2); |
746 | dx20 = _mm_sub_ps(ix2,jx0); |
747 | dy20 = _mm_sub_ps(iy2,jy0); |
748 | dz20 = _mm_sub_ps(iz2,jz0); |
749 | dx21 = _mm_sub_ps(ix2,jx1); |
750 | dy21 = _mm_sub_ps(iy2,jy1); |
751 | dz21 = _mm_sub_ps(iz2,jz1); |
752 | dx22 = _mm_sub_ps(ix2,jx2); |
753 | dy22 = _mm_sub_ps(iy2,jy2); |
754 | dz22 = _mm_sub_ps(iz2,jz2); |
755 | |
756 | /* Calculate squared distance and things based on it */ |
757 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
758 | rsq01 = gmx_mm_calc_rsq_ps(dx01,dy01,dz01); |
759 | rsq02 = gmx_mm_calc_rsq_ps(dx02,dy02,dz02); |
760 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
761 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
762 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
763 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
764 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
765 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
766 | |
767 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
768 | rinv01 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq01); |
769 | rinv02 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq02); |
770 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
771 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
772 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
773 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
774 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
775 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
776 | |
777 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
778 | rinvsq01 = _mm_mul_ps(rinv01,rinv01); |
779 | rinvsq02 = _mm_mul_ps(rinv02,rinv02); |
780 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
781 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
782 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
783 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
784 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
785 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
786 | |
787 | fjx0 = _mm_setzero_ps(); |
788 | fjy0 = _mm_setzero_ps(); |
789 | fjz0 = _mm_setzero_ps(); |
790 | fjx1 = _mm_setzero_ps(); |
791 | fjy1 = _mm_setzero_ps(); |
792 | fjz1 = _mm_setzero_ps(); |
793 | fjx2 = _mm_setzero_ps(); |
794 | fjy2 = _mm_setzero_ps(); |
795 | fjz2 = _mm_setzero_ps(); |
796 | |
797 | /************************** |
798 | * CALCULATE INTERACTIONS * |
799 | **************************/ |
800 | |
801 | r00 = _mm_mul_ps(rsq00,rinv00); |
802 | r00 = _mm_andnot_ps(dummy_mask,r00); |
803 | |
804 | /* EWALD ELECTROSTATICS */ |
805 | |
806 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
807 | ewrt = _mm_mul_ps(r00,ewtabscale); |
808 | ewitab = _mm_cvttps_epi32(ewrt); |
809 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
810 | ewitab = _mm_slli_epi32(ewitab,2); |
811 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
812 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
813 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
814 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
815 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
816 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
817 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
818 | velec = _mm_mul_ps(qq00,_mm_sub_ps(rinv00,velec)); |
819 | felec = _mm_mul_ps(_mm_mul_ps(qq00,rinv00),_mm_sub_ps(rinvsq00,felec)); |
820 | |
821 | /* LENNARD-JONES DISPERSION/REPULSION */ |
822 | |
823 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
824 | vvdw6 = _mm_mul_ps(c6_00,rinvsix); |
825 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
826 | vvdw = _mm_sub_ps( _mm_mul_ps(vvdw12,one_twelfth) , _mm_mul_ps(vvdw6,one_sixth) ); |
827 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq00); |
828 | |
829 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
830 | velec = _mm_andnot_ps(dummy_mask,velec); |
831 | velecsum = _mm_add_ps(velecsum,velec); |
832 | vvdw = _mm_andnot_ps(dummy_mask,vvdw); |
833 | vvdwsum = _mm_add_ps(vvdwsum,vvdw); |
834 | |
835 | fscal = _mm_add_ps(felec,fvdw); |
836 | |
837 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
838 | |
839 | /* Calculate temporary vectorial force */ |
840 | tx = _mm_mul_ps(fscal,dx00); |
841 | ty = _mm_mul_ps(fscal,dy00); |
842 | tz = _mm_mul_ps(fscal,dz00); |
843 | |
844 | /* Update vectorial force */ |
845 | fix0 = _mm_add_ps(fix0,tx); |
846 | fiy0 = _mm_add_ps(fiy0,ty); |
847 | fiz0 = _mm_add_ps(fiz0,tz); |
848 | |
849 | fjx0 = _mm_add_ps(fjx0,tx); |
850 | fjy0 = _mm_add_ps(fjy0,ty); |
851 | fjz0 = _mm_add_ps(fjz0,tz); |
852 | |
853 | /************************** |
854 | * CALCULATE INTERACTIONS * |
855 | **************************/ |
856 | |
857 | r01 = _mm_mul_ps(rsq01,rinv01); |
858 | r01 = _mm_andnot_ps(dummy_mask,r01); |
859 | |
860 | /* EWALD ELECTROSTATICS */ |
861 | |
862 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
863 | ewrt = _mm_mul_ps(r01,ewtabscale); |
864 | ewitab = _mm_cvttps_epi32(ewrt); |
865 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
866 | ewitab = _mm_slli_epi32(ewitab,2); |
867 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
868 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
869 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
870 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
871 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
872 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
873 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
874 | velec = _mm_mul_ps(qq01,_mm_sub_ps(rinv01,velec)); |
875 | felec = _mm_mul_ps(_mm_mul_ps(qq01,rinv01),_mm_sub_ps(rinvsq01,felec)); |
876 | |
877 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
878 | velec = _mm_andnot_ps(dummy_mask,velec); |
879 | velecsum = _mm_add_ps(velecsum,velec); |
880 | |
881 | fscal = felec; |
882 | |
883 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
884 | |
885 | /* Calculate temporary vectorial force */ |
886 | tx = _mm_mul_ps(fscal,dx01); |
887 | ty = _mm_mul_ps(fscal,dy01); |
888 | tz = _mm_mul_ps(fscal,dz01); |
889 | |
890 | /* Update vectorial force */ |
891 | fix0 = _mm_add_ps(fix0,tx); |
892 | fiy0 = _mm_add_ps(fiy0,ty); |
893 | fiz0 = _mm_add_ps(fiz0,tz); |
894 | |
895 | fjx1 = _mm_add_ps(fjx1,tx); |
896 | fjy1 = _mm_add_ps(fjy1,ty); |
897 | fjz1 = _mm_add_ps(fjz1,tz); |
898 | |
899 | /************************** |
900 | * CALCULATE INTERACTIONS * |
901 | **************************/ |
902 | |
903 | r02 = _mm_mul_ps(rsq02,rinv02); |
904 | r02 = _mm_andnot_ps(dummy_mask,r02); |
905 | |
906 | /* EWALD ELECTROSTATICS */ |
907 | |
908 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
909 | ewrt = _mm_mul_ps(r02,ewtabscale); |
910 | ewitab = _mm_cvttps_epi32(ewrt); |
911 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
912 | ewitab = _mm_slli_epi32(ewitab,2); |
913 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
914 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
915 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
916 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
917 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
918 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
919 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
920 | velec = _mm_mul_ps(qq02,_mm_sub_ps(rinv02,velec)); |
921 | felec = _mm_mul_ps(_mm_mul_ps(qq02,rinv02),_mm_sub_ps(rinvsq02,felec)); |
922 | |
923 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
924 | velec = _mm_andnot_ps(dummy_mask,velec); |
925 | velecsum = _mm_add_ps(velecsum,velec); |
926 | |
927 | fscal = felec; |
928 | |
929 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
930 | |
931 | /* Calculate temporary vectorial force */ |
932 | tx = _mm_mul_ps(fscal,dx02); |
933 | ty = _mm_mul_ps(fscal,dy02); |
934 | tz = _mm_mul_ps(fscal,dz02); |
935 | |
936 | /* Update vectorial force */ |
937 | fix0 = _mm_add_ps(fix0,tx); |
938 | fiy0 = _mm_add_ps(fiy0,ty); |
939 | fiz0 = _mm_add_ps(fiz0,tz); |
940 | |
941 | fjx2 = _mm_add_ps(fjx2,tx); |
942 | fjy2 = _mm_add_ps(fjy2,ty); |
943 | fjz2 = _mm_add_ps(fjz2,tz); |
944 | |
945 | /************************** |
946 | * CALCULATE INTERACTIONS * |
947 | **************************/ |
948 | |
949 | r10 = _mm_mul_ps(rsq10,rinv10); |
950 | r10 = _mm_andnot_ps(dummy_mask,r10); |
951 | |
952 | /* EWALD ELECTROSTATICS */ |
953 | |
954 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
955 | ewrt = _mm_mul_ps(r10,ewtabscale); |
956 | ewitab = _mm_cvttps_epi32(ewrt); |
957 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
958 | ewitab = _mm_slli_epi32(ewitab,2); |
959 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
960 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
961 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
962 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
963 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
964 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
965 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
966 | velec = _mm_mul_ps(qq10,_mm_sub_ps(rinv10,velec)); |
967 | felec = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec)); |
968 | |
969 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
970 | velec = _mm_andnot_ps(dummy_mask,velec); |
971 | velecsum = _mm_add_ps(velecsum,velec); |
972 | |
973 | fscal = felec; |
974 | |
975 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
976 | |
977 | /* Calculate temporary vectorial force */ |
978 | tx = _mm_mul_ps(fscal,dx10); |
979 | ty = _mm_mul_ps(fscal,dy10); |
980 | tz = _mm_mul_ps(fscal,dz10); |
981 | |
982 | /* Update vectorial force */ |
983 | fix1 = _mm_add_ps(fix1,tx); |
984 | fiy1 = _mm_add_ps(fiy1,ty); |
985 | fiz1 = _mm_add_ps(fiz1,tz); |
986 | |
987 | fjx0 = _mm_add_ps(fjx0,tx); |
988 | fjy0 = _mm_add_ps(fjy0,ty); |
989 | fjz0 = _mm_add_ps(fjz0,tz); |
990 | |
991 | /************************** |
992 | * CALCULATE INTERACTIONS * |
993 | **************************/ |
994 | |
995 | r11 = _mm_mul_ps(rsq11,rinv11); |
996 | r11 = _mm_andnot_ps(dummy_mask,r11); |
997 | |
998 | /* EWALD ELECTROSTATICS */ |
999 | |
1000 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1001 | ewrt = _mm_mul_ps(r11,ewtabscale); |
1002 | ewitab = _mm_cvttps_epi32(ewrt); |
1003 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1004 | ewitab = _mm_slli_epi32(ewitab,2); |
1005 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1006 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1007 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1008 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1009 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1010 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1011 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1012 | velec = _mm_mul_ps(qq11,_mm_sub_ps(rinv11,velec)); |
1013 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
1014 | |
1015 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1016 | velec = _mm_andnot_ps(dummy_mask,velec); |
1017 | velecsum = _mm_add_ps(velecsum,velec); |
1018 | |
1019 | fscal = felec; |
1020 | |
1021 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1022 | |
1023 | /* Calculate temporary vectorial force */ |
1024 | tx = _mm_mul_ps(fscal,dx11); |
1025 | ty = _mm_mul_ps(fscal,dy11); |
1026 | tz = _mm_mul_ps(fscal,dz11); |
1027 | |
1028 | /* Update vectorial force */ |
1029 | fix1 = _mm_add_ps(fix1,tx); |
1030 | fiy1 = _mm_add_ps(fiy1,ty); |
1031 | fiz1 = _mm_add_ps(fiz1,tz); |
1032 | |
1033 | fjx1 = _mm_add_ps(fjx1,tx); |
1034 | fjy1 = _mm_add_ps(fjy1,ty); |
1035 | fjz1 = _mm_add_ps(fjz1,tz); |
1036 | |
1037 | /************************** |
1038 | * CALCULATE INTERACTIONS * |
1039 | **************************/ |
1040 | |
1041 | r12 = _mm_mul_ps(rsq12,rinv12); |
1042 | r12 = _mm_andnot_ps(dummy_mask,r12); |
1043 | |
1044 | /* EWALD ELECTROSTATICS */ |
1045 | |
1046 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1047 | ewrt = _mm_mul_ps(r12,ewtabscale); |
1048 | ewitab = _mm_cvttps_epi32(ewrt); |
1049 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1050 | ewitab = _mm_slli_epi32(ewitab,2); |
1051 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1052 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1053 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1054 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1055 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1056 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1057 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1058 | velec = _mm_mul_ps(qq12,_mm_sub_ps(rinv12,velec)); |
1059 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
1060 | |
1061 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1062 | velec = _mm_andnot_ps(dummy_mask,velec); |
1063 | velecsum = _mm_add_ps(velecsum,velec); |
1064 | |
1065 | fscal = felec; |
1066 | |
1067 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1068 | |
1069 | /* Calculate temporary vectorial force */ |
1070 | tx = _mm_mul_ps(fscal,dx12); |
1071 | ty = _mm_mul_ps(fscal,dy12); |
1072 | tz = _mm_mul_ps(fscal,dz12); |
1073 | |
1074 | /* Update vectorial force */ |
1075 | fix1 = _mm_add_ps(fix1,tx); |
1076 | fiy1 = _mm_add_ps(fiy1,ty); |
1077 | fiz1 = _mm_add_ps(fiz1,tz); |
1078 | |
1079 | fjx2 = _mm_add_ps(fjx2,tx); |
1080 | fjy2 = _mm_add_ps(fjy2,ty); |
1081 | fjz2 = _mm_add_ps(fjz2,tz); |
1082 | |
1083 | /************************** |
1084 | * CALCULATE INTERACTIONS * |
1085 | **************************/ |
1086 | |
1087 | r20 = _mm_mul_ps(rsq20,rinv20); |
1088 | r20 = _mm_andnot_ps(dummy_mask,r20); |
1089 | |
1090 | /* EWALD ELECTROSTATICS */ |
1091 | |
1092 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1093 | ewrt = _mm_mul_ps(r20,ewtabscale); |
1094 | ewitab = _mm_cvttps_epi32(ewrt); |
1095 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1096 | ewitab = _mm_slli_epi32(ewitab,2); |
1097 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1098 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1099 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1100 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1101 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1102 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1103 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1104 | velec = _mm_mul_ps(qq20,_mm_sub_ps(rinv20,velec)); |
1105 | felec = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec)); |
1106 | |
1107 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1108 | velec = _mm_andnot_ps(dummy_mask,velec); |
1109 | velecsum = _mm_add_ps(velecsum,velec); |
1110 | |
1111 | fscal = felec; |
1112 | |
1113 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1114 | |
1115 | /* Calculate temporary vectorial force */ |
1116 | tx = _mm_mul_ps(fscal,dx20); |
1117 | ty = _mm_mul_ps(fscal,dy20); |
1118 | tz = _mm_mul_ps(fscal,dz20); |
1119 | |
1120 | /* Update vectorial force */ |
1121 | fix2 = _mm_add_ps(fix2,tx); |
1122 | fiy2 = _mm_add_ps(fiy2,ty); |
1123 | fiz2 = _mm_add_ps(fiz2,tz); |
1124 | |
1125 | fjx0 = _mm_add_ps(fjx0,tx); |
1126 | fjy0 = _mm_add_ps(fjy0,ty); |
1127 | fjz0 = _mm_add_ps(fjz0,tz); |
1128 | |
1129 | /************************** |
1130 | * CALCULATE INTERACTIONS * |
1131 | **************************/ |
1132 | |
1133 | r21 = _mm_mul_ps(rsq21,rinv21); |
1134 | r21 = _mm_andnot_ps(dummy_mask,r21); |
1135 | |
1136 | /* EWALD ELECTROSTATICS */ |
1137 | |
1138 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1139 | ewrt = _mm_mul_ps(r21,ewtabscale); |
1140 | ewitab = _mm_cvttps_epi32(ewrt); |
1141 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1142 | ewitab = _mm_slli_epi32(ewitab,2); |
1143 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1144 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1145 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1146 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1147 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1148 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1149 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1150 | velec = _mm_mul_ps(qq21,_mm_sub_ps(rinv21,velec)); |
1151 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
1152 | |
1153 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1154 | velec = _mm_andnot_ps(dummy_mask,velec); |
1155 | velecsum = _mm_add_ps(velecsum,velec); |
1156 | |
1157 | fscal = felec; |
1158 | |
1159 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1160 | |
1161 | /* Calculate temporary vectorial force */ |
1162 | tx = _mm_mul_ps(fscal,dx21); |
1163 | ty = _mm_mul_ps(fscal,dy21); |
1164 | tz = _mm_mul_ps(fscal,dz21); |
1165 | |
1166 | /* Update vectorial force */ |
1167 | fix2 = _mm_add_ps(fix2,tx); |
1168 | fiy2 = _mm_add_ps(fiy2,ty); |
1169 | fiz2 = _mm_add_ps(fiz2,tz); |
1170 | |
1171 | fjx1 = _mm_add_ps(fjx1,tx); |
1172 | fjy1 = _mm_add_ps(fjy1,ty); |
1173 | fjz1 = _mm_add_ps(fjz1,tz); |
1174 | |
1175 | /************************** |
1176 | * CALCULATE INTERACTIONS * |
1177 | **************************/ |
1178 | |
1179 | r22 = _mm_mul_ps(rsq22,rinv22); |
1180 | r22 = _mm_andnot_ps(dummy_mask,r22); |
1181 | |
1182 | /* EWALD ELECTROSTATICS */ |
1183 | |
1184 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1185 | ewrt = _mm_mul_ps(r22,ewtabscale); |
1186 | ewitab = _mm_cvttps_epi32(ewrt); |
1187 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1188 | ewitab = _mm_slli_epi32(ewitab,2); |
1189 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1190 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1191 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1192 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1193 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1194 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1195 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1196 | velec = _mm_mul_ps(qq22,_mm_sub_ps(rinv22,velec)); |
1197 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
1198 | |
1199 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1200 | velec = _mm_andnot_ps(dummy_mask,velec); |
1201 | velecsum = _mm_add_ps(velecsum,velec); |
1202 | |
1203 | fscal = felec; |
1204 | |
1205 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1206 | |
1207 | /* Calculate temporary vectorial force */ |
1208 | tx = _mm_mul_ps(fscal,dx22); |
1209 | ty = _mm_mul_ps(fscal,dy22); |
1210 | tz = _mm_mul_ps(fscal,dz22); |
1211 | |
1212 | /* Update vectorial force */ |
1213 | fix2 = _mm_add_ps(fix2,tx); |
1214 | fiy2 = _mm_add_ps(fiy2,ty); |
1215 | fiz2 = _mm_add_ps(fiz2,tz); |
1216 | |
1217 | fjx2 = _mm_add_ps(fjx2,tx); |
1218 | fjy2 = _mm_add_ps(fjy2,ty); |
1219 | fjz2 = _mm_add_ps(fjz2,tz); |
1220 | |
1221 | fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch; |
1222 | fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch; |
1223 | fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch; |
1224 | fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch; |
1225 | |
1226 | gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
1227 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2); |
1228 | |
1229 | /* Inner loop uses 390 flops */ |
1230 | } |
1231 | |
1232 | /* End of innermost loop */ |
1233 | |
1234 | gmx_mm_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2, |
1235 | f+i_coord_offset,fshift+i_shift_offset); |
1236 | |
1237 | ggid = gid[iidx]; |
1238 | /* Update potential energies */ |
1239 | gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid); |
1240 | gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid); |
1241 | |
1242 | /* Increment number of inner iterations */ |
1243 | inneriter += j_index_end - j_index_start; |
1244 | |
1245 | /* Outer loop uses 20 flops */ |
1246 | } |
1247 | |
1248 | /* Increment number of outer iterations */ |
1249 | outeriter += nri; |
1250 | |
1251 | /* Update outer/inner flops */ |
1252 | |
1253 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*20 + inneriter*390)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_VF] += outeriter*20 + inneriter*390; |
1254 | } |
1255 | /* |
1256 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJ_GeomW3W3_F_sse4_1_single |
1257 | * Electrostatics interaction: Ewald |
1258 | * VdW interaction: LennardJones |
1259 | * Geometry: Water3-Water3 |
1260 | * Calculate force/pot: Force |
1261 | */ |
1262 | void |
1263 | nb_kernel_ElecEw_VdwLJ_GeomW3W3_F_sse4_1_single |
1264 | (t_nblist * gmx_restrict nlist, |
1265 | rvec * gmx_restrict xx, |
1266 | rvec * gmx_restrict ff, |
1267 | t_forcerec * gmx_restrict fr, |
1268 | t_mdatoms * gmx_restrict mdatoms, |
1269 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict kernel_data, |
1270 | t_nrnb * gmx_restrict nrnb) |
1271 | { |
1272 | /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or |
1273 | * just 0 for non-waters. |
1274 | * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different |
1275 | * jnr indices corresponding to data put in the four positions in the SIMD register. |
1276 | */ |
1277 | int i_shift_offset,i_coord_offset,outeriter,inneriter; |
1278 | int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx; |
1279 | int jnrA,jnrB,jnrC,jnrD; |
1280 | int jnrlistA,jnrlistB,jnrlistC,jnrlistD; |
1281 | int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD; |
1282 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
1283 | real rcutoff_scalar; |
1284 | real *shiftvec,*fshift,*x,*f; |
1285 | real *fjptrA,*fjptrB,*fjptrC,*fjptrD; |
1286 | real scratch[4*DIM3]; |
1287 | __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall; |
1288 | int vdwioffset0; |
1289 | __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
1290 | int vdwioffset1; |
1291 | __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
1292 | int vdwioffset2; |
1293 | __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
1294 | int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D; |
1295 | __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
1296 | int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D; |
1297 | __m128 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
1298 | int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D; |
1299 | __m128 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
1300 | __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00; |
1301 | __m128 dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01; |
1302 | __m128 dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02; |
1303 | __m128 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10; |
1304 | __m128 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11; |
1305 | __m128 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12; |
1306 | __m128 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20; |
1307 | __m128 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21; |
1308 | __m128 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22; |
1309 | __m128 velec,felec,velecsum,facel,crf,krf,krf2; |
1310 | real *charge; |
1311 | int nvdwtype; |
1312 | __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6; |
1313 | int *vdwtype; |
1314 | real *vdwparam; |
1315 | __m128 one_sixth = _mm_set1_ps(1.0/6.0); |
1316 | __m128 one_twelfth = _mm_set1_ps(1.0/12.0); |
1317 | __m128i ewitab; |
1318 | __m128 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV; |
1319 | real *ewtab; |
1320 | __m128 dummy_mask,cutoff_mask; |
1321 | __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) ); |
1322 | __m128 one = _mm_set1_ps(1.0); |
1323 | __m128 two = _mm_set1_ps(2.0); |
1324 | x = xx[0]; |
1325 | f = ff[0]; |
1326 | |
1327 | nri = nlist->nri; |
1328 | iinr = nlist->iinr; |
1329 | jindex = nlist->jindex; |
1330 | jjnr = nlist->jjnr; |
1331 | shiftidx = nlist->shift; |
1332 | gid = nlist->gid; |
1333 | shiftvec = fr->shift_vec[0]; |
1334 | fshift = fr->fshift[0]; |
1335 | facel = _mm_set1_ps(fr->epsfac); |
1336 | charge = mdatoms->chargeA; |
1337 | nvdwtype = fr->ntype; |
1338 | vdwparam = fr->nbfp; |
1339 | vdwtype = mdatoms->typeA; |
1340 | |
1341 | sh_ewald = _mm_set1_ps(fr->ic->sh_ewald); |
1342 | ewtab = fr->ic->tabq_coul_F; |
1343 | ewtabscale = _mm_set1_ps(fr->ic->tabq_scale); |
1344 | ewtabhalfspace = _mm_set1_ps(0.5/fr->ic->tabq_scale); |
1345 | |
1346 | /* Setup water-specific parameters */ |
1347 | inr = nlist->iinr[0]; |
1348 | iq0 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+0])); |
1349 | iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1])); |
1350 | iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2])); |
1351 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
1352 | |
1353 | jq0 = _mm_set1_ps(charge[inr+0]); |
1354 | jq1 = _mm_set1_ps(charge[inr+1]); |
1355 | jq2 = _mm_set1_ps(charge[inr+2]); |
1356 | vdwjidx0A = 2*vdwtype[inr+0]; |
1357 | qq00 = _mm_mul_ps(iq0,jq0); |
1358 | c6_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A]); |
1359 | c12_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A+1]); |
1360 | qq01 = _mm_mul_ps(iq0,jq1); |
1361 | qq02 = _mm_mul_ps(iq0,jq2); |
1362 | qq10 = _mm_mul_ps(iq1,jq0); |
1363 | qq11 = _mm_mul_ps(iq1,jq1); |
1364 | qq12 = _mm_mul_ps(iq1,jq2); |
1365 | qq20 = _mm_mul_ps(iq2,jq0); |
1366 | qq21 = _mm_mul_ps(iq2,jq1); |
1367 | qq22 = _mm_mul_ps(iq2,jq2); |
1368 | |
1369 | /* Avoid stupid compiler warnings */ |
1370 | jnrA = jnrB = jnrC = jnrD = 0; |
1371 | j_coord_offsetA = 0; |
1372 | j_coord_offsetB = 0; |
1373 | j_coord_offsetC = 0; |
1374 | j_coord_offsetD = 0; |
1375 | |
1376 | outeriter = 0; |
1377 | inneriter = 0; |
1378 | |
1379 | for(iidx=0;iidx<4*DIM3;iidx++) |
1380 | { |
1381 | scratch[iidx] = 0.0; |
1382 | } |
1383 | |
1384 | /* Start outer loop over neighborlists */ |
1385 | for(iidx=0; iidx<nri; iidx++) |
1386 | { |
1387 | /* Load shift vector for this list */ |
1388 | i_shift_offset = DIM3*shiftidx[iidx]; |
1389 | |
1390 | /* Load limits for loop over neighbors */ |
1391 | j_index_start = jindex[iidx]; |
1392 | j_index_end = jindex[iidx+1]; |
1393 | |
1394 | /* Get outer coordinate index */ |
1395 | inr = iinr[iidx]; |
1396 | i_coord_offset = DIM3*inr; |
1397 | |
1398 | /* Load i particle coords and add shift vector */ |
1399 | gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset, |
1400 | &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2); |
1401 | |
1402 | fix0 = _mm_setzero_ps(); |
1403 | fiy0 = _mm_setzero_ps(); |
1404 | fiz0 = _mm_setzero_ps(); |
1405 | fix1 = _mm_setzero_ps(); |
1406 | fiy1 = _mm_setzero_ps(); |
1407 | fiz1 = _mm_setzero_ps(); |
1408 | fix2 = _mm_setzero_ps(); |
1409 | fiy2 = _mm_setzero_ps(); |
1410 | fiz2 = _mm_setzero_ps(); |
1411 | |
1412 | /* Start inner kernel loop */ |
1413 | for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4) |
1414 | { |
1415 | |
1416 | /* Get j neighbor index, and coordinate index */ |
1417 | jnrA = jjnr[jidx]; |
1418 | jnrB = jjnr[jidx+1]; |
1419 | jnrC = jjnr[jidx+2]; |
1420 | jnrD = jjnr[jidx+3]; |
1421 | j_coord_offsetA = DIM3*jnrA; |
1422 | j_coord_offsetB = DIM3*jnrB; |
1423 | j_coord_offsetC = DIM3*jnrC; |
1424 | j_coord_offsetD = DIM3*jnrD; |
1425 | |
1426 | /* load j atom coordinates */ |
1427 | gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
1428 | x+j_coord_offsetC,x+j_coord_offsetD, |
1429 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2); |
1430 | |
1431 | /* Calculate displacement vector */ |
1432 | dx00 = _mm_sub_ps(ix0,jx0); |
1433 | dy00 = _mm_sub_ps(iy0,jy0); |
1434 | dz00 = _mm_sub_ps(iz0,jz0); |
1435 | dx01 = _mm_sub_ps(ix0,jx1); |
1436 | dy01 = _mm_sub_ps(iy0,jy1); |
1437 | dz01 = _mm_sub_ps(iz0,jz1); |
1438 | dx02 = _mm_sub_ps(ix0,jx2); |
1439 | dy02 = _mm_sub_ps(iy0,jy2); |
1440 | dz02 = _mm_sub_ps(iz0,jz2); |
1441 | dx10 = _mm_sub_ps(ix1,jx0); |
1442 | dy10 = _mm_sub_ps(iy1,jy0); |
1443 | dz10 = _mm_sub_ps(iz1,jz0); |
1444 | dx11 = _mm_sub_ps(ix1,jx1); |
1445 | dy11 = _mm_sub_ps(iy1,jy1); |
1446 | dz11 = _mm_sub_ps(iz1,jz1); |
1447 | dx12 = _mm_sub_ps(ix1,jx2); |
1448 | dy12 = _mm_sub_ps(iy1,jy2); |
1449 | dz12 = _mm_sub_ps(iz1,jz2); |
1450 | dx20 = _mm_sub_ps(ix2,jx0); |
1451 | dy20 = _mm_sub_ps(iy2,jy0); |
1452 | dz20 = _mm_sub_ps(iz2,jz0); |
1453 | dx21 = _mm_sub_ps(ix2,jx1); |
1454 | dy21 = _mm_sub_ps(iy2,jy1); |
1455 | dz21 = _mm_sub_ps(iz2,jz1); |
1456 | dx22 = _mm_sub_ps(ix2,jx2); |
1457 | dy22 = _mm_sub_ps(iy2,jy2); |
1458 | dz22 = _mm_sub_ps(iz2,jz2); |
1459 | |
1460 | /* Calculate squared distance and things based on it */ |
1461 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
1462 | rsq01 = gmx_mm_calc_rsq_ps(dx01,dy01,dz01); |
1463 | rsq02 = gmx_mm_calc_rsq_ps(dx02,dy02,dz02); |
1464 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
1465 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
1466 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
1467 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
1468 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
1469 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
1470 | |
1471 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
1472 | rinv01 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq01); |
1473 | rinv02 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq02); |
1474 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
1475 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
1476 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
1477 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
1478 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
1479 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
1480 | |
1481 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
1482 | rinvsq01 = _mm_mul_ps(rinv01,rinv01); |
1483 | rinvsq02 = _mm_mul_ps(rinv02,rinv02); |
1484 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
1485 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
1486 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
1487 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
1488 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
1489 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
1490 | |
1491 | fjx0 = _mm_setzero_ps(); |
1492 | fjy0 = _mm_setzero_ps(); |
1493 | fjz0 = _mm_setzero_ps(); |
1494 | fjx1 = _mm_setzero_ps(); |
1495 | fjy1 = _mm_setzero_ps(); |
1496 | fjz1 = _mm_setzero_ps(); |
1497 | fjx2 = _mm_setzero_ps(); |
1498 | fjy2 = _mm_setzero_ps(); |
1499 | fjz2 = _mm_setzero_ps(); |
1500 | |
1501 | /************************** |
1502 | * CALCULATE INTERACTIONS * |
1503 | **************************/ |
1504 | |
1505 | r00 = _mm_mul_ps(rsq00,rinv00); |
1506 | |
1507 | /* EWALD ELECTROSTATICS */ |
1508 | |
1509 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1510 | ewrt = _mm_mul_ps(r00,ewtabscale); |
1511 | ewitab = _mm_cvttps_epi32(ewrt); |
1512 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1513 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1514 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1515 | &ewtabF,&ewtabFn); |
1516 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1517 | felec = _mm_mul_ps(_mm_mul_ps(qq00,rinv00),_mm_sub_ps(rinvsq00,felec)); |
1518 | |
1519 | /* LENNARD-JONES DISPERSION/REPULSION */ |
1520 | |
1521 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
1522 | fvdw = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_00,rinvsix),c6_00),_mm_mul_ps(rinvsix,rinvsq00)); |
1523 | |
1524 | fscal = _mm_add_ps(felec,fvdw); |
1525 | |
1526 | /* Calculate temporary vectorial force */ |
1527 | tx = _mm_mul_ps(fscal,dx00); |
1528 | ty = _mm_mul_ps(fscal,dy00); |
1529 | tz = _mm_mul_ps(fscal,dz00); |
1530 | |
1531 | /* Update vectorial force */ |
1532 | fix0 = _mm_add_ps(fix0,tx); |
1533 | fiy0 = _mm_add_ps(fiy0,ty); |
1534 | fiz0 = _mm_add_ps(fiz0,tz); |
1535 | |
1536 | fjx0 = _mm_add_ps(fjx0,tx); |
1537 | fjy0 = _mm_add_ps(fjy0,ty); |
1538 | fjz0 = _mm_add_ps(fjz0,tz); |
1539 | |
1540 | /************************** |
1541 | * CALCULATE INTERACTIONS * |
1542 | **************************/ |
1543 | |
1544 | r01 = _mm_mul_ps(rsq01,rinv01); |
1545 | |
1546 | /* EWALD ELECTROSTATICS */ |
1547 | |
1548 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1549 | ewrt = _mm_mul_ps(r01,ewtabscale); |
1550 | ewitab = _mm_cvttps_epi32(ewrt); |
1551 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1552 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1553 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1554 | &ewtabF,&ewtabFn); |
1555 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1556 | felec = _mm_mul_ps(_mm_mul_ps(qq01,rinv01),_mm_sub_ps(rinvsq01,felec)); |
1557 | |
1558 | fscal = felec; |
1559 | |
1560 | /* Calculate temporary vectorial force */ |
1561 | tx = _mm_mul_ps(fscal,dx01); |
1562 | ty = _mm_mul_ps(fscal,dy01); |
1563 | tz = _mm_mul_ps(fscal,dz01); |
1564 | |
1565 | /* Update vectorial force */ |
1566 | fix0 = _mm_add_ps(fix0,tx); |
1567 | fiy0 = _mm_add_ps(fiy0,ty); |
1568 | fiz0 = _mm_add_ps(fiz0,tz); |
1569 | |
1570 | fjx1 = _mm_add_ps(fjx1,tx); |
1571 | fjy1 = _mm_add_ps(fjy1,ty); |
1572 | fjz1 = _mm_add_ps(fjz1,tz); |
1573 | |
1574 | /************************** |
1575 | * CALCULATE INTERACTIONS * |
1576 | **************************/ |
1577 | |
1578 | r02 = _mm_mul_ps(rsq02,rinv02); |
1579 | |
1580 | /* EWALD ELECTROSTATICS */ |
1581 | |
1582 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1583 | ewrt = _mm_mul_ps(r02,ewtabscale); |
1584 | ewitab = _mm_cvttps_epi32(ewrt); |
1585 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1586 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1587 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1588 | &ewtabF,&ewtabFn); |
1589 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1590 | felec = _mm_mul_ps(_mm_mul_ps(qq02,rinv02),_mm_sub_ps(rinvsq02,felec)); |
1591 | |
1592 | fscal = felec; |
1593 | |
1594 | /* Calculate temporary vectorial force */ |
1595 | tx = _mm_mul_ps(fscal,dx02); |
1596 | ty = _mm_mul_ps(fscal,dy02); |
1597 | tz = _mm_mul_ps(fscal,dz02); |
1598 | |
1599 | /* Update vectorial force */ |
1600 | fix0 = _mm_add_ps(fix0,tx); |
1601 | fiy0 = _mm_add_ps(fiy0,ty); |
1602 | fiz0 = _mm_add_ps(fiz0,tz); |
1603 | |
1604 | fjx2 = _mm_add_ps(fjx2,tx); |
1605 | fjy2 = _mm_add_ps(fjy2,ty); |
1606 | fjz2 = _mm_add_ps(fjz2,tz); |
1607 | |
1608 | /************************** |
1609 | * CALCULATE INTERACTIONS * |
1610 | **************************/ |
1611 | |
1612 | r10 = _mm_mul_ps(rsq10,rinv10); |
1613 | |
1614 | /* EWALD ELECTROSTATICS */ |
1615 | |
1616 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1617 | ewrt = _mm_mul_ps(r10,ewtabscale); |
1618 | ewitab = _mm_cvttps_epi32(ewrt); |
1619 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1620 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1621 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1622 | &ewtabF,&ewtabFn); |
1623 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1624 | felec = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec)); |
1625 | |
1626 | fscal = felec; |
1627 | |
1628 | /* Calculate temporary vectorial force */ |
1629 | tx = _mm_mul_ps(fscal,dx10); |
1630 | ty = _mm_mul_ps(fscal,dy10); |
1631 | tz = _mm_mul_ps(fscal,dz10); |
1632 | |
1633 | /* Update vectorial force */ |
1634 | fix1 = _mm_add_ps(fix1,tx); |
1635 | fiy1 = _mm_add_ps(fiy1,ty); |
1636 | fiz1 = _mm_add_ps(fiz1,tz); |
1637 | |
1638 | fjx0 = _mm_add_ps(fjx0,tx); |
1639 | fjy0 = _mm_add_ps(fjy0,ty); |
1640 | fjz0 = _mm_add_ps(fjz0,tz); |
1641 | |
1642 | /************************** |
1643 | * CALCULATE INTERACTIONS * |
1644 | **************************/ |
1645 | |
1646 | r11 = _mm_mul_ps(rsq11,rinv11); |
1647 | |
1648 | /* EWALD ELECTROSTATICS */ |
1649 | |
1650 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1651 | ewrt = _mm_mul_ps(r11,ewtabscale); |
1652 | ewitab = _mm_cvttps_epi32(ewrt); |
1653 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1654 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1655 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1656 | &ewtabF,&ewtabFn); |
1657 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1658 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
1659 | |
1660 | fscal = felec; |
1661 | |
1662 | /* Calculate temporary vectorial force */ |
1663 | tx = _mm_mul_ps(fscal,dx11); |
1664 | ty = _mm_mul_ps(fscal,dy11); |
1665 | tz = _mm_mul_ps(fscal,dz11); |
1666 | |
1667 | /* Update vectorial force */ |
1668 | fix1 = _mm_add_ps(fix1,tx); |
1669 | fiy1 = _mm_add_ps(fiy1,ty); |
1670 | fiz1 = _mm_add_ps(fiz1,tz); |
1671 | |
1672 | fjx1 = _mm_add_ps(fjx1,tx); |
1673 | fjy1 = _mm_add_ps(fjy1,ty); |
1674 | fjz1 = _mm_add_ps(fjz1,tz); |
1675 | |
1676 | /************************** |
1677 | * CALCULATE INTERACTIONS * |
1678 | **************************/ |
1679 | |
1680 | r12 = _mm_mul_ps(rsq12,rinv12); |
1681 | |
1682 | /* EWALD ELECTROSTATICS */ |
1683 | |
1684 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1685 | ewrt = _mm_mul_ps(r12,ewtabscale); |
1686 | ewitab = _mm_cvttps_epi32(ewrt); |
1687 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1688 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1689 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1690 | &ewtabF,&ewtabFn); |
1691 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1692 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
1693 | |
1694 | fscal = felec; |
1695 | |
1696 | /* Calculate temporary vectorial force */ |
1697 | tx = _mm_mul_ps(fscal,dx12); |
1698 | ty = _mm_mul_ps(fscal,dy12); |
1699 | tz = _mm_mul_ps(fscal,dz12); |
1700 | |
1701 | /* Update vectorial force */ |
1702 | fix1 = _mm_add_ps(fix1,tx); |
1703 | fiy1 = _mm_add_ps(fiy1,ty); |
1704 | fiz1 = _mm_add_ps(fiz1,tz); |
1705 | |
1706 | fjx2 = _mm_add_ps(fjx2,tx); |
1707 | fjy2 = _mm_add_ps(fjy2,ty); |
1708 | fjz2 = _mm_add_ps(fjz2,tz); |
1709 | |
1710 | /************************** |
1711 | * CALCULATE INTERACTIONS * |
1712 | **************************/ |
1713 | |
1714 | r20 = _mm_mul_ps(rsq20,rinv20); |
1715 | |
1716 | /* EWALD ELECTROSTATICS */ |
1717 | |
1718 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1719 | ewrt = _mm_mul_ps(r20,ewtabscale); |
1720 | ewitab = _mm_cvttps_epi32(ewrt); |
1721 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1722 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1723 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1724 | &ewtabF,&ewtabFn); |
1725 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1726 | felec = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec)); |
1727 | |
1728 | fscal = felec; |
1729 | |
1730 | /* Calculate temporary vectorial force */ |
1731 | tx = _mm_mul_ps(fscal,dx20); |
1732 | ty = _mm_mul_ps(fscal,dy20); |
1733 | tz = _mm_mul_ps(fscal,dz20); |
1734 | |
1735 | /* Update vectorial force */ |
1736 | fix2 = _mm_add_ps(fix2,tx); |
1737 | fiy2 = _mm_add_ps(fiy2,ty); |
1738 | fiz2 = _mm_add_ps(fiz2,tz); |
1739 | |
1740 | fjx0 = _mm_add_ps(fjx0,tx); |
1741 | fjy0 = _mm_add_ps(fjy0,ty); |
1742 | fjz0 = _mm_add_ps(fjz0,tz); |
1743 | |
1744 | /************************** |
1745 | * CALCULATE INTERACTIONS * |
1746 | **************************/ |
1747 | |
1748 | r21 = _mm_mul_ps(rsq21,rinv21); |
1749 | |
1750 | /* EWALD ELECTROSTATICS */ |
1751 | |
1752 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1753 | ewrt = _mm_mul_ps(r21,ewtabscale); |
1754 | ewitab = _mm_cvttps_epi32(ewrt); |
1755 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1756 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1757 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1758 | &ewtabF,&ewtabFn); |
1759 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1760 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
1761 | |
1762 | fscal = felec; |
1763 | |
1764 | /* Calculate temporary vectorial force */ |
1765 | tx = _mm_mul_ps(fscal,dx21); |
1766 | ty = _mm_mul_ps(fscal,dy21); |
1767 | tz = _mm_mul_ps(fscal,dz21); |
1768 | |
1769 | /* Update vectorial force */ |
1770 | fix2 = _mm_add_ps(fix2,tx); |
1771 | fiy2 = _mm_add_ps(fiy2,ty); |
1772 | fiz2 = _mm_add_ps(fiz2,tz); |
1773 | |
1774 | fjx1 = _mm_add_ps(fjx1,tx); |
1775 | fjy1 = _mm_add_ps(fjy1,ty); |
1776 | fjz1 = _mm_add_ps(fjz1,tz); |
1777 | |
1778 | /************************** |
1779 | * CALCULATE INTERACTIONS * |
1780 | **************************/ |
1781 | |
1782 | r22 = _mm_mul_ps(rsq22,rinv22); |
1783 | |
1784 | /* EWALD ELECTROSTATICS */ |
1785 | |
1786 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1787 | ewrt = _mm_mul_ps(r22,ewtabscale); |
1788 | ewitab = _mm_cvttps_epi32(ewrt); |
1789 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1790 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1791 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1792 | &ewtabF,&ewtabFn); |
1793 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1794 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
1795 | |
1796 | fscal = felec; |
1797 | |
1798 | /* Calculate temporary vectorial force */ |
1799 | tx = _mm_mul_ps(fscal,dx22); |
1800 | ty = _mm_mul_ps(fscal,dy22); |
1801 | tz = _mm_mul_ps(fscal,dz22); |
1802 | |
1803 | /* Update vectorial force */ |
1804 | fix2 = _mm_add_ps(fix2,tx); |
1805 | fiy2 = _mm_add_ps(fiy2,ty); |
1806 | fiz2 = _mm_add_ps(fiz2,tz); |
1807 | |
1808 | fjx2 = _mm_add_ps(fjx2,tx); |
1809 | fjy2 = _mm_add_ps(fjy2,ty); |
1810 | fjz2 = _mm_add_ps(fjz2,tz); |
1811 | |
1812 | fjptrA = f+j_coord_offsetA; |
1813 | fjptrB = f+j_coord_offsetB; |
1814 | fjptrC = f+j_coord_offsetC; |
1815 | fjptrD = f+j_coord_offsetD; |
1816 | |
1817 | gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
1818 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2); |
1819 | |
1820 | /* Inner loop uses 331 flops */ |
1821 | } |
1822 | |
1823 | if(jidx<j_index_end) |
1824 | { |
1825 | |
1826 | /* Get j neighbor index, and coordinate index */ |
1827 | jnrlistA = jjnr[jidx]; |
1828 | jnrlistB = jjnr[jidx+1]; |
1829 | jnrlistC = jjnr[jidx+2]; |
1830 | jnrlistD = jjnr[jidx+3]; |
1831 | /* Sign of each element will be negative for non-real atoms. |
1832 | * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones, |
1833 | * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries. |
1834 | */ |
1835 | dummy_mask = gmx_mm_castsi128_ps_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())); |
1836 | jnrA = (jnrlistA>=0) ? jnrlistA : 0; |
1837 | jnrB = (jnrlistB>=0) ? jnrlistB : 0; |
1838 | jnrC = (jnrlistC>=0) ? jnrlistC : 0; |
1839 | jnrD = (jnrlistD>=0) ? jnrlistD : 0; |
1840 | j_coord_offsetA = DIM3*jnrA; |
1841 | j_coord_offsetB = DIM3*jnrB; |
1842 | j_coord_offsetC = DIM3*jnrC; |
1843 | j_coord_offsetD = DIM3*jnrD; |
1844 | |
1845 | /* load j atom coordinates */ |
1846 | gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
1847 | x+j_coord_offsetC,x+j_coord_offsetD, |
1848 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2); |
1849 | |
1850 | /* Calculate displacement vector */ |
1851 | dx00 = _mm_sub_ps(ix0,jx0); |
1852 | dy00 = _mm_sub_ps(iy0,jy0); |
1853 | dz00 = _mm_sub_ps(iz0,jz0); |
1854 | dx01 = _mm_sub_ps(ix0,jx1); |
1855 | dy01 = _mm_sub_ps(iy0,jy1); |
1856 | dz01 = _mm_sub_ps(iz0,jz1); |
1857 | dx02 = _mm_sub_ps(ix0,jx2); |
1858 | dy02 = _mm_sub_ps(iy0,jy2); |
1859 | dz02 = _mm_sub_ps(iz0,jz2); |
1860 | dx10 = _mm_sub_ps(ix1,jx0); |
1861 | dy10 = _mm_sub_ps(iy1,jy0); |
1862 | dz10 = _mm_sub_ps(iz1,jz0); |
1863 | dx11 = _mm_sub_ps(ix1,jx1); |
1864 | dy11 = _mm_sub_ps(iy1,jy1); |
1865 | dz11 = _mm_sub_ps(iz1,jz1); |
1866 | dx12 = _mm_sub_ps(ix1,jx2); |
1867 | dy12 = _mm_sub_ps(iy1,jy2); |
1868 | dz12 = _mm_sub_ps(iz1,jz2); |
1869 | dx20 = _mm_sub_ps(ix2,jx0); |
1870 | dy20 = _mm_sub_ps(iy2,jy0); |
1871 | dz20 = _mm_sub_ps(iz2,jz0); |
1872 | dx21 = _mm_sub_ps(ix2,jx1); |
1873 | dy21 = _mm_sub_ps(iy2,jy1); |
1874 | dz21 = _mm_sub_ps(iz2,jz1); |
1875 | dx22 = _mm_sub_ps(ix2,jx2); |
1876 | dy22 = _mm_sub_ps(iy2,jy2); |
1877 | dz22 = _mm_sub_ps(iz2,jz2); |
1878 | |
1879 | /* Calculate squared distance and things based on it */ |
1880 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
1881 | rsq01 = gmx_mm_calc_rsq_ps(dx01,dy01,dz01); |
1882 | rsq02 = gmx_mm_calc_rsq_ps(dx02,dy02,dz02); |
1883 | rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10); |
1884 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
1885 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
1886 | rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20); |
1887 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
1888 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
1889 | |
1890 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
1891 | rinv01 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq01); |
1892 | rinv02 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq02); |
1893 | rinv10 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq10); |
1894 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
1895 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
1896 | rinv20 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq20); |
1897 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
1898 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
1899 | |
1900 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
1901 | rinvsq01 = _mm_mul_ps(rinv01,rinv01); |
1902 | rinvsq02 = _mm_mul_ps(rinv02,rinv02); |
1903 | rinvsq10 = _mm_mul_ps(rinv10,rinv10); |
1904 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
1905 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
1906 | rinvsq20 = _mm_mul_ps(rinv20,rinv20); |
1907 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
1908 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
1909 | |
1910 | fjx0 = _mm_setzero_ps(); |
1911 | fjy0 = _mm_setzero_ps(); |
1912 | fjz0 = _mm_setzero_ps(); |
1913 | fjx1 = _mm_setzero_ps(); |
1914 | fjy1 = _mm_setzero_ps(); |
1915 | fjz1 = _mm_setzero_ps(); |
1916 | fjx2 = _mm_setzero_ps(); |
1917 | fjy2 = _mm_setzero_ps(); |
1918 | fjz2 = _mm_setzero_ps(); |
1919 | |
1920 | /************************** |
1921 | * CALCULATE INTERACTIONS * |
1922 | **************************/ |
1923 | |
1924 | r00 = _mm_mul_ps(rsq00,rinv00); |
1925 | r00 = _mm_andnot_ps(dummy_mask,r00); |
1926 | |
1927 | /* EWALD ELECTROSTATICS */ |
1928 | |
1929 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1930 | ewrt = _mm_mul_ps(r00,ewtabscale); |
1931 | ewitab = _mm_cvttps_epi32(ewrt); |
1932 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1933 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1934 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1935 | &ewtabF,&ewtabFn); |
1936 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1937 | felec = _mm_mul_ps(_mm_mul_ps(qq00,rinv00),_mm_sub_ps(rinvsq00,felec)); |
1938 | |
1939 | /* LENNARD-JONES DISPERSION/REPULSION */ |
1940 | |
1941 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
1942 | fvdw = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_00,rinvsix),c6_00),_mm_mul_ps(rinvsix,rinvsq00)); |
1943 | |
1944 | fscal = _mm_add_ps(felec,fvdw); |
1945 | |
1946 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1947 | |
1948 | /* Calculate temporary vectorial force */ |
1949 | tx = _mm_mul_ps(fscal,dx00); |
1950 | ty = _mm_mul_ps(fscal,dy00); |
1951 | tz = _mm_mul_ps(fscal,dz00); |
1952 | |
1953 | /* Update vectorial force */ |
1954 | fix0 = _mm_add_ps(fix0,tx); |
1955 | fiy0 = _mm_add_ps(fiy0,ty); |
1956 | fiz0 = _mm_add_ps(fiz0,tz); |
1957 | |
1958 | fjx0 = _mm_add_ps(fjx0,tx); |
1959 | fjy0 = _mm_add_ps(fjy0,ty); |
1960 | fjz0 = _mm_add_ps(fjz0,tz); |
1961 | |
1962 | /************************** |
1963 | * CALCULATE INTERACTIONS * |
1964 | **************************/ |
1965 | |
1966 | r01 = _mm_mul_ps(rsq01,rinv01); |
1967 | r01 = _mm_andnot_ps(dummy_mask,r01); |
1968 | |
1969 | /* EWALD ELECTROSTATICS */ |
1970 | |
1971 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1972 | ewrt = _mm_mul_ps(r01,ewtabscale); |
1973 | ewitab = _mm_cvttps_epi32(ewrt); |
1974 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1975 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1976 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1977 | &ewtabF,&ewtabFn); |
1978 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1979 | felec = _mm_mul_ps(_mm_mul_ps(qq01,rinv01),_mm_sub_ps(rinvsq01,felec)); |
1980 | |
1981 | fscal = felec; |
1982 | |
1983 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1984 | |
1985 | /* Calculate temporary vectorial force */ |
1986 | tx = _mm_mul_ps(fscal,dx01); |
1987 | ty = _mm_mul_ps(fscal,dy01); |
1988 | tz = _mm_mul_ps(fscal,dz01); |
1989 | |
1990 | /* Update vectorial force */ |
1991 | fix0 = _mm_add_ps(fix0,tx); |
1992 | fiy0 = _mm_add_ps(fiy0,ty); |
1993 | fiz0 = _mm_add_ps(fiz0,tz); |
1994 | |
1995 | fjx1 = _mm_add_ps(fjx1,tx); |
1996 | fjy1 = _mm_add_ps(fjy1,ty); |
1997 | fjz1 = _mm_add_ps(fjz1,tz); |
1998 | |
1999 | /************************** |
2000 | * CALCULATE INTERACTIONS * |
2001 | **************************/ |
2002 | |
2003 | r02 = _mm_mul_ps(rsq02,rinv02); |
2004 | r02 = _mm_andnot_ps(dummy_mask,r02); |
2005 | |
2006 | /* EWALD ELECTROSTATICS */ |
2007 | |
2008 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2009 | ewrt = _mm_mul_ps(r02,ewtabscale); |
2010 | ewitab = _mm_cvttps_epi32(ewrt); |
2011 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2012 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2013 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2014 | &ewtabF,&ewtabFn); |
2015 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2016 | felec = _mm_mul_ps(_mm_mul_ps(qq02,rinv02),_mm_sub_ps(rinvsq02,felec)); |
2017 | |
2018 | fscal = felec; |
2019 | |
2020 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2021 | |
2022 | /* Calculate temporary vectorial force */ |
2023 | tx = _mm_mul_ps(fscal,dx02); |
2024 | ty = _mm_mul_ps(fscal,dy02); |
2025 | tz = _mm_mul_ps(fscal,dz02); |
2026 | |
2027 | /* Update vectorial force */ |
2028 | fix0 = _mm_add_ps(fix0,tx); |
2029 | fiy0 = _mm_add_ps(fiy0,ty); |
2030 | fiz0 = _mm_add_ps(fiz0,tz); |
2031 | |
2032 | fjx2 = _mm_add_ps(fjx2,tx); |
2033 | fjy2 = _mm_add_ps(fjy2,ty); |
2034 | fjz2 = _mm_add_ps(fjz2,tz); |
2035 | |
2036 | /************************** |
2037 | * CALCULATE INTERACTIONS * |
2038 | **************************/ |
2039 | |
2040 | r10 = _mm_mul_ps(rsq10,rinv10); |
2041 | r10 = _mm_andnot_ps(dummy_mask,r10); |
2042 | |
2043 | /* EWALD ELECTROSTATICS */ |
2044 | |
2045 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2046 | ewrt = _mm_mul_ps(r10,ewtabscale); |
2047 | ewitab = _mm_cvttps_epi32(ewrt); |
2048 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2049 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2050 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2051 | &ewtabF,&ewtabFn); |
2052 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2053 | felec = _mm_mul_ps(_mm_mul_ps(qq10,rinv10),_mm_sub_ps(rinvsq10,felec)); |
2054 | |
2055 | fscal = felec; |
2056 | |
2057 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2058 | |
2059 | /* Calculate temporary vectorial force */ |
2060 | tx = _mm_mul_ps(fscal,dx10); |
2061 | ty = _mm_mul_ps(fscal,dy10); |
2062 | tz = _mm_mul_ps(fscal,dz10); |
2063 | |
2064 | /* Update vectorial force */ |
2065 | fix1 = _mm_add_ps(fix1,tx); |
2066 | fiy1 = _mm_add_ps(fiy1,ty); |
2067 | fiz1 = _mm_add_ps(fiz1,tz); |
2068 | |
2069 | fjx0 = _mm_add_ps(fjx0,tx); |
2070 | fjy0 = _mm_add_ps(fjy0,ty); |
2071 | fjz0 = _mm_add_ps(fjz0,tz); |
2072 | |
2073 | /************************** |
2074 | * CALCULATE INTERACTIONS * |
2075 | **************************/ |
2076 | |
2077 | r11 = _mm_mul_ps(rsq11,rinv11); |
2078 | r11 = _mm_andnot_ps(dummy_mask,r11); |
2079 | |
2080 | /* EWALD ELECTROSTATICS */ |
2081 | |
2082 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2083 | ewrt = _mm_mul_ps(r11,ewtabscale); |
2084 | ewitab = _mm_cvttps_epi32(ewrt); |
2085 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2086 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2087 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2088 | &ewtabF,&ewtabFn); |
2089 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2090 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
2091 | |
2092 | fscal = felec; |
2093 | |
2094 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2095 | |
2096 | /* Calculate temporary vectorial force */ |
2097 | tx = _mm_mul_ps(fscal,dx11); |
2098 | ty = _mm_mul_ps(fscal,dy11); |
2099 | tz = _mm_mul_ps(fscal,dz11); |
2100 | |
2101 | /* Update vectorial force */ |
2102 | fix1 = _mm_add_ps(fix1,tx); |
2103 | fiy1 = _mm_add_ps(fiy1,ty); |
2104 | fiz1 = _mm_add_ps(fiz1,tz); |
2105 | |
2106 | fjx1 = _mm_add_ps(fjx1,tx); |
2107 | fjy1 = _mm_add_ps(fjy1,ty); |
2108 | fjz1 = _mm_add_ps(fjz1,tz); |
2109 | |
2110 | /************************** |
2111 | * CALCULATE INTERACTIONS * |
2112 | **************************/ |
2113 | |
2114 | r12 = _mm_mul_ps(rsq12,rinv12); |
2115 | r12 = _mm_andnot_ps(dummy_mask,r12); |
2116 | |
2117 | /* EWALD ELECTROSTATICS */ |
2118 | |
2119 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2120 | ewrt = _mm_mul_ps(r12,ewtabscale); |
2121 | ewitab = _mm_cvttps_epi32(ewrt); |
2122 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2123 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2124 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2125 | &ewtabF,&ewtabFn); |
2126 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2127 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
2128 | |
2129 | fscal = felec; |
2130 | |
2131 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2132 | |
2133 | /* Calculate temporary vectorial force */ |
2134 | tx = _mm_mul_ps(fscal,dx12); |
2135 | ty = _mm_mul_ps(fscal,dy12); |
2136 | tz = _mm_mul_ps(fscal,dz12); |
2137 | |
2138 | /* Update vectorial force */ |
2139 | fix1 = _mm_add_ps(fix1,tx); |
2140 | fiy1 = _mm_add_ps(fiy1,ty); |
2141 | fiz1 = _mm_add_ps(fiz1,tz); |
2142 | |
2143 | fjx2 = _mm_add_ps(fjx2,tx); |
2144 | fjy2 = _mm_add_ps(fjy2,ty); |
2145 | fjz2 = _mm_add_ps(fjz2,tz); |
2146 | |
2147 | /************************** |
2148 | * CALCULATE INTERACTIONS * |
2149 | **************************/ |
2150 | |
2151 | r20 = _mm_mul_ps(rsq20,rinv20); |
2152 | r20 = _mm_andnot_ps(dummy_mask,r20); |
2153 | |
2154 | /* EWALD ELECTROSTATICS */ |
2155 | |
2156 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2157 | ewrt = _mm_mul_ps(r20,ewtabscale); |
2158 | ewitab = _mm_cvttps_epi32(ewrt); |
2159 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2160 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2161 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2162 | &ewtabF,&ewtabFn); |
2163 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2164 | felec = _mm_mul_ps(_mm_mul_ps(qq20,rinv20),_mm_sub_ps(rinvsq20,felec)); |
2165 | |
2166 | fscal = felec; |
2167 | |
2168 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2169 | |
2170 | /* Calculate temporary vectorial force */ |
2171 | tx = _mm_mul_ps(fscal,dx20); |
2172 | ty = _mm_mul_ps(fscal,dy20); |
2173 | tz = _mm_mul_ps(fscal,dz20); |
2174 | |
2175 | /* Update vectorial force */ |
2176 | fix2 = _mm_add_ps(fix2,tx); |
2177 | fiy2 = _mm_add_ps(fiy2,ty); |
2178 | fiz2 = _mm_add_ps(fiz2,tz); |
2179 | |
2180 | fjx0 = _mm_add_ps(fjx0,tx); |
2181 | fjy0 = _mm_add_ps(fjy0,ty); |
2182 | fjz0 = _mm_add_ps(fjz0,tz); |
2183 | |
2184 | /************************** |
2185 | * CALCULATE INTERACTIONS * |
2186 | **************************/ |
2187 | |
2188 | r21 = _mm_mul_ps(rsq21,rinv21); |
2189 | r21 = _mm_andnot_ps(dummy_mask,r21); |
2190 | |
2191 | /* EWALD ELECTROSTATICS */ |
2192 | |
2193 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2194 | ewrt = _mm_mul_ps(r21,ewtabscale); |
2195 | ewitab = _mm_cvttps_epi32(ewrt); |
2196 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2197 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2198 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2199 | &ewtabF,&ewtabFn); |
2200 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2201 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
2202 | |
2203 | fscal = felec; |
2204 | |
2205 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2206 | |
2207 | /* Calculate temporary vectorial force */ |
2208 | tx = _mm_mul_ps(fscal,dx21); |
2209 | ty = _mm_mul_ps(fscal,dy21); |
2210 | tz = _mm_mul_ps(fscal,dz21); |
2211 | |
2212 | /* Update vectorial force */ |
2213 | fix2 = _mm_add_ps(fix2,tx); |
2214 | fiy2 = _mm_add_ps(fiy2,ty); |
2215 | fiz2 = _mm_add_ps(fiz2,tz); |
2216 | |
2217 | fjx1 = _mm_add_ps(fjx1,tx); |
2218 | fjy1 = _mm_add_ps(fjy1,ty); |
2219 | fjz1 = _mm_add_ps(fjz1,tz); |
2220 | |
2221 | /************************** |
2222 | * CALCULATE INTERACTIONS * |
2223 | **************************/ |
2224 | |
2225 | r22 = _mm_mul_ps(rsq22,rinv22); |
2226 | r22 = _mm_andnot_ps(dummy_mask,r22); |
2227 | |
2228 | /* EWALD ELECTROSTATICS */ |
2229 | |
2230 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2231 | ewrt = _mm_mul_ps(r22,ewtabscale); |
2232 | ewitab = _mm_cvttps_epi32(ewrt); |
2233 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2234 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2235 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2236 | &ewtabF,&ewtabFn); |
2237 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2238 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
2239 | |
2240 | fscal = felec; |
2241 | |
2242 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2243 | |
2244 | /* Calculate temporary vectorial force */ |
2245 | tx = _mm_mul_ps(fscal,dx22); |
2246 | ty = _mm_mul_ps(fscal,dy22); |
2247 | tz = _mm_mul_ps(fscal,dz22); |
2248 | |
2249 | /* Update vectorial force */ |
2250 | fix2 = _mm_add_ps(fix2,tx); |
2251 | fiy2 = _mm_add_ps(fiy2,ty); |
2252 | fiz2 = _mm_add_ps(fiz2,tz); |
2253 | |
2254 | fjx2 = _mm_add_ps(fjx2,tx); |
2255 | fjy2 = _mm_add_ps(fjy2,ty); |
2256 | fjz2 = _mm_add_ps(fjz2,tz); |
2257 | |
2258 | fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch; |
2259 | fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch; |
2260 | fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch; |
2261 | fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch; |
2262 | |
2263 | gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
2264 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2); |
2265 | |
2266 | /* Inner loop uses 340 flops */ |
2267 | } |
2268 | |
2269 | /* End of innermost loop */ |
2270 | |
2271 | gmx_mm_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2, |
2272 | f+i_coord_offset,fshift+i_shift_offset); |
2273 | |
2274 | /* Increment number of inner iterations */ |
2275 | inneriter += j_index_end - j_index_start; |
2276 | |
2277 | /* Outer loop uses 18 flops */ |
2278 | } |
2279 | |
2280 | /* Increment number of outer iterations */ |
2281 | outeriter += nri; |
2282 | |
2283 | /* Update outer/inner flops */ |
2284 | |
2285 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*18 + inneriter*340)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W3W3_F] += outeriter*18 + inneriter *340; |
2286 | } |