| 1 | /*
|
|---|
| 2 | A simple 2D hydro code
|
|---|
| 3 | (C) Romain Teyssier : CEA/IRFU -- original F90 code
|
|---|
| 4 | (C) Pierre-Francois Lavallee : IDRIS -- original F90 code
|
|---|
| 5 | (C) Guillaume Colin de Verdiere : CEA/DAM -- for the C version
|
|---|
| 6 | */
|
|---|
| 7 |
|
|---|
| 8 | /*
|
|---|
| 9 |
|
|---|
| 10 | This software is governed by the CeCILL license under French law and
|
|---|
| 11 | abiding by the rules of distribution of free software. You can use,
|
|---|
| 12 | modify and/ or redistribute the software under the terms of the CeCILL
|
|---|
| 13 | license as circulated by CEA, CNRS and INRIA at the following URL
|
|---|
| 14 | "http://www.cecill.info".
|
|---|
| 15 |
|
|---|
| 16 | As a counterpart to the access to the source code and rights to copy,
|
|---|
| 17 | modify and redistribute granted by the license, users are provided only
|
|---|
| 18 | with a limited warranty and the software's author, the holder of the
|
|---|
| 19 | economic rights, and the successive licensors have only limited
|
|---|
| 20 | liability.
|
|---|
| 21 |
|
|---|
| 22 | In this respect, the user's attention is drawn to the risks associated
|
|---|
| 23 | with loading, using, modifying and/or developing or reproducing the
|
|---|
| 24 | software by the user in light of its specific status of free software,
|
|---|
| 25 | that may mean that it is complicated to manipulate, and that also
|
|---|
| 26 | therefore means that it is reserved for developers and experienced
|
|---|
| 27 | professionals having in-depth computer knowledge. Users are therefore
|
|---|
| 28 | encouraged to load and test the software's suitability as regards their
|
|---|
| 29 | requirements in conditions enabling the security of their systems and/or
|
|---|
| 30 | data to be ensured and, more generally, to use and operate it in the
|
|---|
| 31 | same conditions as regards security.
|
|---|
| 32 |
|
|---|
| 33 | The fact that you are presently reading this means that you have had
|
|---|
| 34 | knowledge of the CeCILL license and that you accept its terms.
|
|---|
| 35 |
|
|---|
| 36 | */
|
|---|
| 37 |
|
|---|
| 38 | #include <stdlib.h>
|
|---|
| 39 | #include <unistd.h>
|
|---|
| 40 | #include <math.h>
|
|---|
| 41 | #include <stdio.h>
|
|---|
| 42 |
|
|---|
| 43 | #include "parametres.h"
|
|---|
| 44 | #include "perfcnt.h"
|
|---|
| 45 | #include "utils.h"
|
|---|
| 46 | #include "riemann.h"
|
|---|
| 47 |
|
|---|
| 48 | #ifdef HMPP
|
|---|
| 49 | #undef HMPP
|
|---|
| 50 | #include "constoprim.c"
|
|---|
| 51 | #include "equation_of_state.c"
|
|---|
| 52 | #include "slope.c"
|
|---|
| 53 | #include "trace.c"
|
|---|
| 54 | #include "qleftright.c"
|
|---|
| 55 | #include "cmpflx.c"
|
|---|
| 56 | #include "conservar.c"
|
|---|
| 57 | #define HMPP
|
|---|
| 58 | #endif
|
|---|
| 59 |
|
|---|
| 60 | #define PRECISION 1e-6
|
|---|
| 61 |
|
|---|
| 62 | void
|
|---|
| 63 | Dmemset(size_t nbr, real_t t[nbr], real_t motif) {
|
|---|
| 64 | int i;
|
|---|
| 65 | for (i = 0; i < nbr; i++) {
|
|---|
| 66 | t[i] = motif;
|
|---|
| 67 | }
|
|---|
| 68 | }
|
|---|
| 69 |
|
|---|
| 70 |
|
|---|
| 71 | #define DABS(x) (real_t) fabs((x))
|
|---|
| 72 | #ifdef HMPP
|
|---|
| 73 | #define MAX(x,y) fmax(x,y)
|
|---|
| 74 | #endif
|
|---|
| 75 |
|
|---|
| 76 | #define MYSQRT sqrt
|
|---|
| 77 |
|
|---|
| 78 | void
|
|---|
| 79 | riemann(int narray, const real_t Hsmallr,
|
|---|
| 80 | const real_t Hsmallc, const real_t Hgamma,
|
|---|
| 81 | const int Hniter_riemann, const int Hnvar,
|
|---|
| 82 | const int Hnxyt, const int slices,
|
|---|
| 83 | const int Hstep,
|
|---|
| 84 | real_t qleft[Hnvar][Hstep][Hnxyt],
|
|---|
| 85 | real_t qright[Hnvar][Hstep][Hnxyt], //
|
|---|
| 86 | real_t qgdnv[Hnvar][Hstep][Hnxyt], //
|
|---|
| 87 | int sgnm[Hstep][Hnxyt],
|
|---|
| 88 | hydrowork_t * Hw)
|
|---|
| 89 | {
|
|---|
| 90 | int i, s, ii, iimx;
|
|---|
| 91 | real_t smallp_ = Square(Hsmallc) / Hgamma;
|
|---|
| 92 | real_t gamma6_ = (Hgamma + one) / (two * Hgamma);
|
|---|
| 93 | real_t smallpp_ = Hsmallr * smallp_;
|
|---|
| 94 |
|
|---|
| 95 | FLOPS(4, 2, 0, 0);
|
|---|
| 96 | // __declspec(align(256)) thevariable
|
|---|
| 97 |
|
|---|
| 98 | int *Fgoon = Hw->goon;
|
|---|
| 99 | real_t *Fpstar = Hw->pstar;
|
|---|
| 100 | real_t *Frl = Hw->rl;
|
|---|
| 101 | real_t *Ful = Hw->ul;
|
|---|
| 102 | real_t *Fpl = Hw->pl;
|
|---|
| 103 | real_t *Fur = Hw->ur;
|
|---|
| 104 | real_t *Fpr = Hw->pr;
|
|---|
| 105 | real_t *Fcl = Hw->cl;
|
|---|
| 106 | real_t *Fcr = Hw->cr;
|
|---|
| 107 | real_t *Frr = Hw->rr;
|
|---|
| 108 |
|
|---|
| 109 | real_t smallp = smallp_;
|
|---|
| 110 | real_t gamma6 = gamma6_;
|
|---|
| 111 | real_t smallpp = smallpp_;
|
|---|
| 112 |
|
|---|
| 113 | // fprintf(stderr, "%d\n", __ICC );
|
|---|
| 114 | #pragma message "active pragma simd "
|
|---|
| 115 | #define SIMDNEEDED 1
|
|---|
| 116 | #if __ICC < 1300
|
|---|
| 117 | #define SIMD ivdep
|
|---|
| 118 | #else
|
|---|
| 119 | #define SIMD simd
|
|---|
| 120 | #endif
|
|---|
| 121 | // #define SIMD novector
|
|---|
| 122 |
|
|---|
| 123 | // Pressure, density and velocity
|
|---|
| 124 | #pragma omp parallel for private(s, i), shared(qgdnv, sgnm) reduction(+:flopsAri), reduction(+:flopsSqr), reduction(+:flopsMin), reduction(+:flopsTra)
|
|---|
| 125 | for (s = 0; s < slices; s++) {
|
|---|
| 126 | int ii, iimx;
|
|---|
| 127 | int *goon;
|
|---|
| 128 | real_t *pstar, *rl, *ul, *pl, *rr, *ur, *pr, *cl, *cr;
|
|---|
| 129 | int iter;
|
|---|
| 130 | pstar = &Fpstar[s * narray];
|
|---|
| 131 | rl = &Frl[s * narray];
|
|---|
| 132 | ul = &Ful[s * narray];
|
|---|
| 133 | pl = &Fpl[s * narray];
|
|---|
| 134 | rr = &Frr[s * narray];
|
|---|
| 135 | ur = &Fur[s * narray];
|
|---|
| 136 | pr = &Fpr[s * narray];
|
|---|
| 137 | cl = &Fcl[s * narray];
|
|---|
| 138 | cr = &Fcr[s * narray];
|
|---|
| 139 | goon = &Fgoon[s * narray];
|
|---|
| 140 |
|
|---|
| 141 | // Precompute values for this slice
|
|---|
| 142 |
|
|---|
| 143 | #ifdef SIMDNEEDED
|
|---|
| 144 | #if __ICC < 1300
|
|---|
| 145 | #pragma ivdep
|
|---|
| 146 | #else
|
|---|
| 147 | #pragma SIMD
|
|---|
| 148 | #endif
|
|---|
| 149 | #endif
|
|---|
| 150 | for (i = 0; i < narray; i++) {
|
|---|
| 151 | rl[i] = fmax(qleft[ID][s][i], Hsmallr);
|
|---|
| 152 | ul[i] = qleft[IU][s][i];
|
|---|
| 153 | pl[i] = fmax(qleft[IP][s][i], (real_t) (rl[i] * smallp));
|
|---|
| 154 | rr[i] = fmax(qright[ID][s][i], Hsmallr);
|
|---|
| 155 | ur[i] = qright[IU][s][i];
|
|---|
| 156 | pr[i] = fmax(qright[IP][s][i], (real_t) (rr[i] * smallp));
|
|---|
| 157 |
|
|---|
| 158 | // Lagrangian sound speed
|
|---|
| 159 | cl[i] = Hgamma * pl[i] * rl[i];
|
|---|
| 160 | cr[i] = Hgamma * pr[i] * rr[i];
|
|---|
| 161 | // First guess
|
|---|
| 162 |
|
|---|
| 163 | real_t wl_i = MYSQRT(cl[i]);
|
|---|
| 164 | real_t wr_i = MYSQRT(cr[i]);
|
|---|
| 165 | pstar[i] = fmax(((wr_i * pl[i] + wl_i * pr[i]) + wl_i * wr_i * (ul[i] - ur[i])) / (wl_i + wr_i), 0.0);
|
|---|
| 166 | goon[i] = 1;
|
|---|
| 167 | }
|
|---|
| 168 |
|
|---|
| 169 | #define Fmax(a,b) (((a) > (b)) ? (a): (b))
|
|---|
| 170 | #define Fabs(a) (((a) > 0) ? (a): -(a))
|
|---|
| 171 |
|
|---|
| 172 | // solve the riemann problem on the interfaces of this slice
|
|---|
| 173 | for (iter = 0; iter < Hniter_riemann; iter++) {
|
|---|
| 174 | #ifdef SIMDNEEDED
|
|---|
| 175 | #if __ICC < 1300
|
|---|
| 176 | #pragma simd
|
|---|
| 177 | #else
|
|---|
| 178 | #pragma SIMD
|
|---|
| 179 | #endif
|
|---|
| 180 | #endif
|
|---|
| 181 | for (i = 0; i < narray; i++) {
|
|---|
| 182 | if (goon[i]) {
|
|---|
| 183 | real_t pst = pstar[i];
|
|---|
| 184 | // Newton-Raphson iterations to find pstar at the required accuracy
|
|---|
| 185 | real_t wwl = MYSQRT(cl[i] * (one + gamma6 * (pst - pl[i]) / pl[i]));
|
|---|
| 186 | real_t wwr = MYSQRT(cr[i] * (one + gamma6 * (pst - pr[i]) / pr[i]));
|
|---|
| 187 | real_t swwl = Square(wwl);
|
|---|
| 188 | real_t ql = two * wwl * swwl / (swwl + cl[i]);
|
|---|
| 189 | real_t qr = two * wwr * Square(wwr) / (Square(wwr) + cr[i]);
|
|---|
| 190 | real_t usl = ul[i] - (pst - pl[i]) / wwl;
|
|---|
| 191 | real_t usr = ur[i] + (pst - pr[i]) / wwr;
|
|---|
| 192 | real_t tmp = (qr * ql / (qr + ql) * (usl - usr));
|
|---|
| 193 | real_t delp_i = Fmax(tmp, (-pst));
|
|---|
| 194 | // pstar[i] = pstar[i] + delp_i;
|
|---|
| 195 | pst += delp_i;
|
|---|
| 196 | // Convergence indicator
|
|---|
| 197 | real_t tmp2 = delp_i / (pst + smallpp);
|
|---|
| 198 | real_t uo_i = Fabs(tmp2);
|
|---|
| 199 | goon[i] = uo_i > PRECISION;
|
|---|
| 200 | // FLOPS(29, 10, 2, 0);
|
|---|
| 201 | pstar[i] = pst;
|
|---|
| 202 | }
|
|---|
| 203 | }
|
|---|
| 204 | } // iter_riemann
|
|---|
| 205 |
|
|---|
| 206 | #ifdef SIMDNEEDED
|
|---|
| 207 | #pragma SIMD
|
|---|
| 208 | #endif
|
|---|
| 209 | for (i = 0; i < narray; i++) {
|
|---|
| 210 | real_t wl_i = MYSQRT(cl[i]);
|
|---|
| 211 | real_t wr_i = MYSQRT(cr[i]);
|
|---|
| 212 |
|
|---|
| 213 | wr_i = MYSQRT(cr[i] * (one + gamma6 * (pstar[i] - pr[i]) / pr[i]));
|
|---|
| 214 | wl_i = MYSQRT(cl[i] * (one + gamma6 * (pstar[i] - pl[i]) / pl[i]));
|
|---|
| 215 |
|
|---|
| 216 | real_t ustar_i = half * (ul[i] + (pl[i] - pstar[i]) / wl_i + ur[i] - (pr[i] - pstar[i]) / wr_i);
|
|---|
| 217 |
|
|---|
| 218 | int left = ustar_i > 0;
|
|---|
| 219 |
|
|---|
| 220 | real_t ro_i, uo_i, po_i, wo_i;
|
|---|
| 221 |
|
|---|
| 222 | if (left) {
|
|---|
| 223 | sgnm[s][i] = 1;
|
|---|
| 224 | ro_i = rl[i];
|
|---|
| 225 | uo_i = ul[i];
|
|---|
| 226 | po_i = pl[i];
|
|---|
| 227 | wo_i = wl_i;
|
|---|
| 228 | } else {
|
|---|
| 229 | sgnm[s][i] = -1;
|
|---|
| 230 | ro_i = rr[i];
|
|---|
| 231 | uo_i = ur[i];
|
|---|
| 232 | po_i = pr[i];
|
|---|
| 233 | wo_i = wr_i;
|
|---|
| 234 | }
|
|---|
| 235 |
|
|---|
| 236 | real_t co_i = MYSQRT(fabs(Hgamma * po_i / ro_i));
|
|---|
| 237 | co_i = fmax(Hsmallc, co_i);
|
|---|
| 238 |
|
|---|
| 239 | real_t rstar_i = ro_i / (one + ro_i * (po_i - pstar[i]) / Square(wo_i));
|
|---|
| 240 | rstar_i = fmax(rstar_i, Hsmallr);
|
|---|
| 241 |
|
|---|
| 242 | real_t cstar_i = MYSQRT(fabs(Hgamma * pstar[i] / rstar_i));
|
|---|
| 243 | cstar_i = fmax(Hsmallc, cstar_i);
|
|---|
| 244 |
|
|---|
| 245 | real_t spout_i = co_i - sgnm[s][i] * uo_i;
|
|---|
| 246 | real_t spin_i = cstar_i - sgnm[s][i] * ustar_i;
|
|---|
| 247 | real_t ushock_i = wo_i / ro_i - sgnm[s][i] * uo_i;
|
|---|
| 248 |
|
|---|
| 249 | if (pstar[i] >= po_i) {
|
|---|
| 250 | spin_i = ushock_i;
|
|---|
| 251 | spout_i = ushock_i;
|
|---|
| 252 | }
|
|---|
| 253 |
|
|---|
| 254 | real_t scr_i = fmax((real_t) (spout_i - spin_i), (real_t) (Hsmallc + fabs(spout_i + spin_i)));
|
|---|
| 255 |
|
|---|
| 256 | real_t frac_i = (one + (spout_i + spin_i) / scr_i) * half;
|
|---|
| 257 | frac_i = fmax(zero, (real_t) (fmin(one, frac_i)));
|
|---|
| 258 |
|
|---|
| 259 | int addSpout = spout_i < zero;
|
|---|
| 260 | int addSpin = spin_i > zero;
|
|---|
| 261 | // real_t originalQgdnv = !addSpout & !addSpin;
|
|---|
| 262 | real_t qgdnv_ID, qgdnv_IU, qgdnv_IP;
|
|---|
| 263 |
|
|---|
| 264 | if (addSpout) {
|
|---|
| 265 | qgdnv_ID = ro_i;
|
|---|
| 266 | qgdnv_IU = uo_i;
|
|---|
| 267 | qgdnv_IP = po_i;
|
|---|
| 268 | } else if (addSpin) {
|
|---|
| 269 | qgdnv_ID = rstar_i;
|
|---|
| 270 | qgdnv_IU = ustar_i;
|
|---|
| 271 | qgdnv_IP = pstar[i];
|
|---|
| 272 | } else {
|
|---|
| 273 | qgdnv_ID = (frac_i * rstar_i + (one - frac_i) * ro_i);
|
|---|
| 274 | qgdnv_IU = (frac_i * ustar_i + (one - frac_i) * uo_i);
|
|---|
| 275 | qgdnv_IP = (frac_i * pstar[i] + (one - frac_i) * po_i);
|
|---|
| 276 | }
|
|---|
| 277 |
|
|---|
| 278 | qgdnv[ID][s][i] = qgdnv_ID;
|
|---|
| 279 | qgdnv[IU][s][i] = qgdnv_IU;
|
|---|
| 280 | qgdnv[IP][s][i] = qgdnv_IP;
|
|---|
| 281 |
|
|---|
| 282 | // transverse velocity
|
|---|
| 283 | if (left) {
|
|---|
| 284 | qgdnv[IV][s][i] = qleft[IV][s][i];
|
|---|
| 285 | } else {
|
|---|
| 286 | qgdnv[IV][s][i] = qright[IV][s][i];
|
|---|
| 287 | }
|
|---|
| 288 | }
|
|---|
| 289 | }
|
|---|
| 290 | {
|
|---|
| 291 | int nops = slices * narray;
|
|---|
| 292 | FLOPS(57 * nops, 17 * nops, 14 * nops, 0 * nops);
|
|---|
| 293 | }
|
|---|
| 294 |
|
|---|
| 295 | // other passive variables
|
|---|
| 296 | if (Hnvar > IP) {
|
|---|
| 297 | int invar;
|
|---|
| 298 | for (invar = IP + 1; invar < Hnvar; invar++) {
|
|---|
| 299 | for (s = 0; s < slices; s++) {
|
|---|
| 300 | #ifdef SIMDNEEDED
|
|---|
| 301 | #pragma SIMD
|
|---|
| 302 | #endif
|
|---|
| 303 | for (i = 0; i < narray; i++) {
|
|---|
| 304 | int left = (sgnm[s][i] == 1);
|
|---|
| 305 | qgdnv[invar][s][i] = qleft[invar][s][i] * left + qright[invar][s][i] * !left;
|
|---|
| 306 | }
|
|---|
| 307 | }
|
|---|
| 308 | }
|
|---|
| 309 | }
|
|---|
| 310 | } // riemann_vec
|
|---|
| 311 |
|
|---|
| 312 | //EOF
|
|---|
