/usr/share/doc/libsundials-serial-dev/examples/idas/parallel/idasBruss_ASAp_kry_bbd_p.c is in libsundials-serial-dev 2.5.0-3+b3.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 | /* * -----------------------------------------------------------------
* $Revision:
* $Date:
* -----------------------------------------------------------------
* Programmer(s): Cosmin Petra and Radu Serban @ LLNL
* -----------------------------------------------------------------
* Example program for IDAS: Brusselator, parallel, GMRES, IDABBD
* preconditioner, ASA
*
* This example program for IDAS uses IDASPGMR as the linear solver.
* It is written for a parallel computer system and uses the
* IDABBDPRE band-block-diagonal preconditioner module for the
* IDASPGMR package.
*
* The mathematical problem solved in this example is a DAE system
* that arises from a system of partial differential equations after
* spatial discretization.
*
* The PDE system is a two-species time-dependent PDE known as
* Brusselator PDE and models a chemically reacting system.
*
*
* du/dt = eps(u + u) + u^2 v -(B+1)u + A
* xx yy
* domain Omega = [0,L]X[0,L]
* dv/dt = eps(v + v) - u^2 v + Bu
* xx yy
*
* B.C. : Neumann
* I.C : u(x,y,t0) = u0(x,y) = 1 - 0.5*cos(pi*y/L)
* v(x,y,t0) = v0(x,y) = 3.5 - 2.5*cos(pi*x/L)
*
* The PDEs are discretized by central differencing on a MX by MY
* mesh, and so the system size Neq is the product MX*MY*NUM_SPECIES.
* The system is actually implemented on submeshes, processor by
* processor, with an MXSUB by MYSUB mesh on each of NPEX * NPEY
* processors.
*
*
* The sensitivity of the output functional
* 1 /
* g(t) = ----- | u(x,y,t) ,
* |L^2| /
* Omega
* with respect to initial conditions u0 and v0 is also computed.
* Given the perturbations du0 and dv0 in the IC, the sensitivity of
* of g at final time tf is
* 1 /
* dg(tf) = ----- | ( lambda(0,x,y) du0(x,y) + mu(0,x,y) dv0(x,y) ),
* |L^2| /
* Omega
* where lambda and mu are the solutions of the adjoint PDEs:
*
* dl/dt = - eps(l + l) - (2uv - B - 1)l + (2uv - B)m
* xx yy
* domain Omega = [0,L]X[0,L]
* dm/dt = - eps(m + m) - u^2 l + u^2 m
* xx yy
* B.C. : Neumann
* I.C. : l(x,y,tf) = 1
* m(x,y,tf) = 0
*
* The adjoint PDEs are discretized and solved in the same way as
* the Brusselator PDEs.
*/
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#include <idas/idas.h>
#include <idas/idas_spgmr.h>
#include <idas/idas_bbdpre.h>
#include <nvector/nvector_parallel.h>
#include <sundials/sundials_dense.h>
#include <sundials/sundials_types.h>
#include <sundials/sundials_math.h>
#include <mpi.h>
/* Problem Constants */
#define NUM_SPECIES 2
#define ctL RCONST(1.0) /* Domain =[0,L]^2 */
#define ctA RCONST(1.0)
#define ctB RCONST(3.4)
#define ctEps RCONST(2.0e-3)
#define PI RCONST(3.1415926535898) /* pi */
#define MXSUB 21 /* Number of x mesh points per processor subgrid */
#define MYSUB 21 /* Number of y mesh points per processor subgrid */
#define NPEX 2 /* Number of subgrids in the x direction */
#define NPEY 2 /* Number of subgrids in the y direction */
#define MX (MXSUB*NPEX) /* MX = number of x mesh points */
#define MY (MYSUB*NPEY) /* MY = number of y mesh points */
#define NSMXSUB (NUM_SPECIES * MXSUB)
#define NEQ (NUM_SPECIES*MX*MY) /* Number of equations in system */
#define RTOL RCONST(1.e-5) /* rtol tolerance */
#define ATOL RCONST(1.e-5) /* atol tolerance */
#define TBEGIN RCONST(0.0) /* Multiplier for tout values */
#define TEND RCONST(1.0) /* Increment for tout values */
#define STEPS 50
#define ZERO RCONST(0.0)
#define HALF RCONST(0.5)
#define ONE RCONST(1.0)
#define TWO RCONST(2.0)
/* User-defined vector accessor macro IJ_Vptr. */
/*
* IJ_Vptr is defined in order to express the underlying 3-d structure of the
* dependent variable vector from its underlying 1-d storage (an N_Vector).
* IJ_Vptr(vv,i,j) returns a pointer to the location in vv corresponding to
* species index is = 0, x-index ix = i, and y-index jy = j.
*/
#define IJ_Vptr(vv,i,j) (&NV_Ith_P(vv, (i)*NUM_SPECIES + (j)*NSMXSUB ))
/* Type: UserData. Contains problem constants, preconditioner data, etc. */
typedef struct {
int ns, thispe, npes, ixsub, jysub, npex, npey;
int mxsub, mysub, nsmxsub, nsmxsub2;
realtype A, B, L, eps[NUM_SPECIES];
realtype dx, dy;
realtype cox[NUM_SPECIES], coy[NUM_SPECIES];
realtype gridext[(MXSUB+2)*(MYSUB+2)*NUM_SPECIES];
realtype rhs[NUM_SPECIES];
MPI_Comm comm;
realtype rates[2];
long int n_local;
} *UserData;
/* Prototypes for functions called by the IDA Solver. */
static int res(realtype tt,
N_Vector uv, N_Vector uvp, N_Vector rr,
void *user_data);
static int reslocal(long int Nlocal, realtype tt,
N_Vector uv, N_Vector uvp, N_Vector res,
void *user_data);
static int rescomm(long int Nlocal, realtype tt,
N_Vector uv, N_Vector uvp,
void *user_data);
/* Prototypes for supporting functions */
static void BSend(MPI_Comm comm, int thispe, int ixsub, int jysub,
int dsizex, int dsizey, realtype carray[]);
static void BRecvPost(MPI_Comm comm, MPI_Request request[], int thispe,
int ixsub, int jysub,
int dsizex, int dsizey,
realtype cext[], realtype buffer[]);
static void BRecvWait(MPI_Request request[], int ixsub, int jysub,
int dsizex, realtype cext[], realtype buffer[]);
static void ReactRates(realtype xx, realtype yy, realtype *cxy, realtype *ratesxy,
UserData data);
/* ADJOINT */
static int resB(realtype tt,
N_Vector yy, N_Vector yp,
N_Vector yyB, N_Vector ypB, N_Vector rrB,
void *user_dataB);
static int resBlocal(long int Nlocal, realtype tt,
N_Vector uv, N_Vector uvp,
N_Vector yyB, N_Vector ypB,
N_Vector res, void *user_dataB);
/* Prototypes for private functions */
static void InitUserData(UserData data, int thispe, int npes,
MPI_Comm comm);
static void SetInitialProfiles(N_Vector uv, N_Vector uvp, N_Vector id,
N_Vector resid, UserData data);
static void SetInitialProfilesB(N_Vector uv, N_Vector uvp,
N_Vector uvB, N_Vector uvpB,
N_Vector residB, UserData data);
static void PrintHeader(int SystemSize, int maxl,
long int mudq, long int mldq,
long int mukeep, long int mlkeep,
realtype rtol, realtype atol);
static void PrintOutput(void *mem, N_Vector uv, realtype time,
UserData data, MPI_Comm comm);
static void PrintSol(void* mem, N_Vector uv, N_Vector uvp, UserData data,
MPI_Comm comm);
static void PrintAdjSol(N_Vector uvB, N_Vector uvpB, UserData data);
static void PrintFinalStats(void *mem);
static int check_flag(void *flagvalue, char *funcname, int opt, int id);
/*
*--------------------------------------------------------------------
* MAIN PROGRAM
*--------------------------------------------------------------------
*/
int main(int argc, char *argv[])
{
MPI_Comm comm;
void *mem;
UserData data;
long int SystemSize, local_N, mudq, mldq, mukeep, mlkeep;
realtype rtol, atol, t0, tout, tret;
N_Vector uv, uvp, resid, id, uvB, uvpB, residB, qB;
int thispe, npes, maxl, retval;
int nckpnt, indexB;
uv = uvp = resid = id = NULL;
data = NULL;
mem = NULL;
/* Set communicator, and get processor number and total number of PE's. */
MPI_Init(&argc, &argv);
comm = MPI_COMM_WORLD;
MPI_Comm_rank(comm, &thispe);
MPI_Comm_size(comm, &npes);
if (npes != NPEX*NPEY) {
if (thispe == 0)
fprintf(stderr,
"\nMPI_ERROR(0): npes = %d not equal to NPEX*NPEY = %d\n",
npes, NPEX*NPEY);
MPI_Finalize();
return(1);
}
/* Set local length (local_N) and global length (SystemSize). */
local_N = MXSUB*MYSUB*NUM_SPECIES;
SystemSize = NEQ;
/* Set up user data block data. */
data = (UserData) malloc(sizeof *data);
InitUserData(data, thispe, npes, comm);
/* Create needed vectors, and load initial values.
The vector resid is used temporarily only. */
uv = N_VNew_Parallel(comm, local_N, SystemSize);
if(check_flag((void *)uv, "N_VNew_Parallel", 0, thispe)) MPI_Abort(comm, 1);
uvp = N_VNew_Parallel(comm, local_N, SystemSize);
if(check_flag((void *)uvp, "N_VNew_Parallel", 0, thispe)) MPI_Abort(comm, 1);
resid = N_VNew_Parallel(comm, local_N, SystemSize);
if(check_flag((void *)resid, "N_VNew_Parallel", 0, thispe)) MPI_Abort(comm, 1);
id = N_VNew_Parallel(comm, local_N, SystemSize);
if(check_flag((void *)id, "N_VNew_Parallel", 0, thispe)) MPI_Abort(comm, 1);
SetInitialProfiles(uv, uvp, id, resid, data);
res(ZERO, uv, uvp, resid, data);
/* Set remaining inputs to IDAS. */
t0 = ZERO;
rtol = RTOL;
atol = ATOL;
/* Call IDACreate and IDAInit to initialize solution */
mem = IDACreate();
if(check_flag((void *)mem, "IDACreate", 0, thispe)) MPI_Abort(comm, 1);
retval = IDASetUserData(mem, data);
if(check_flag(&retval, "IDASetUserData", 1, thispe)) MPI_Abort(comm, 1);
retval = IDASetId(mem, id);
if(check_flag(&retval, "IDASetId", 1, thispe)) MPI_Abort(comm, 1);
retval = IDAInit(mem, res, t0, uv, uvp);
if(check_flag(&retval, "IDAInit", 1, thispe)) MPI_Abort(comm, 1);
retval = IDASStolerances(mem, rtol, atol);
if(check_flag(&retval, "IDASStolerances", 1, thispe)) MPI_Abort(comm, 1);
/* Call IDASpgmr to specify the IDAS LINEAR SOLVER IDASPGMR */
maxl = 16;
retval = IDASpgmr(mem, maxl);
if(check_flag(&retval, "IDASpgmr", 1, thispe)) MPI_Abort(comm, 1);
/* Call IDABBDPrecInit to initialize the band-block-diagonal preconditioner.
The half-bandwidths for the difference quotient evaluation are exact
for the system Jacobian, but only a 5-diagonal band matrix is retained. */
mudq = mldq = NSMXSUB;
mukeep = mlkeep = 2;
retval = IDABBDPrecInit(mem, local_N, mudq, mldq, mukeep, mlkeep,
ZERO, reslocal, NULL);
if(check_flag(&retval, "IDABBDPrecInit", 1, thispe)) MPI_Abort(comm, 1);
/* Initialize adjoint module. */
retval = IDAAdjInit(mem, STEPS, IDA_POLYNOMIAL);
if(check_flag(&retval, "IDAAdjInit", 1, thispe)) MPI_Abort(comm, 1);
/* Call IDACalcIC (with default options) to correct the initial values. */
tout = RCONST(0.001);
retval = IDACalcIC(mem, IDA_YA_YDP_INIT, tout);
if(check_flag(&retval, "IDACalcIC", 1, thispe)) MPI_Abort(comm, 1);
if (thispe == 0) printf("\nStarting integration of the FORWARD problem\n\n");
/* On PE 0, print heading, basic parameters, initial values. */
if (thispe == 0) PrintHeader(SystemSize, maxl,
mudq, mldq, mukeep, mlkeep,
rtol, atol);
/* Call IDAS in tout loop, normal mode, and print selected output. */
retval = IDASolveF(mem, TEND, &tret, uv, uvp, IDA_NORMAL, &nckpnt);
if(check_flag(&retval, "IDASolveF", 1, thispe)) MPI_Abort(comm, 1);
PrintOutput(mem, uv, tret, data, comm);
/* Print each PE's portion of the solution in a separate file. */
//PrintSol(mem, uv, uvp, data, comm);
/* On PE 0, print final set of statistics. */
if (thispe == 0) {
PrintFinalStats(mem);
}
/*******************************************************
* ADJOINT *
*******************************************************/
if (thispe == 0) printf("\n\t\t BACKWARD problem\n");
uvB = N_VNew_Parallel(comm, local_N, SystemSize);
uvpB = N_VNew_Parallel(comm, local_N, SystemSize);
residB = N_VNew_Parallel(comm, local_N, SystemSize);
qB = N_VNew_Parallel(comm, local_N, SystemSize);
retval = IDACreateB(mem, &indexB);
/*Get consistent IC */
SetInitialProfilesB(uv, uvp, uvB, uvpB, residB, data);
retval = IDAInitB(mem, indexB, resB, TEND, uvB, uvpB);
if(check_flag(&retval, "IDAInitB", 1, thispe)) MPI_Abort(comm, 1);
retval = IDASetUserDataB(mem, indexB, data);
if(check_flag(&retval, "IDASetUserDataB", 1, thispe)) MPI_Abort(comm, 1);
retval = IDASetIdB(mem, indexB, id);
if(check_flag(&retval, "IDASetIdBIDAInitB", 1, thispe)) MPI_Abort(comm, 1);
retval = IDASStolerancesB(mem, indexB, rtol, atol);
if(check_flag(&retval, "IDASStolerancesB", 1, thispe)) MPI_Abort(comm, 1);
/* Call IDASpgmr to specify the IDAS LINEAR SOLVER IDASPGMR */
maxl = 16;
retval = IDASpgmrB(mem, indexB, maxl);
mudq = mldq = NSMXSUB;
mukeep = mlkeep = 2;
retval = IDABBDPrecInitB(mem, indexB, local_N, mudq, mldq, mukeep, mlkeep,
ZERO, resBlocal, NULL);
if(check_flag(&retval, "IDABBDPrecInitB", 1, thispe)) MPI_Abort(comm, 1);
retval = IDASolveB(mem, TBEGIN, IDA_NORMAL);
if(check_flag(&retval, "IDASolveB", 1, thispe)) MPI_Abort(comm, 1);
retval = IDAGetB(mem, indexB, &tret, uvB, uvpB);
if(check_flag(&retval, "IDAGetB", 1, thispe)) MPI_Abort(comm, 1);
/* Print each PE's portion of solution in a separate file. */
/* PrintAdjSol(uvB, uvpB, data); */
/* On PE 0, print final set of statistics. */
if (thispe == 0) {
PrintFinalStats(IDAGetAdjIDABmem(mem, indexB));
}
/* Free memory. */
N_VDestroy_Parallel(uv);
N_VDestroy_Parallel(uvp);
N_VDestroy_Parallel(id);
N_VDestroy_Parallel(resid);
N_VDestroy_Parallel(uvB);
N_VDestroy_Parallel(uvpB);
N_VDestroy_Parallel(residB);
IDAFree(&mem);
free(data);
MPI_Finalize();
return(0);
}
/*
*--------------------------------------------------------------------
* PRIVATE FUNCTIONS
*--------------------------------------------------------------------
*/
/*
* InitUserData: Load problem constants in data (of type UserData).
*/
static void InitUserData(UserData data, int thispe, int npes,
MPI_Comm comm)
{
data->jysub = thispe / NPEX;
data->ixsub = thispe - (data->jysub)*NPEX;
data->mxsub = MXSUB;
data->mysub = MYSUB;
data->npex = NPEX;
data->npey = NPEY;
data->ns = NUM_SPECIES;
data->dx = ctL/(MX-1);
data->dy = ctL/(MY-1);
data->thispe = thispe;
data->npes = npes;
data->nsmxsub = MXSUB * NUM_SPECIES;
data->nsmxsub2 = (MXSUB+2)*NUM_SPECIES;
data->comm = comm;
data->n_local = MXSUB*MYSUB*NUM_SPECIES;
data->A = ctA;
data->B = ctB;
data->L = ctL;
data->eps[0] = data->eps[1] = ctEps;
}
/*
* SetInitialProfiles: Set initial conditions in uv, uvp, and id.
*/
static void SetInitialProfiles(N_Vector uv, N_Vector uvp, N_Vector id,
N_Vector resid, UserData data)
{
int ixsub, jysub, mxsub, mysub, nsmxsub, ix, jy;
realtype *idxy, dx, dy, x, y, *uvxy, *uvxy1, L, npex, npey;
ixsub = data->ixsub;
jysub = data->jysub;
mxsub = data->mxsub;
mysub = data->mysub;
nsmxsub = data->nsmxsub;
npex = data->npex;
npey = data->npey;
dx = data->dx;
dy = data->dy;
L = data->L;
/* Loop over grid, load uv values and id values. */
for (jy = 0; jy < mysub; jy++) {
y = (jy + jysub*mysub) * dy;
for (ix = 0; ix < mxsub; ix++) {
x = (ix + ixsub*mxsub) * dx;
uvxy = IJ_Vptr(uv,ix,jy);
uvxy[0] = RCONST(1.0) - HALF*cos(PI*y/L);
uvxy[1] = RCONST(3.5) - RCONST(2.5)*cos(PI*x/L);
}
}
N_VConst(ONE, id);
if (jysub == 0) {
for (ix=0; ix<mxsub; ix++) {
idxy = IJ_Vptr(id,ix,0);
idxy[0] = idxy[1] = ZERO;
uvxy = IJ_Vptr(uv,ix,0);
uvxy1 = IJ_Vptr(uv,ix,1);
uvxy[0] = uvxy1[0];
uvxy[1] = uvxy1[1];
}
}
if (ixsub == npex-1) {
for (jy = 0; jy < mysub; jy++) {
idxy = IJ_Vptr(id,mxsub-1,jy);
idxy[0] = idxy[1] = ZERO;
uvxy = IJ_Vptr(uv,mxsub-1,jy);
uvxy1 = IJ_Vptr(uv,mxsub-2,jy);
uvxy[0] = uvxy1[0];
uvxy[1] = uvxy1[1];
}
}
if (ixsub == 0) {
for (jy = 0; jy < mysub; jy++) {
idxy = IJ_Vptr(id,0,jy);
idxy[0] = idxy[1] = ZERO;
uvxy = IJ_Vptr(uv,0,jy);
uvxy1 = IJ_Vptr(uv,1,jy);
uvxy[0] = uvxy1[0];
uvxy[1] = uvxy1[1];
}
}
if (jysub == npey-1) {
for (ix=0; ix<mxsub; ix++) {
idxy = IJ_Vptr(id,ix,jysub);
idxy[0] = idxy[1] = ZERO;
uvxy = IJ_Vptr(uv,ix,mysub-1);
uvxy1 = IJ_Vptr(uv,ix,mysub-2);
uvxy[0] = uvxy1[0];
uvxy[1] = uvxy1[1];
}
}
/* Derivative found by calling the residual function with uvp = 0. */
N_VConst(ZERO, uvp);
res(ZERO, uv, uvp, resid, data);
N_VScale(-ONE, resid, uvp);
}
/*
* SetInitialProfilesB: Set initial conditions in uvB, uvpB
*/
static void SetInitialProfilesB(N_Vector uv, N_Vector uvp,
N_Vector uvB, N_Vector uvpB,
N_Vector residB, UserData data)
{
int ixsub, jysub, mxsub, mysub, nsmxsub, ix, jy;
realtype dx, dy, *uvxy, *uvBxy, *uvpBxy, npex, npey;
realtype B;
ixsub = data->ixsub;
jysub = data->jysub;
mxsub = data->mxsub;
mysub = data->mxsub;
nsmxsub = data->nsmxsub;
npex = data->npex;
npey = data->npey;
dx = data->dx;
dy = data->dy;
B = data->B;
/* Loop over grid, load (lambda, mu) values. */
for (jy = 0; jy < mysub; jy++) {
for (ix = 0; ix < mxsub; ix++) {
uvBxy = IJ_Vptr(uvB, ix,jy);
uvpBxy = IJ_Vptr(uvpB,ix,jy);
uvxy = IJ_Vptr(uv,ix,jy);
uvBxy[0] = ONE;
uvBxy[1] = ZERO;
uvpBxy[0] = -TWO*uvxy[0]*uvxy[1]+(B+1);
uvpBxy[1] = -uvxy[0]*uvxy[0];
}
}
if (jysub == 0) {
for (ix=0; ix<mxsub; ix++) {
uvBxy = IJ_Vptr(uvB,ix,0);
uvpBxy = IJ_Vptr(uvpB,ix,0);
uvpBxy[0] = uvpBxy[1] = ZERO;
}
}
if (ixsub == npex-1) {
for (jy = 0; jy < mysub; jy++) {
uvBxy = IJ_Vptr(uvB,mxsub-1,jy);
uvpBxy = IJ_Vptr(uvpB,mxsub-1,jy);
uvpBxy[0] = uvpBxy[1] = ZERO;
}
}
if (ixsub == 0) {
for (jy = 0; jy < mysub; jy++) {
uvBxy = IJ_Vptr(uvB,0,jy);
uvpBxy = IJ_Vptr(uvpB,0,jy);
uvpBxy[0] = uvpBxy[1] = ZERO;
}
}
if (jysub == npey-1) {
for (ix=0; ix<mxsub; ix++) {
uvBxy = IJ_Vptr(uvB,ix,mysub-1);
uvpBxy = IJ_Vptr(uvpB,ix,mysub-1);
uvpBxy[0] = uvpBxy[1] = ZERO;
}
}
}
/*
* Print first lines of output (problem description)
* and table headerr
*/
static void PrintHeader(int SystemSize, int maxl,
long int mudq, long int mldq,
long int mukeep, long int mlkeep,
realtype rtol, realtype atol)
{
printf("\n BRUSSELATOR: chemically reacting system\n\n");
printf("Number of species ns: %d", NUM_SPECIES);
printf(" Mesh dimensions: %d x %d\n", MX, MY);
printf("Total system size: %d\n",SystemSize);
printf("Subgrid dimensions: %d x %d", MXSUB, MYSUB);
printf(" Processor array: %d x %d\n", NPEX, NPEY);
#if defined(SUNDIALS_EXTENDED_PRECISION)
printf("Tolerance parameters: rtol = %Lg atol = %Lg\n", rtol, atol);
#elif defined(SUNDIALS_DOUBLE_PRECISION)
printf("Tolerance parameters: rtol = %lg atol = %lg\n", rtol, atol);
#else
printf("Tolerance parameters: rtol = %g atol = %g\n", rtol, atol);
#endif
printf("Linear solver: IDASPGMR Max. Krylov dimension maxl: %d\n", maxl);
printf("Preconditioner: band-block-diagonal (IDABBDPRE), with parameters\n");
printf(" mudq = %d, mldq = %d, mukeep = %d, mlkeep = %d\n",
mudq, mldq, mukeep, mlkeep);
printf("-----------------------------------------------------------\n");
printf(" t bottom-left top-right");
printf(" | nst k h\n");
printf("-----------------------------------------------------------\n\n");
}
/*
* PrintOutput: Print output values at output time t = tt.
* Selected run statistics are printed. Then values of c1 and c2
* are printed for the bottom left and top right grid points only.
*/
static void PrintOutput(void *mem, N_Vector uv, realtype tt,
UserData data, MPI_Comm comm)
{
MPI_Status status;
realtype *cdata, clast[2], hused;
long int nst;
int i, kused, flag, thispe, npelast, ilast;;
thispe = data->thispe;
npelast = data->npes - 1;
cdata = NV_DATA_P(uv);
/* Send conc. at top right mesh point from PE npes-1 to PE 0. */
if (thispe == npelast) {
ilast = NUM_SPECIES*MXSUB*MYSUB - 2;
if (npelast != 0)
MPI_Send(&cdata[ilast], 2, PVEC_REAL_MPI_TYPE, 0, 0, comm);
else { clast[0] = cdata[ilast]; clast[1] = cdata[ilast+1]; }
}
/* On PE 0, receive conc. at top right from PE npes - 1.
Then print performance data and sampled solution values. */
if (thispe == 0) {
if (npelast != 0)
MPI_Recv(&clast[0], 2, PVEC_REAL_MPI_TYPE, npelast, 0, comm, &status);
flag = IDAGetLastOrder(mem, &kused);
check_flag(&flag, "IDAGetLastOrder", 1, thispe);
flag = IDAGetNumSteps(mem, &nst);
check_flag(&flag, "IDAGetNumSteps", 1, thispe);
flag = IDAGetLastStep(mem, &hused);
check_flag(&flag, "IDAGetLastStep", 1, thispe);
#if defined(SUNDIALS_EXTENDED_PRECISION)
printf("%8.2Le %12.4Le %12.4Le | %3ld %1d %12.4Le\n",
tt, cdata[0], clast[0], nst, kused, hused);
for (i=1;i<NUM_SPECIES;i++)
printf(" %12.4Le %12.4Le |\n",cdata[i],clast[i]);
#elif defined(SUNDIALS_DOUBLE_PRECISION)
printf("%8.2le %12.4le %12.4le | %3ld %1d %12.4le\n",
tt, cdata[0], clast[0], nst, kused, hused);
for (i=1;i<NUM_SPECIES;i++)
printf(" %12.4le %12.4le |\n",cdata[i],clast[i]);
#else
printf("%8.2e %12.4e %12.4e | %3ld %1d %12.4e\n",
tt, cdata[0], clast[0], nst, kused, hused);
for (i=1;i<NUM_SPECIES;i++)
printf(" %12.4e %12.4e |\n",cdata[i],clast[i]);
#endif
printf("\n");
}
}
static void PrintSol(void* mem, N_Vector uv, N_Vector uvp,
UserData data, MPI_Comm comm)
{
FILE* fout;
realtype *uvxy;
int ix, jy, mxsub, mysub, npex, npey, ixsub, jysub, nsmxsub, thispe, i, j;
char szFilename[128];
thispe = data->thispe;
sprintf(szFilename, "ysol%da.txt", thispe);
fout = fopen(szFilename, "w+");
if (fout==NULL) {
printf("PE[% 2d] is unable to write solution to disk!\n", thispe);
return;
}
npex = data->npex;
npey = data->npey;
mxsub = data->mxsub;
mysub = data->mysub;
ixsub = data->ixsub;
jysub = data->jysub;
nsmxsub = data->nsmxsub;
for (jy=0; jy<mysub; jy++) {
j = jysub*mysub+jy;
for (ix=0; ix<mxsub; ix++) {
i = ix + mxsub*ixsub;
if(MXSUB<5 && MYSUB<5)
printf("PE 2D[% 2d][% 2d] -- 1D[% 2d] subgrid[%d][%d] uv[%d][%d] uv[[%d]]\n",
ixsub, jysub, thispe, ix, jy, i, j,
(i)*NUM_SPECIES + (j)*NSMXSUB*npex);
uvxy = IJ_Vptr(uv, ix, jy);
//uvxy = (&NV_Ith_P(uv, (i)*NUM_SPECIES + (j)*NSMXSUB*npex ));
fprintf(fout, "%g\n%g\n", uvxy[0], uvxy[1]);
}
}
fclose(fout);
}
static void PrintAdjSol(N_Vector uvB, N_Vector uvpB, UserData data)
{
FILE* fout;
realtype *uvxy;
int ix, jy, mxsub, mysub, npex, npey, ixsub, jysub, nsmxsub, thispe;
char szFilename[128];
thispe = data->thispe;
sprintf(szFilename, "ysol%dadj.txt", thispe);
fout = fopen(szFilename, "w+");
if (fout==NULL) {
printf("PE[% 2d] is unable to write adj solution to disk!\n", thispe);
return;
}
npex = data->npex;
npey = data->npey;
mxsub = data->mxsub;
mysub = data->mysub;
ixsub = data->ixsub;
jysub = data->jysub;
nsmxsub = data->nsmxsub;
for (jy=0; jy<mysub; jy++) {
for (ix=0; ix<mxsub; ix++) {
uvxy = IJ_Vptr(uvB, ix, jy);
fprintf(fout, "%g\n%g\n", uvxy[0], uvxy[1]);
}
}
fclose(fout);
}
/*
* PrintFinalStats: Print final run data contained in iopt.
*/
static void PrintFinalStats(void *mem)
{
long int nst, nre, nreLS, netf, ncfn, nni, ncfl, nli, npe, nps, nge;
int flag;
flag = IDAGetNumSteps(mem, &nst);
check_flag(&flag, "IDAGetNumSteps", 1, 0);
flag = IDAGetNumResEvals(mem, &nre);
check_flag(&flag, "IDAGetNumResEvals", 1, 0);
flag = IDAGetNumErrTestFails(mem, &netf);
check_flag(&flag, "IDAGetNumErrTestFails", 1, 0);
flag = IDAGetNumNonlinSolvConvFails(mem, &ncfn);
check_flag(&flag, "IDAGetNumNonlinSolvConvFails", 1, 0);
flag = IDAGetNumNonlinSolvIters(mem, &nni);
check_flag(&flag, "IDAGetNumNonlinSolvIters", 1, 0);
flag = IDASpilsGetNumConvFails(mem, &ncfl);
check_flag(&flag, "IDASpilsGetNumConvFails", 1, 0);
flag = IDASpilsGetNumLinIters(mem, &nli);
check_flag(&flag, "IDASpilsGetNumLinIters", 1, 0);
flag = IDASpilsGetNumPrecEvals(mem, &npe);
check_flag(&flag, "IDASpilsGetNumPrecEvals", 1, 0);
flag = IDASpilsGetNumPrecSolves(mem, &nps);
check_flag(&flag, "IDASpilsGetNumPrecSolves", 1, 0);
flag = IDASpilsGetNumResEvals(mem, &nreLS);
check_flag(&flag, "IDASpilsGetNumResEvals", 1, 0);
flag = IDABBDPrecGetNumGfnEvals(mem, &nge);
check_flag(&flag, "IDABBDPrecGetNumGfnEvals", 1, 0);
printf("-----------------------------------------------------------\n");
printf("\nFinal statistics: \n\n");
printf("Number of steps = %ld\n", nst);
printf("Number of residual evaluations = %ld\n", nre+nreLS);
printf("Number of nonlinear iterations = %ld\n", nni);
printf("Number of error test failures = %ld\n", netf);
printf("Number of nonlinear conv. failures = %ld\n\n", ncfn);
printf("Number of linear iterations = %ld\n", nli);
printf("Number of linear conv. failures = %ld\n\n", ncfl);
printf("Number of preconditioner setups = %ld\n", npe);
printf("Number of preconditioner solves = %ld\n", nps);
printf("Number of local residual evals. = %ld\n", nge);
}
/*
* Check function return value...
* opt == 0 means SUNDIALS function allocates memory so check if
* returned NULL pointer
* opt == 1 means SUNDIALS function returns a flag so check if
* flag >= 0
* opt == 2 means function allocates memory so check if returned
* NULL pointer
*/
static int check_flag(void *flagvalue, char *funcname, int opt, int id)
{
int *errflag;
if (opt == 0 && flagvalue == NULL) {
/* Check if SUNDIALS function returned NULL pointer - no memory allocated */
fprintf(stderr,
"\nSUNDIALS_ERROR(%d): %s() failed - returned NULL pointer\n\n",
id, funcname);
return(1);
} else if (opt == 1) {
/* Check if flag < 0 */
errflag = (int *) flagvalue;
if (*errflag < 0) {
fprintf(stderr,
"\nSUNDIALS_ERROR(%d): %s() failed with flag = %d\n\n",
id, funcname, *errflag);
return(1);
}
} else if (opt == 2 && flagvalue == NULL) {
/* Check if function returned NULL pointer - no memory allocated */
fprintf(stderr,
"\nMEMORY_ERROR(%d): %s() failed - returned NULL pointer\n\n",
id, funcname);
return(1);
}
return(0);
}
/*
*--------------------------------------------------------------------
* FUNCTIONS CALLED BY IDA & SUPPORTING FUNCTIONS
*--------------------------------------------------------------------
*/
/*
* res: System residual function
*
* To compute the residual function F, this routine calls:
* rescomm, for needed communication, and then
* reslocal, for computation of the residuals on this processor.
*/
static int res(realtype tt,
N_Vector uv, N_Vector uvp, N_Vector rr,
void *user_data)
{
int retval;
UserData data;
long int Nlocal;
data = (UserData) user_data;
Nlocal = data->n_local;
/* Call rescomm to do inter-processor communication. */
retval = rescomm(Nlocal, tt, uv, uvp, user_data);
/* Call reslocal to calculate the local portion of residual vector. */
retval = reslocal(Nlocal, tt, uv, uvp, rr, user_data);
return(0);
}
/*
* rescomm: Communication routine in support of resweb.
* This routine performs all inter-processor communication of components
* of the uv vector needed to calculate F, namely the components at all
* interior subgrid boundaries (ghost cell data). It loads this data
* into a work array cext (the local portion of c, extended).
* The message-passing uses blocking sends, non-blocking receives,
* and receive-waiting, in routines BRecvPost, BSend, BRecvWait.
*/
static int rescomm(long int Nlocal, realtype tt,
N_Vector uv, N_Vector uvp,
void *user_data)
{
UserData data;
realtype *cdata, *gridext, buffer[2*NUM_SPECIES*MYSUB];
int thispe, ixsub, jysub, nsmxsub, nsmysub;
MPI_Comm comm;
MPI_Request request[4];
data = (UserData) user_data;
cdata = NV_DATA_P(uv);
/* Get comm, thispe, subgrid indices, data sizes, extended array cext. */
comm = data->comm;
thispe = data->thispe;
ixsub = data->ixsub;
jysub = data->jysub;
gridext = data->gridext;
nsmxsub = data->nsmxsub;
nsmysub = (data->ns)*(data->mysub);
/* Start receiving boundary data from neighboring PEs. */
BRecvPost(comm, request, thispe, ixsub, jysub, nsmxsub, nsmysub,
gridext, buffer);
/* Send data from boundary of local grid to neighboring PEs. */
BSend(comm, thispe, ixsub, jysub, nsmxsub, nsmysub, cdata);
/* Finish receiving boundary data from neighboring PEs. */
BRecvWait(request, ixsub, jysub, nsmxsub, gridext, buffer);
return(0);
}
/*
* BRecvPost: Start receiving boundary data from neighboring PEs.
* (1) buffer should be able to hold 2*NUM_SPECIES*MYSUB realtype entries,
* should be passed to both the BRecvPost and BRecvWait functions, and
* should not be manipulated between the two calls.
* (2) request should have 4 entries, and is also passed in both calls.
*/
static void BRecvPost(MPI_Comm comm, MPI_Request request[], int my_pe,
int ixsub, int jysub,
int dsizex, int dsizey,
realtype cext[], realtype buffer[])
{
int offsetce;
/* Have bufleft and bufright use the same buffer. */
realtype *bufleft = buffer, *bufright = buffer+NUM_SPECIES*MYSUB;
/* If jysub > 0, receive data for bottom x-line of cext. */
if (jysub != 0)
MPI_Irecv(&cext[NUM_SPECIES], dsizex, PVEC_REAL_MPI_TYPE,
my_pe-NPEX, 0, comm, &request[0]);
/* If jysub < NPEY-1, receive data for top x-line of cext. */
if (jysub != NPEY-1) {
offsetce = NUM_SPECIES*(1 + (MYSUB+1)*(MXSUB+2));
MPI_Irecv(&cext[offsetce], dsizex, PVEC_REAL_MPI_TYPE,
my_pe+NPEX, 0, comm, &request[1]);
}
/* If ixsub > 0, receive data for left y-line of cext (via bufleft). */
if (ixsub != 0) {
MPI_Irecv(&bufleft[0], dsizey, PVEC_REAL_MPI_TYPE,
my_pe-1, 0, comm, &request[2]);
}
/* If ixsub < NPEX-1, receive data for right y-line of cext (via bufright). */
if (ixsub != NPEX-1) {
MPI_Irecv(&bufright[0], dsizey, PVEC_REAL_MPI_TYPE,
my_pe+1, 0, comm, &request[3]);
}
}
/*
* BRecvWait: Finish receiving boundary data from neighboring PEs.
* (1) buffer should be able to hold 2*NUM_SPECIES*MYSUB realtype entries,
* should be passed to both the BRecvPost and BRecvWait functions, and
* should not be manipulated between the two calls.
* (2) request should have 4 entries, and is also passed in both calls.
*/
static void BRecvWait(MPI_Request request[], int ixsub, int jysub,
int dsizex, realtype cext[], realtype buffer[])
{
int i;
int ly, dsizex2, offsetce, offsetbuf;
realtype *bufleft = buffer, *bufright = buffer+NUM_SPECIES*MYSUB;
MPI_Status status;
dsizex2 = dsizex + 2*NUM_SPECIES;
/* If jysub > 0, receive data for bottom x-line of cext. */
if (jysub != 0)
MPI_Wait(&request[0],&status);
/* If jysub < NPEY-1, receive data for top x-line of cext. */
if (jysub != NPEY-1)
MPI_Wait(&request[1],&status);
/* If ixsub > 0, receive data for left y-line of cext (via bufleft). */
if (ixsub != 0) {
MPI_Wait(&request[2],&status);
/* Copy the buffer to cext */
for (ly = 0; ly < MYSUB; ly++) {
offsetbuf = ly*NUM_SPECIES;
offsetce = (ly+1)*dsizex2;
for (i = 0; i < NUM_SPECIES; i++)
cext[offsetce+i] = bufleft[offsetbuf+i];
}
}
/* If ixsub < NPEX-1, receive data for right y-line of cext (via bufright). */
if (ixsub != NPEX-1) {
MPI_Wait(&request[3],&status);
/* Copy the buffer to cext */
for (ly = 0; ly < MYSUB; ly++) {
offsetbuf = ly*NUM_SPECIES;
offsetce = (ly+2)*dsizex2 - NUM_SPECIES;
for (i = 0; i < NUM_SPECIES; i++)
cext[offsetce+i] = bufright[offsetbuf+i];
}
}
}
/*
* BSend: Send boundary data to neighboring PEs.
* This routine sends components of uv from internal subgrid boundaries
* to the appropriate neighbor PEs.
*/
static void BSend(MPI_Comm comm, int my_pe, int ixsub, int jysub,
int dsizex, int dsizey, realtype cdata[])
{
int i;
int ly, offsetc, offsetbuf;
realtype bufleft[NUM_SPECIES*MYSUB], bufright[NUM_SPECIES*MYSUB];
/* If jysub > 0, send data from bottom x-line of uv. */
if (jysub != 0)
MPI_Send(&cdata[0], dsizex, PVEC_REAL_MPI_TYPE, my_pe-NPEX, 0, comm);
/* If jysub < NPEY-1, send data from top x-line of uv. */
if (jysub != NPEY-1) {
offsetc = (MYSUB-1)*dsizex;
MPI_Send(&cdata[offsetc], dsizex, PVEC_REAL_MPI_TYPE, my_pe+NPEX, 0, comm);
}
/* If ixsub > 0, send data from left y-line of uv (via bufleft). */
if (ixsub != 0) {
for (ly = 0; ly < MYSUB; ly++) {
offsetbuf = ly*NUM_SPECIES;
offsetc = ly*dsizex;
for (i = 0; i < NUM_SPECIES; i++)
bufleft[offsetbuf+i] = cdata[offsetc+i];
}
MPI_Send(&bufleft[0], dsizey, PVEC_REAL_MPI_TYPE, my_pe-1, 0, comm);
}
/* If ixsub < NPEX-1, send data from right y-line of uv (via bufright). */
if (ixsub != NPEX-1) {
for (ly = 0; ly < MYSUB; ly++) {
offsetbuf = ly*NUM_SPECIES;
offsetc = offsetbuf*MXSUB + (MXSUB-1)*NUM_SPECIES;
for (i = 0; i < NUM_SPECIES; i++)
bufright[offsetbuf+i] = cdata[offsetc+i];
}
MPI_Send(&bufright[0], dsizey, PVEC_REAL_MPI_TYPE, my_pe+1, 0, comm);
}
}
/* Define lines are for ease of readability in the following functions. */
#define mxsub (data->mxsub)
#define mysub (data->mysub)
#define npex (data->npex)
#define npey (data->npey)
#define ixsub (data->ixsub)
#define jysub (data->jysub)
#define nsmxsub (data->nsmxsub)
#define nsmxsub2 (data->nsmxsub2)
#define dx (data->dx)
#define dy (data->dy)
#define cox (data->cox)
#define coy (data->coy)
#define gridext (data->gridext)
#define eps (data->eps)
#define ns (data->ns)
/*
* reslocal: Compute res = F(t,uv,uvp).
* This routine assumes that all inter-processor communication of data
* needed to calculate F has already been done. Components at interior
* subgrid boundaries are assumed to be in the work array cext.
* The local portion of the uv vector is first copied into cext.
* The exterior Neumann boundary conditions are explicitly handled here
* by copying data from the first interior mesh line to the ghost cell
* locations in cext. Then the reaction and diffusion terms are
* evaluated in terms of the cext array, and the residuals are formed.
* The reaction terms are saved separately in the vector data->rates
* for use by the preconditioner setup routine.
*/
static int reslocal(long int Nlocal, realtype tt,
N_Vector uv, N_Vector uvp, N_Vector rr,
void *user_data)
{
realtype *uvdata, *uvpxy, *resxy, xx, yy, dcyli, dcyui, dcxli, dcxui, dx2, dy2;
realtype ixend, ixstart, jystart, jyend;
int ix, jy, is, i, locc, ylocce, locce;
realtype rates[2];
UserData data;
data = (UserData) user_data;
/* Get data pointers, subgrid data, array sizes, work array cext. */
uvdata = NV_DATA_P(uv);
dx2 = dx * dx;
dy2 = dy * dy;
/* Copy local segment of uv vector into the working extended array gridext. */
locc = 0;
locce = nsmxsub2 + NUM_SPECIES;
for (jy = 0; jy < mysub; jy++) {
for (i = 0; i < nsmxsub; i++) gridext[locce+i] = uvdata[locc+i];
locc = locc + nsmxsub;
locce = locce + nsmxsub2;
}
/* To facilitate homogeneous Neumann boundary conditions, when this is
a boundary PE, copy data from the first interior mesh line of uv to gridext. */
/* If jysub = 0, copy x-line 2 of uv to gridext. */
if (jysub == 0)
{ for (i = 0; i < nsmxsub; i++) gridext[NUM_SPECIES+i] = uvdata[nsmxsub+i]; }
/* If jysub = npey-1, copy x-line mysub-1 of uv to gridext. */
if (jysub == npey-1) {
locc = (mysub-2)*nsmxsub;
locce = (mysub+1)*nsmxsub2 + NUM_SPECIES;
for (i = 0; i < nsmxsub; i++) gridext[locce+i] = uvdata[locc+i];
}
/* If ixsub = 0, copy y-line 2 of uv to gridext. */
if (ixsub == 0) {
for (jy = 0; jy < mysub; jy++) {
locc = jy*nsmxsub + NUM_SPECIES;
locce = (jy+1)*nsmxsub2;
for (i = 0; i < NUM_SPECIES; i++) gridext[locce+i] = uvdata[locc+i];
}
}
/* If ixsub = npex-1, copy y-line mxsub-1 of uv to gridext. */
if (ixsub == npex-1) {
for (jy = 0; jy < mysub; jy++) {
locc = (jy+1)*nsmxsub - 2*NUM_SPECIES;
locce = (jy+2)*nsmxsub2 - NUM_SPECIES;
for (i = 0; i < NUM_SPECIES; i++) gridext[locce+i] = uvdata[locc+i];
}
}
/* Loop over all grid points, setting local array rates to right-hand sides.
Then set rr values appropriately (ODE in the interior and DAE on the boundary)*/
ixend = ixstart = jystart = jyend = 0;
if (jysub==0) jystart = 1;
if (jysub==npey-1) jyend = 1;
if (ixsub==0) ixstart = 1;
if (ixsub==npex-1) ixend = 1;
for (jy = jystart; jy < mysub-jyend; jy++) {
ylocce = (jy+1)*nsmxsub2;
yy = (jy+jysub*mysub)*dy;
for (ix = ixstart; ix < mxsub-ixend; ix++) {
locce = ylocce + (ix+1)*NUM_SPECIES;
xx = (ix + ixsub*mxsub)*dx;
ReactRates(xx, yy, &(gridext[locce]), rates, data);
resxy = IJ_Vptr(rr,ix,jy);
uvpxy = IJ_Vptr(uvp,ix,jy);
for (is = 0; is < NUM_SPECIES; is++) {
dcyli = gridext[locce+is] - gridext[locce+is-nsmxsub2];
dcyui = gridext[locce+is+nsmxsub2] - gridext[locce+is];
dcxli = gridext[locce+is] - gridext[locce+is-NUM_SPECIES];
dcxui = gridext[locce+is+NUM_SPECIES] - gridext[locce+is];
resxy[is] = uvpxy[is]-eps[is]*((dcxui-dcxli)/dx2+(dcyui-dcyli)/dy2)-rates[is];
}
}
}
if (jysub==0) {
for (ix=0; ix<mxsub; ix++) {
locce = nsmxsub2 + NUM_SPECIES * (ix+1);
resxy = IJ_Vptr(rr,ix,0);
for (is=0; is<NUM_SPECIES; is++)
resxy[is] = gridext[locce+is+nsmxsub2] - gridext[locce+is];
}
}
if (ixsub==npex-1) {
for(jy=0; jy<mysub; jy++) {
locce = (jy+1)*nsmxsub2 + nsmxsub2-NUM_SPECIES;
resxy = IJ_Vptr(rr,mxsub-1,jy);
for (is=0; is<NUM_SPECIES; is++)
resxy[is] = gridext[locce+is-NUM_SPECIES] - gridext[locce+is];
}
}
if (ixsub==0) {
for (jy=0; jy<mysub; jy++) {
locce = (jy+1)*nsmxsub2 + NUM_SPECIES;
resxy = IJ_Vptr(rr,0,jy);
for (is=0; is<NUM_SPECIES; is++)
resxy[is] = gridext[locce+is-NUM_SPECIES] - gridext[locce+is];
}
}
if (jysub==npey-1) {
for(ix=0; ix<mxsub; ix++) {
locce = nsmxsub2*mysub + (ix+1)*NUM_SPECIES;
resxy = IJ_Vptr(rr,ix, mysub-1);
for (is=0; is<NUM_SPECIES; is++)
resxy[is] = gridext[locce+is-nsmxsub2] - gridext[locce+is];
}
}
return(0);
}
/*
* ReactRates: Evaluate reaction rates at a given spatial point.
* At a given (x,y), evaluate the array of ns reaction terms R.
*/
static void ReactRates(realtype xx, realtype yy, realtype *uvval, realtype *rates,
UserData data)
{
realtype A, B;
A = data->A; B = data->B;
rates[0] = uvval[0]*uvval[0]*uvval[1];
rates[1] = - rates[0];
rates[0] += A-(B+1)*uvval[0];
rates[1] += B*uvval[0];
}
static int resB(realtype tt, N_Vector yy, N_Vector yp,
N_Vector yyB, N_Vector ypB, N_Vector rrB,
void *user_dataB)
{
UserData data;
int retval;
long int Nlocal;
data = (UserData) user_dataB;
Nlocal = data->n_local;
/* Call rescomm to do inter-processor communication. */
retval = rescomm(Nlocal, tt, yyB, ypB, data);
/* Call reslocal to calculate the local portion of residual vector. */
retval = resBlocal(Nlocal, tt, yy, yp, yyB, ypB, rrB, user_dataB);
return(0);
}
static int resBlocal(long int Nlocal, realtype tt,
N_Vector uv, N_Vector uvp,
N_Vector uvB, N_Vector uvpB, N_Vector rrB,
void *user_dataB)
{
realtype *uvBdata, *uvBxy, *uvpBxy, *uvxy, *rrBxy;
realtype dx2, dy2, xx, yy;
realtype dcxli, dcxui, dcyli, dcyui;
int locc, locce, ylocce;
int ix, jy, i, ixstart, ixend, jystart, jyend, is;
UserData data;
realtype A, B;
data = (UserData) user_dataB;
A = data->A; B = data->B;
/* Get data pointers, subgrid data, array sizes, work array cext. */
uvBdata = NV_DATA_P(uvB);
dx2 = dx * dx;
dy2 = dy * dy;
/* Copy local segment of uv vector into the working extended array gridext. */
locc = 0;
locce = nsmxsub2 + NUM_SPECIES;
for (jy = 0; jy < mysub; jy++) {
for (i = 0; i < nsmxsub; i++) gridext[locce+i] = uvBdata[locc+i];
locc = locc + nsmxsub;
locce = locce + nsmxsub2;
}
/* If jysub = 0, copy x-line 2 of uv to gridext. */
if (jysub == 0)
{ for (i = 0; i < nsmxsub; i++) gridext[NUM_SPECIES+i] = uvBdata[nsmxsub+i]; }
/* If jysub = npey-1, copy x-line mysub-1 of uv to gridext. */
if (jysub == npey-1) {
locc = (mysub-2)*nsmxsub;
locce = (mysub+1)*nsmxsub2 + NUM_SPECIES;
for (i = 0; i < nsmxsub; i++) gridext[locce+i] = uvBdata[locc+i];
}
/* If ixsub = 0, copy y-line 2 of uv to gridext. */
if (ixsub == 0) {
for (jy = 0; jy < mysub; jy++) {
locc = jy*nsmxsub + NUM_SPECIES;
locce = (jy+1)*nsmxsub2;
for (i = 0; i < NUM_SPECIES; i++) gridext[locce+i] = uvBdata[locc+i];
}
}
/* If ixsub = npex-1, copy y-line mxsub-1 of uv to gridext. */
if (ixsub == npex-1) {
for (jy = 0; jy < mysub; jy++) {
locc = (jy+1)*nsmxsub - 2*NUM_SPECIES;
locce = (jy+2)*nsmxsub2 - NUM_SPECIES;
for (i = 0; i < NUM_SPECIES; i++) gridext[locce+i] = uvBdata[locc+i];
}
}
/* Loop over all grid points, setting local array rates to right-hand sides.
Then set rr values appropriately (ODE in the interior and DAE on the boundary)*/
ixend = ixstart = jystart = jyend = 0;
if (jysub==0) jystart = 1;
if (jysub==npey-1) jyend = 1;
if (ixsub==0) ixstart = 1;
if (ixsub==npex-1) ixend = 1;
for (jy = jystart; jy < mysub-jyend; jy++) {
ylocce = (jy+1)*nsmxsub2;
yy = (jy+jysub*mysub)*dy;
for (ix = ixstart; ix < mxsub-ixend; ix++) {
locce = ylocce + (ix+1)*NUM_SPECIES;
xx = (ix + ixsub*mxsub)*dx;
uvxy = IJ_Vptr(uv ,ix,jy);
uvBxy = IJ_Vptr(uvB ,ix,jy);
uvpBxy= IJ_Vptr(uvpB,ix,jy);
rrBxy = IJ_Vptr(rrB ,ix,jy);
for (is = 0; is < NUM_SPECIES; is++) {
dcyli = gridext[locce+is] - gridext[locce+is-nsmxsub2];
dcyui = gridext[locce+is+nsmxsub2] - gridext[locce+is];
dcxli = gridext[locce+is] - gridext[locce+is-NUM_SPECIES];
dcxui = gridext[locce+is+NUM_SPECIES] - gridext[locce+is];
rrBxy[is] = uvpBxy[is] + eps[is]*( (dcxui-dcxli)/dx2 + (dcyui-dcyli)/dy2 );
}
//now add rates
rrBxy[0] += (uvBxy[0]-uvBxy[1])*(2*uvxy[0]*uvxy[1] - B) - uvBxy[0];
rrBxy[1] += uvxy[0]*uvxy[0]*(uvBxy[0]-uvBxy[1]);
}
}
if (jysub==0) {
for (ix=0; ix<mxsub; ix++) {
locce = nsmxsub2 + NUM_SPECIES * (ix+1);
rrBxy = IJ_Vptr(rrB,ix,0);
for (is=0; is<NUM_SPECIES; is++)
rrBxy[is] = gridext[locce+is+nsmxsub2] - gridext[locce+is];
}
}
if (ixsub==npex-1) {
for(jy=0; jy<mysub; jy++) {
locce = (jy+1)*nsmxsub2 + nsmxsub2-NUM_SPECIES;
rrBxy = IJ_Vptr(rrB,mxsub-1,jy);
for (is=0; is<NUM_SPECIES; is++)
rrBxy[is] = gridext[locce+is-NUM_SPECIES] - gridext[locce+is];
}
}
if (ixsub==0) {
for (jy=0; jy<mysub; jy++) {
locce = (jy+1)*nsmxsub2 + NUM_SPECIES;
rrBxy = IJ_Vptr(rrB,0,jy);
for (is=0; is<NUM_SPECIES; is++)
rrBxy[is] = gridext[locce+is-NUM_SPECIES] - gridext[locce+is];
}
}
if (jysub==npey-1) {
for(ix=0; ix<mxsub; ix++) {
locce = nsmxsub2*mysub + (ix+1)*NUM_SPECIES;
rrBxy = IJ_Vptr(rrB,ix, mysub-1);
for (is=0; is<NUM_SPECIES; is++)
rrBxy[is] = gridext[locce+is-nsmxsub2] - gridext[locce+is];
}
}
return(0);
}
|