| 1 |
C $Header: /u/gcmpack/MITgcm/model/src/cg3d.F,v 1.19 2007/09/04 14:54:58 jmc Exp $ |
| 2 |
C $Name: $ |
| 3 |
|
| 4 |
#include "CPP_OPTIONS.h" |
| 5 |
|
| 6 |
CBOP |
| 7 |
C !ROUTINE: CG3D |
| 8 |
C !INTERFACE: |
| 9 |
SUBROUTINE CG3D( |
| 10 |
I cg3d_b, |
| 11 |
U cg3d_x, |
| 12 |
O firstResidual, |
| 13 |
O lastResidual, |
| 14 |
U numIters, |
| 15 |
I myThid ) |
| 16 |
C !DESCRIPTION: \bv |
| 17 |
C *==========================================================* |
| 18 |
C | SUBROUTINE CG3D |
| 19 |
C | o Three-dimensional grid problem conjugate-gradient |
| 20 |
C | inverter (with preconditioner). |
| 21 |
C *==========================================================* |
| 22 |
C | Con. grad is an iterative procedure for solving Ax = b. |
| 23 |
C | It requires the A be symmetric. |
| 24 |
C | This implementation assumes A is a seven-diagonal |
| 25 |
C | matrix of the form that arises in the discrete |
| 26 |
C | representation of the del^2 operator in a |
| 27 |
C | three-dimensional space. |
| 28 |
C | Notes: |
| 29 |
C | ====== |
| 30 |
C | This implementation can support shared-memory |
| 31 |
C | multi-threaded execution. In order to do this COMMON |
| 32 |
C | blocks are used for many of the arrays - even ones that |
| 33 |
C | are only used for intermedaite results. This design is |
| 34 |
C | OK if you want to all the threads to collaborate on |
| 35 |
C | solving the same problem. On the other hand if you want |
| 36 |
C | the threads to solve several different problems |
| 37 |
C | concurrently this implementation will not work. |
| 38 |
C *==========================================================* |
| 39 |
C \ev |
| 40 |
|
| 41 |
C !USES: |
| 42 |
IMPLICIT NONE |
| 43 |
C === Global data === |
| 44 |
#include "SIZE.h" |
| 45 |
#include "EEPARAMS.h" |
| 46 |
#include "PARAMS.h" |
| 47 |
#include "GRID.h" |
| 48 |
#include "CG3D.h" |
| 49 |
|
| 50 |
C !INPUT/OUTPUT PARAMETERS: |
| 51 |
C === Routine arguments === |
| 52 |
C myThid - Thread on which I am working. |
| 53 |
C cg3d_b - The source term or "right hand side" |
| 54 |
C cg3d_x - The solution |
| 55 |
C firstResidual - the initial residual before any iterations |
| 56 |
C lastResidual - the actual residual reached |
| 57 |
C numIters - Entry: the maximum number of iterations allowed |
| 58 |
C Exit: the actual number of iterations used |
| 59 |
_RL cg3d_b(1-OLx:sNx+OLx,1-OLy:sNy+OLy,Nr,nSx,nSy) |
| 60 |
_RL cg3d_x(1-OLx:sNx+OLx,1-OLy:sNy+OLy,Nr,nSx,nSy) |
| 61 |
_RL firstResidual |
| 62 |
_RL lastResidual |
| 63 |
INTEGER numIters |
| 64 |
INTEGER myThid |
| 65 |
|
| 66 |
|
| 67 |
#ifdef ALLOW_NONHYDROSTATIC |
| 68 |
|
| 69 |
C !LOCAL VARIABLES: |
| 70 |
C === Local variables ==== |
| 71 |
C actualIts - Number of iterations taken |
| 72 |
C actualResidual - residual |
| 73 |
C bi - Block index in X and Y. |
| 74 |
C bj |
| 75 |
C eta_qrN - Used in computing search directions |
| 76 |
C eta_qrNM1 suffix N and NM1 denote current and |
| 77 |
C cgBeta previous iterations respectively. |
| 78 |
C alpha |
| 79 |
C sumRHS - Sum of right-hand-side. Sometimes this is a |
| 80 |
C useful debuggin/trouble shooting diagnostic. |
| 81 |
C For neumann problems sumRHS needs to be ~0. |
| 82 |
C or they converge at a non-zero residual. |
| 83 |
C err - Measure of residual of Ax - b, usually the norm. |
| 84 |
C I, J, K, N - Loop counters ( N counts CG iterations ) |
| 85 |
INTEGER actualIts |
| 86 |
_RL actualResidual |
| 87 |
INTEGER bi, bj |
| 88 |
INTEGER I, J, K, it3d |
| 89 |
INTEGER Km1, Kp1 |
| 90 |
_RL maskM1, maskP1 |
| 91 |
_RL err, errTile(nSx,nSy) |
| 92 |
_RL eta_qrN,eta_qrNtile(nSx,nSy) |
| 93 |
_RL eta_qrNM1 |
| 94 |
_RL cgBeta |
| 95 |
_RL alpha , alphaTile(nSx,nSy) |
| 96 |
_RL sumRHS, sumRHStile(nSx,nSy) |
| 97 |
_RL rhsMax |
| 98 |
_RL rhsNorm |
| 99 |
CEOP |
| 100 |
|
| 101 |
|
| 102 |
C-- Initialise inverter |
| 103 |
eta_qrNM1 = 1. D0 |
| 104 |
|
| 105 |
C-- Normalise RHS |
| 106 |
rhsMax = 0. _d 0 |
| 107 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 108 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 109 |
DO K=1,Nr |
| 110 |
DO J=1,sNy |
| 111 |
DO I=1,sNx |
| 112 |
cg3d_b(I,J,K,bi,bj) = cg3d_b(I,J,K,bi,bj)*cg3dNorm |
| 113 |
& * maskC(I,J,K,bi,bj) |
| 114 |
rhsMax = MAX(ABS(cg3d_b(I,J,K,bi,bj)),rhsMax) |
| 115 |
ENDDO |
| 116 |
ENDDO |
| 117 |
ENDDO |
| 118 |
ENDDO |
| 119 |
ENDDO |
| 120 |
_GLOBAL_MAX_RL( rhsMax, myThid ) |
| 121 |
rhsNorm = 1. _d 0 |
| 122 |
IF ( rhsMax .NE. 0. ) rhsNorm = 1. _d 0 / rhsMax |
| 123 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 124 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 125 |
DO K=1,Nr |
| 126 |
DO J=1,sNy |
| 127 |
DO I=1,sNx |
| 128 |
cg3d_b(I,J,K,bi,bj) = cg3d_b(I,J,K,bi,bj)*rhsNorm |
| 129 |
cg3d_x(I,J,K,bi,bj) = cg3d_x(I,J,K,bi,bj)*rhsNorm |
| 130 |
ENDDO |
| 131 |
ENDDO |
| 132 |
ENDDO |
| 133 |
ENDDO |
| 134 |
ENDDO |
| 135 |
|
| 136 |
C-- Update overlaps |
| 137 |
c _EXCH_XYZ_RL( cg3d_b, myThid ) |
| 138 |
_EXCH_XYZ_RL( cg3d_x, myThid ) |
| 139 |
|
| 140 |
C-- Initial residual calculation (with free-Surface term) |
| 141 |
err = 0. _d 0 |
| 142 |
sumRHS = 0. _d 0 |
| 143 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 144 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 145 |
errTile(bi,bj) = 0. _d 0 |
| 146 |
sumRHStile(bi,bj) = 0. _d 0 |
| 147 |
DO K=1,Nr |
| 148 |
Km1 = MAX(K-1, 1 ) |
| 149 |
Kp1 = MIN(K+1, Nr) |
| 150 |
maskM1 = 1. _d 0 |
| 151 |
maskP1 = 1. _d 0 |
| 152 |
IF ( K .EQ. 1 ) maskM1 = 0. _d 0 |
| 153 |
IF ( K .EQ. Nr) maskP1 = 0. _d 0 |
| 154 |
|
| 155 |
DO J=1,sNy |
| 156 |
DO I=1,sNx |
| 157 |
cg3d_r(I,J,K,bi,bj) = cg3d_b(I,J,K,bi,bj) -( 0. |
| 158 |
& +aW3d(I ,J ,K ,bi,bj)*cg3d_x(I-1,J ,K ,bi,bj) |
| 159 |
& +aW3d(I+1,J ,K ,bi,bj)*cg3d_x(I+1,J ,K ,bi,bj) |
| 160 |
& +aS3d(I ,J ,K ,bi,bj)*cg3d_x(I ,J-1,K ,bi,bj) |
| 161 |
& +aS3d(I ,J+1,K ,bi,bj)*cg3d_x(I ,J+1,K ,bi,bj) |
| 162 |
& +aV3d(I ,J ,K ,bi,bj)*cg3d_x(I ,J ,Km1,bi,bj)*maskM1 |
| 163 |
& +aV3d(I ,J ,Kp1,bi,bj)*cg3d_x(I ,J ,Kp1,bi,bj)*maskP1 |
| 164 |
& +aC3d(I ,J ,K ,bi,bj)*cg3d_x(I ,J ,K ,bi,bj) |
| 165 |
& ) |
| 166 |
errTile(bi,bj) = errTile(bi,bj) |
| 167 |
& +cg3d_r(I,J,K,bi,bj)*cg3d_r(I,J,K,bi,bj) |
| 168 |
sumRHStile(bi,bj) = sumRHStile(bi,bj) |
| 169 |
& +cg3d_b(I,J,K,bi,bj) |
| 170 |
ENDDO |
| 171 |
ENDDO |
| 172 |
DO J=1-1,sNy+1 |
| 173 |
DO I=1-1,sNx+1 |
| 174 |
cg3d_s(I,J,K,bi,bj) = 0. |
| 175 |
ENDDO |
| 176 |
ENDDO |
| 177 |
ENDDO |
| 178 |
c err = err + errTile(bi,bj) |
| 179 |
c sumRHS = sumRHS + sumRHStile(bi,bj) |
| 180 |
ENDDO |
| 181 |
ENDDO |
| 182 |
CALL EXCH_S3D_RL( cg3d_r, Nr, myThid ) |
| 183 |
c CALL EXCH_S3D_RL( cg3d_s, Nr, myThid ) |
| 184 |
c _GLOBAL_SUM_RL( sumRHS, myThid ) |
| 185 |
c _GLOBAL_SUM_RL( err , myThid ) |
| 186 |
CALL GLOBAL_SUM_TILE_RL( sumRHStile, sumRHS, myThid ) |
| 187 |
CALL GLOBAL_SUM_TILE_RL( errTile, err, myThid ) |
| 188 |
|
| 189 |
IF ( debugLevel .GE. debLevZero ) THEN |
| 190 |
_BEGIN_MASTER( myThid ) |
| 191 |
write(standardmessageunit,'(A,1P2E22.14)') |
| 192 |
& ' cg3d: Sum(rhs),rhsMax = ',sumRHS,rhsMax |
| 193 |
_END_MASTER( myThid ) |
| 194 |
ENDIF |
| 195 |
|
| 196 |
actualIts = 0 |
| 197 |
actualResidual = SQRT(err) |
| 198 |
C _BARRIER |
| 199 |
c _BEGIN_MASTER( myThid ) |
| 200 |
c WRITE(*,'(A,I6,1PE30.14)') ' CG3D iters, err = ', |
| 201 |
c & actualIts, actualResidual |
| 202 |
c _END_MASTER( myThid ) |
| 203 |
firstResidual=actualResidual |
| 204 |
|
| 205 |
C >>>>>>>>>>>>>>> BEGIN SOLVER <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< |
| 206 |
DO 10 it3d=1, numIters |
| 207 |
|
| 208 |
CcnhDebugStarts |
| 209 |
c IF ( mod(it3d-1,10).EQ.0) |
| 210 |
c & WRITE(*,*) ' CG3D: Iteration ',it3d-1, |
| 211 |
c & ' residual = ',actualResidual |
| 212 |
CcnhDebugEnds |
| 213 |
IF ( actualResidual .LT. cg3dTargetResidual ) GOTO 11 |
| 214 |
C-- Solve preconditioning equation and update |
| 215 |
C-- conjugate direction vector "s". |
| 216 |
C Note. On the next to loops over all tiles the inner loop ranges |
| 217 |
C in sNx and sNy are expanded by 1 to avoid a communication |
| 218 |
C step. However this entails a bit of gynamastics because we only |
| 219 |
C want eta_qrN for the interior points. |
| 220 |
eta_qrN = 0. _d 0 |
| 221 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 222 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 223 |
eta_qrNtile(bi,bj) = 0. _d 0 |
| 224 |
DO K=1,1 |
| 225 |
DO J=1-1,sNy+1 |
| 226 |
DO I=1-1,sNx+1 |
| 227 |
cg3d_q(I,J,K,bi,bj) = |
| 228 |
& zMC(I ,J ,K,bi,bj)*cg3d_r(I ,J ,K,bi,bj) |
| 229 |
ENDDO |
| 230 |
ENDDO |
| 231 |
ENDDO |
| 232 |
DO K=2,Nr |
| 233 |
DO J=1-1,sNy+1 |
| 234 |
DO I=1-1,sNx+1 |
| 235 |
cg3d_q(I,J,K,bi,bj) = |
| 236 |
& zMC(I,J,K,bi,bj)*(cg3d_r(I,J,K ,bi,bj) |
| 237 |
& -zML(I,J,K,bi,bj)*cg3d_q(I,J,K-1,bi,bj)) |
| 238 |
ENDDO |
| 239 |
ENDDO |
| 240 |
ENDDO |
| 241 |
DO K=Nr,Nr |
| 242 |
caja IF (Nr .GT. 1) THEN |
| 243 |
caja DO J=1-1,sNy+1 |
| 244 |
caja DO I=1-1,sNx+1 |
| 245 |
caja cg3d_q(I,J,K,bi,bj) = |
| 246 |
caja & zMC(i,j,k,bi,bj)*(cg3d_r(i,j,k ,bi,bj) |
| 247 |
caja & -zML(i,j,k,bi,bj)*cg3d_q(i,j,k-1,bi,bj)) |
| 248 |
caja ENDDO |
| 249 |
caja ENDDO |
| 250 |
caja ENDIF |
| 251 |
DO J=1,sNy |
| 252 |
DO I=1,sNx |
| 253 |
eta_qrNtile(bi,bj) = eta_qrNtile(bi,bj) |
| 254 |
& +cg3d_q(I,J,K,bi,bj)*cg3d_r(I,J,K,bi,bj) |
| 255 |
ENDDO |
| 256 |
ENDDO |
| 257 |
ENDDO |
| 258 |
DO K=Nr-1,1,-1 |
| 259 |
DO J=1-1,sNy+1 |
| 260 |
DO I=1-1,sNx+1 |
| 261 |
cg3d_q(I,J,K,bi,bj) = |
| 262 |
& cg3d_q(I,J,K,bi,bj) |
| 263 |
& -zMU(I,J,K,bi,bj)*cg3d_q(I,J,K+1,bi,bj) |
| 264 |
ENDDO |
| 265 |
ENDDO |
| 266 |
DO J=1,sNy |
| 267 |
DO I=1,sNx |
| 268 |
eta_qrNtile(bi,bj) = eta_qrNtile(bi,bj) |
| 269 |
& +cg3d_q(I,J,K,bi,bj)*cg3d_r(I,J,K,bi,bj) |
| 270 |
ENDDO |
| 271 |
ENDDO |
| 272 |
ENDDO |
| 273 |
c eta_qrN = eta_qrN + eta_qrNtile(bi,bj) |
| 274 |
ENDDO |
| 275 |
ENDDO |
| 276 |
caja |
| 277 |
caja eta_qrN=0. |
| 278 |
caja DO bj=myByLo(myThid),myByHi(myThid) |
| 279 |
caja DO bi=myBxLo(myThid),myBxHi(myThid) |
| 280 |
caja DO K=1,Nr |
| 281 |
caja DO J=1,sNy |
| 282 |
caja DO I=1,sNx |
| 283 |
caja eta_qrN = eta_qrN |
| 284 |
caja & +cg3d_q(I,J,K,bi,bj)*cg3d_r(I,J,K,bi,bj) |
| 285 |
caja ENDDO |
| 286 |
caja ENDDO |
| 287 |
caja ENDDO |
| 288 |
caja ENDDO |
| 289 |
caja ENDDO |
| 290 |
caja |
| 291 |
|
| 292 |
c _GLOBAL_SUM_RL(eta_qrN, myThid) |
| 293 |
CALL GLOBAL_SUM_TILE_RL( eta_qrNtile,eta_qrN,myThid ) |
| 294 |
CcnhDebugStarts |
| 295 |
C WRITE(*,*) ' CG3D: Iteration ',it3d-1,' eta_qrN = ',eta_qrN |
| 296 |
CcnhDebugEnds |
| 297 |
cgBeta = eta_qrN/eta_qrNM1 |
| 298 |
CcnhDebugStarts |
| 299 |
C WRITE(*,*) ' CG3D: Iteration ',it3d-1,' beta = ',cgBeta |
| 300 |
CcnhDebugEnds |
| 301 |
eta_qrNM1 = eta_qrN |
| 302 |
|
| 303 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 304 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 305 |
DO K=1,Nr |
| 306 |
DO J=1-1,sNy+1 |
| 307 |
DO I=1-1,sNx+1 |
| 308 |
cg3d_s(I,J,K,bi,bj) = cg3d_q(I,J,K,bi,bj) |
| 309 |
& + cgBeta*cg3d_s(I,J,K,bi,bj) |
| 310 |
ENDDO |
| 311 |
ENDDO |
| 312 |
ENDDO |
| 313 |
ENDDO |
| 314 |
ENDDO |
| 315 |
|
| 316 |
C== Evaluate laplace operator on conjugate gradient vector |
| 317 |
C== q = A.s |
| 318 |
alpha = 0. _d 0 |
| 319 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 320 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 321 |
alphaTile(bi,bj) = 0. _d 0 |
| 322 |
IF ( Nr .GT. 1 ) THEN |
| 323 |
DO K=1,1 |
| 324 |
DO J=1,sNy |
| 325 |
DO I=1,sNx |
| 326 |
cg3d_q(I,J,K,bi,bj) = |
| 327 |
& aW3d(I ,J ,K ,bi,bj)*cg3d_s(I-1,J ,K ,bi,bj) |
| 328 |
& +aW3d(I+1,J ,K ,bi,bj)*cg3d_s(I+1,J ,K ,bi,bj) |
| 329 |
& +aS3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J-1,K ,bi,bj) |
| 330 |
& +aS3d(I ,J+1,K ,bi,bj)*cg3d_s(I ,J+1,K ,bi,bj) |
| 331 |
& +aV3d(I ,J ,K+1,bi,bj)*cg3d_s(I ,J ,K+1,bi,bj) |
| 332 |
& +aC3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J ,K ,bi,bj) |
| 333 |
alphaTile(bi,bj) = alphaTile(bi,bj) |
| 334 |
& +cg3d_s(I,J,K,bi,bj)*cg3d_q(I,J,K,bi,bj) |
| 335 |
ENDDO |
| 336 |
ENDDO |
| 337 |
ENDDO |
| 338 |
ELSE |
| 339 |
DO K=1,1 |
| 340 |
DO J=1,sNy |
| 341 |
DO I=1,sNx |
| 342 |
cg3d_q(I,J,K,bi,bj) = |
| 343 |
& aW3d(I ,J ,K ,bi,bj)*cg3d_s(I-1,J ,K ,bi,bj) |
| 344 |
& +aW3d(I+1,J ,K ,bi,bj)*cg3d_s(I+1,J ,K ,bi,bj) |
| 345 |
& +aS3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J-1,K ,bi,bj) |
| 346 |
& +aS3d(I ,J+1,K ,bi,bj)*cg3d_s(I ,J+1,K ,bi,bj) |
| 347 |
& +aC3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J ,K ,bi,bj) |
| 348 |
alphaTile(bi,bj) = alphaTile(bi,bj) |
| 349 |
& +cg3d_s(I,J,K,bi,bj)*cg3d_q(I,J,K,bi,bj) |
| 350 |
ENDDO |
| 351 |
ENDDO |
| 352 |
ENDDO |
| 353 |
ENDIF |
| 354 |
DO K=2,Nr-1 |
| 355 |
DO J=1,sNy |
| 356 |
DO I=1,sNx |
| 357 |
cg3d_q(I,J,K,bi,bj) = |
| 358 |
& aW3d(I ,J ,K ,bi,bj)*cg3d_s(I-1,J ,K ,bi,bj) |
| 359 |
& +aW3d(I+1,J ,K ,bi,bj)*cg3d_s(I+1,J ,K ,bi,bj) |
| 360 |
& +aS3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J-1,K ,bi,bj) |
| 361 |
& +aS3d(I ,J+1,K ,bi,bj)*cg3d_s(I ,J+1,K ,bi,bj) |
| 362 |
& +aV3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J ,K-1,bi,bj) |
| 363 |
& +aV3d(I ,J ,K+1,bi,bj)*cg3d_s(I ,J ,K+1,bi,bj) |
| 364 |
& +aC3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J ,K ,bi,bj) |
| 365 |
alphaTile(bi,bj) = alphaTile(bi,bj) |
| 366 |
& +cg3d_s(I,J,K,bi,bj)*cg3d_q(I,J,K,bi,bj) |
| 367 |
ENDDO |
| 368 |
ENDDO |
| 369 |
ENDDO |
| 370 |
IF ( Nr .GT. 1 ) THEN |
| 371 |
DO K=Nr,Nr |
| 372 |
DO J=1,sNy |
| 373 |
DO I=1,sNx |
| 374 |
cg3d_q(I,J,K,bi,bj) = |
| 375 |
& aW3d(I ,J ,K ,bi,bj)*cg3d_s(I-1,J ,K ,bi,bj) |
| 376 |
& +aW3d(I+1,J ,K ,bi,bj)*cg3d_s(I+1,J ,K ,bi,bj) |
| 377 |
& +aS3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J-1,K ,bi,bj) |
| 378 |
& +aS3d(I ,J+1,K ,bi,bj)*cg3d_s(I ,J+1,K ,bi,bj) |
| 379 |
& +aV3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J ,K-1,bi,bj) |
| 380 |
& +aC3d(I ,J ,K ,bi,bj)*cg3d_s(I ,J ,K ,bi,bj) |
| 381 |
alphaTile(bi,bj) = alphaTile(bi,bj) |
| 382 |
& +cg3d_s(I,J,K,bi,bj)*cg3d_q(I,J,K,bi,bj) |
| 383 |
ENDDO |
| 384 |
ENDDO |
| 385 |
ENDDO |
| 386 |
ENDIF |
| 387 |
c alpha = alpha + alphaTile(bi,bj) |
| 388 |
ENDDO |
| 389 |
ENDDO |
| 390 |
c _GLOBAL_SUM_RL(alpha,myThid) |
| 391 |
CALL GLOBAL_SUM_TILE_RL( alphaTile, alpha, myThid ) |
| 392 |
CcnhDebugStarts |
| 393 |
C WRITE(*,*) ' CG3D: Iteration ',it3d-1,' SUM(s*q)= ',alpha |
| 394 |
CcnhDebugEnds |
| 395 |
alpha = eta_qrN/alpha |
| 396 |
CcnhDebugStarts |
| 397 |
C WRITE(*,*) ' CG3D: Iteration ',it3d-1,' alpha= ',alpha |
| 398 |
CcnhDebugEnds |
| 399 |
|
| 400 |
C== Update solution and residual vectors |
| 401 |
C Now compute "interior" points. |
| 402 |
err = 0. _d 0 |
| 403 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 404 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 405 |
errTile(bi,bj) = 0. _d 0 |
| 406 |
DO K=1,Nr |
| 407 |
DO J=1,sNy |
| 408 |
DO I=1,sNx |
| 409 |
cg3d_x(I,J,K,bi,bj)=cg3d_x(I,J,K,bi,bj) |
| 410 |
& +alpha*cg3d_s(I,J,K,bi,bj) |
| 411 |
cg3d_r(I,J,K,bi,bj)=cg3d_r(I,J,K,bi,bj) |
| 412 |
& -alpha*cg3d_q(I,J,K,bi,bj) |
| 413 |
errTile(bi,bj) = errTile(bi,bj) |
| 414 |
& +cg3d_r(I,J,K,bi,bj)*cg3d_r(I,J,K,bi,bj) |
| 415 |
ENDDO |
| 416 |
ENDDO |
| 417 |
ENDDO |
| 418 |
c err = err + errTile(bi,bj) |
| 419 |
ENDDO |
| 420 |
ENDDO |
| 421 |
|
| 422 |
c _GLOBAL_SUM_RL( err , myThid ) |
| 423 |
CALL GLOBAL_SUM_TILE_RL( errTile, err, myThid ) |
| 424 |
err = SQRT(err) |
| 425 |
actualIts = it3d |
| 426 |
actualResidual = err |
| 427 |
IF ( actualResidual .LT. cg3dTargetResidual ) GOTO 11 |
| 428 |
CALL EXCH_S3D_RL( cg3d_r, Nr, myThid ) |
| 429 |
|
| 430 |
10 CONTINUE |
| 431 |
11 CONTINUE |
| 432 |
|
| 433 |
C-- Un-normalise the answer |
| 434 |
DO bj=myByLo(myThid),myByHi(myThid) |
| 435 |
DO bi=myBxLo(myThid),myBxHi(myThid) |
| 436 |
DO K=1,Nr |
| 437 |
DO J=1,sNy |
| 438 |
DO I=1,sNx |
| 439 |
cg3d_x(I,J,K,bi,bj) = cg3d_x(I,J,K,bi,bj)/rhsNorm |
| 440 |
ENDDO |
| 441 |
ENDDO |
| 442 |
ENDDO |
| 443 |
ENDDO |
| 444 |
ENDDO |
| 445 |
|
| 446 |
Cadj _EXCH_XYZ_RL(cg3d_x, myThid ) |
| 447 |
c _BEGIN_MASTER( myThid ) |
| 448 |
c WRITE(*,'(A,I6,1PE30.14)') ' CG3D iters, err = ', |
| 449 |
c & actualIts, actualResidual |
| 450 |
c _END_MASTER( myThid ) |
| 451 |
lastResidual=actualResidual |
| 452 |
numIters=actualIts |
| 453 |
|
| 454 |
#endif /* ALLOW_NONHYDROSTATIC */ |
| 455 |
|
| 456 |
RETURN |
| 457 |
END |
Parent Directory
| 
Revision Log
| 
Revision Graph