forked from altMITgcm/MITgcm
-
Notifications
You must be signed in to change notification settings - Fork 0
/
exch_recv_get_vec.F
383 lines (353 loc) · 12.5 KB
/
exch_recv_get_vec.F
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
C $Header: /u/gcmpack/MITgcm/pkg/flt/exch_recv_get_vec.F,v 1.1 2001/09/13 17:43:55 adcroft Exp $
#include "CPP_OPTIONS.h"
#include "CPP_EEOPTIONS.h"
SUBROUTINE EXCH_RL_RECV_GET_VEC_X( arrayE, arrayW,
I myd1, myThid )
C /==========================================================\
C | SUBROUTINE RECV_RL_GET_X |
C | o "Send" or "put" X edges for RL array. |
C |==========================================================|
C | Routine that invokes actual message passing send or |
C | direct "put" of data to update X faces of an XY[R] array.|
C \==========================================================/
IMPLICIT NONE
C == Global variables ==
#include "SIZE.h"
#include "EEPARAMS.h"
#include "EESUPPORT.h"
#include "FLT.h"
#include "EXCH.h"
C == Routine arguments ==
C arrayE - Arrays to exchange be exchanged.
C arrayW
C myd1 - sizes.
C myd2
C theSimulationMode - Forward or reverse mode exchange ( provides
C support for adjoint integration of code. )
C myThid - Thread number of this instance of S/R EXCH...
C eBl - Edge buffer level
INTEGER myd1
INTEGER myd2
_RL arrayE(myd1, nSx, nSy), arrayW(myd1, nSx, nSy)
INTEGER theSimulationMode
INTEGER myThid
CEndOfInterface
C == Local variables ==
C I, J - Loop counters and extents
C bi, bj
C biW, bjW - West tile indices
C biE, bjE - East tile indices
C theProc, theTag, theType, - Variables used in message building
C theSize
C westCommMode - Working variables holding type
C eastCommMode of communication a particular
C tile face uses.
INTEGER I, J
INTEGER bi, bj, biW, bjW, biE, bjE
INTEGER westCommMode
INTEGER eastCommMode
INTEGER spinCount
#ifdef ALLOW_USE_MPI
INTEGER theProc, theTag, theType, theSize
INTEGER mpiStatus(MPI_STATUS_SIZE,4), mpiRc
#endif
C-- Under a "put" scenario we
C-- i. set completetion signal for buffer we put into.
C-- ii. wait for completetion signal indicating data has been put in
C-- our buffer.
C-- Under a messaging mode we "receive" the message.
C-- Under a "get" scenario we
C-- i. Check that the data is ready.
C-- ii. Read the data.
C-- iii. Set data read flag + memory sync.
DO bj=myByLo(myThid),myByHi(myThid)
DO bi=myBxLo(myThid),myBxHi(myThid)
westCommMode = _tileCommModeW(bi,bj)
eastCommMode = _tileCommModeE(bi,bj)
biE = _tileBiE(bi,bj)
bjE = _tileBjE(bi,bj)
biW = _tileBiW(bi,bj)
bjW = _tileBjW(bi,bj)
IF ( westCommMode .EQ. COMM_MSG ) THEN
#ifdef ALLOW_USE_MPI
#ifndef ALWAYS_USE_MPI
IF ( usingMPI ) THEN
#endif
theProc = tilePidW(bi,bj)
theTag = _tileTagRecvW(bi,bj)
theType = MPI_DOUBLE_PRECISION
theSize = myd1
CALL MPI_Recv( arrayW(1,bi,bj), theSize, theType,
& theProc, theTag, MPI_COMM_MODEL,
& mpiStatus, mpiRc )
c if (theProc .eq. 0 .or. theProc .eq. 2) then
c if (arrayW(1,bi,bj) .ne. 0.) then
c write(errormessageunit,*) 'qq2y: ',myprocid,
c & theProc,theTag,theSize,(arrayW(i,bi,bj),i=1,32)
c else
c write(errormessageunit,*) 'qq2n: ',myprocid,
c & theProc,theTag,theSize,(arrayW(i,bi,bj),i=1,32)
c endif
c endif
#ifndef ALWAYS_USE_MPI
ENDIF
#endif
#endif /* ALLOW_USE_MPI */
ENDIF
IF ( eastCommMode .EQ. COMM_MSG ) THEN
#ifdef ALLOW_USE_MPI
#ifndef ALWAYS_USE_MPI
IF ( usingMPI ) THEN
#endif
theProc = tilePidE(bi,bj)
theTag = _tileTagRecvE(bi,bj)
theType = MPI_DOUBLE_PRECISION
theSize = myd1
CALL MPI_Recv( arrayE(1,bi,bj), theSize, theType,
& theProc, theTag, MPI_COMM_MODEL,
& mpiStatus, mpiRc )
#ifndef ALWAYS_USE_MPI
ENDIF
#endif
#endif /* ALLOW_USE_MPI */
ENDIF
ENDDO
ENDDO
C-- Wait for buffers I am going read to be ready.
IF ( exchUsesBarrier ) THEN
C o On some machines ( T90 ) use system barrier rather than spinning.
CALL BARRIER( myThid )
ELSE
C o Spin waiting for completetion flag. This avoids a global-lock
C i.e. we only lock waiting for data that we need.
DO bj=myByLo(myThid),myByHi(myThid)
DO bi=myBxLo(myThid),myBxHi(myThid)
spinCount = 0
westCommMode = _tileCommModeW(bi,bj)
eastCommMode = _tileCommModeE(bi,bj)
10 CONTINUE
CALL FOOL_THE_COMPILER
spinCount = spinCount+1
C IF ( myThid .EQ. 1 .AND. spinCount .GT. _EXCH_SPIN_LIMIT ) THEN
C WRITE(0,*) ' eBl = ', ebl
C STOP ' S/R EXCH_RECV_GET_X: spinCount .GT. _EXCH_SPIN_LIMIT'
C ENDIF
IF ( westRecvAck(1,bi,bj) .EQ. 0. ) GOTO 10
IF ( eastRecvAck(1,bi,bj) .EQ. 0. ) GOTO 10
C Clear outstanding requests
westRecvAck(1,bi,bj) = 0.
eastRecvAck(1,bi,bj) = 0.
c IF ( exchVReqsX(1,bi,bj) .GT. 0 ) THEN
IF ( exchNReqsX(1,bi,bj) .GT. 0 ) THEN
#ifdef ALLOW_USE_MPI
#ifndef ALWAYS_USE_MPI
IF ( usingMPI ) THEN
#endif
c CALL MPI_Waitall( exchVReqsX(1,bi,bj), exchReqVIdX(1,1,bi,bj),
CALL MPI_Waitall( exchNReqsX(1,bi,bj), exchReqIdX(1,1,bi,bj),
& mpiStatus, mpiRC )
#ifndef ALWAYS_USE_MPI
ENDIF
#endif
#endif /* ALLOW_USE_MPI */
ENDIF
C Clear outstanding requests counter
c exchVReqsX(1,bi,bj) = 0
exchNReqsX(1,bi,bj) = 0
C Update statistics
ENDDO
ENDDO
ENDIF
C-- Read from the buffers
DO bj=myByLo(myThid),myByHi(myThid)
DO bi=myBxLo(myThid),myBxHi(myThid)
biE = _tileBiE(bi,bj)
bjE = _tileBjE(bi,bj)
biW = _tileBiW(bi,bj)
bjW = _tileBjW(bi,bj)
westCommMode = _tileCommModeW(bi,bj)
eastCommMode = _tileCommModeE(bi,bj)
IF ( eastCommMode .EQ. COMM_GET ) THEN
DO I=1,myd1
arrayE(I,bi,bj) = arrayW(I,biE,bjE)
ENDDO
ENDIF
IF ( westCommMode .EQ. COMM_GET ) THEN
DO I=1,myd1
arrayW(I,bi,bj) = arrayE(I,biW,bjW)
ENDDO
ENDIF
ENDDO
ENDDO
RETURN
END
SUBROUTINE EXCH_RL_RECV_GET_VEC_Y( arrayN, arrayS,
I myd1, myThid )
C /==========================================================\
C | SUBROUTINE RECV_RL_GET_Y |
C | o "Send" or "put" Y edges for RL array. |
C |==========================================================|
C | Routine that invokes actual message passing send or |
C | direct "put" of data to update Y faces of an XY[R] array.|
C \==========================================================/
IMPLICIT NONE
C == Global variables ==
#include "SIZE.h"
#include "EEPARAMS.h"
#include "EESUPPORT.h"
#include "FLT.h"
#include "EXCH.h"
C == Routine arguments ==
C arrayN - Arrays to exchange be exchanged.
C arrayS
C myd1 - sizes.
C myd2
C theSimulationMode - Forward or reverse mode exchange ( provides
C support for adjoint integration of code. )
C myThid - Thread number of this instance of S/R EXCH...
INTEGER myd1
INTEGER myd2
_RL arrayN(myd1, nSx, nSy), arrayS(myd1, nSx, nSy)
INTEGER theSimulationMode
INTEGER myThid
CEndOfInterface
C == Local variables ==
C I, J - Loop counters and extents
C bi, bj
C biS, bjS - South tile indices
C biE, bjE - North tile indices
C theProc, theTag, theType, - Variables used in message building
C theSize
C southCommMode - Working variables holding type
C northCommMode of communication a particular
C tile face uses.
INTEGER I, J
INTEGER bi, bj, biS, bjS, biN, bjN
INTEGER southCommMode
INTEGER northCommMode
INTEGER spinCount
#ifdef ALLOW_USE_MPI
INTEGER theProc, theTag, theType, theSize
INTEGER mpiStatus(MPI_STATUS_SIZE,4), mpiRc
#endif
C-- Under a "put" scenario we
C-- i. set completetion signal for buffer we put into.
C-- ii. wait for completetion signal indicating data has been put in
C-- our buffer.
C-- Under a messaging mode we "receive" the message.
C-- Under a "get" scenario we
C-- i. Check that the data is ready.
C-- ii. Read the data.
C-- iii. Set data read flag + memory sync.
DO bj=myByLo(myThid),myByHi(myThid)
DO bi=myBxLo(myThid),myBxHi(myThid)
southCommMode = _tileCommModeS(bi,bj)
northCommMode = _tileCommModeN(bi,bj)
biN = _tileBiN(bi,bj)
bjN = _tileBjN(bi,bj)
biS = _tileBiS(bi,bj)
bjS = _tileBjS(bi,bj)
IF ( southCommMode .EQ. COMM_MSG ) THEN
#ifdef ALLOW_USE_MPI
#ifndef ALWAYS_USE_MPI
IF ( usingMPI ) THEN
#endif
theProc = tilePidS(bi,bj)
theTag = _tileTagRecvS(bi,bj)
theType = MPI_DOUBLE_PRECISION
theSize = myd1
CALL MPI_Recv( arrayS(1,bi,bj), theSize, theType,
& theProc, theTag, MPI_COMM_MODEL,
& mpiStatus, mpiRc )
#ifndef ALWAYS_USE_MPI
ENDIF
#endif
#endif /* ALLOW_USE_MPI */
ENDIF
IF ( northCommMode .EQ. COMM_MSG ) THEN
#ifdef ALLOW_USE_MPI
#ifndef ALWAYS_USE_MPI
IF ( usingMPI ) THEN
#endif
theProc = tilePidN(bi,bj)
theTag = _tileTagRecvN(bi,bj)
theType = MPI_DOUBLE_PRECISION
theSize = myd1
CALL MPI_Recv( arrayN(1,bi,bj), theSize, theType,
& theProc, theTag, MPI_COMM_MODEL,
& mpiStatus, mpiRc )
#ifndef ALWAYS_USE_MPI
ENDIF
#endif
#endif /* ALLOW_USE_MPI */
ENDIF
ENDDO
ENDDO
C-- Wait for buffers I am going read to be ready.
IF ( exchUsesBarrier ) THEN
C o On some machines ( T90 ) use system barrier rather than spinning.
CALL BARRIER( myThid )
ELSE
C o Spin waiting for completetion flag. This avoids a global-lock
C i.e. we only lock waiting for data that we need.
DO bj=myByLo(myThid),myByHi(myThid)
DO bi=myBxLo(myThid),myBxHi(myThid)
spinCount = 0
southCommMode = _tileCommModeS(bi,bj)
northCommMode = _tileCommModeN(bi,bj)
10 CONTINUE
CALL FOOL_THE_COMPILER
spinCount = spinCount+1
C IF ( myThid .EQ. 1 .AND. spinCount .GT. _EXCH_SPIN_LIMIT ) THEN
C WRITE(0,*) ' eBl = ', ebl
C STOP ' S/R EXCH_RECV_GET_X: spinCount .GT. _EXCH_SPIN_LIMIT'
C ENDIF
IF ( southRecvAck(1,bi,bj) .EQ. 0. ) GOTO 10
IF ( northRecvAck(1,bi,bj) .EQ. 0. ) GOTO 10
C Clear outstanding requests
southRecvAck(1,bi,bj) = 0.
northRecvAck(1,bi,bj) = 0.
c IF ( exchVReqsY(1,bi,bj) .GT. 0 ) THEN
IF ( exchNReqsY(1,bi,bj) .GT. 0 ) THEN
#ifdef ALLOW_USE_MPI
#ifndef ALWAYS_USE_MPI
IF ( usingMPI ) THEN
#endif
c CALL MPI_Waitall( exchVReqsY(1,bi,bj), exchReqVIdY(1,1,bi,bj),
CALL MPI_Waitall( exchNReqsY(1,bi,bj), exchReqIdY(1,1,bi,bj),
& mpiStatus, mpiRC )
#ifndef ALWAYS_USE_MPI
ENDIF
#endif
#endif /* ALLOW_USE_MPI */
ENDIF
C Clear outstanding requests counter
c exchVReqsY(1,bi,bj) = 0
exchNReqsY(1,bi,bj) = 0
C Update statistics
ENDDO
ENDDO
ENDIF
C-- Read from the buffers
DO bj=myByLo(myThid),myByHi(myThid)
DO bi=myBxLo(myThid),myBxHi(myThid)
biN = _tileBiN(bi,bj)
bjN = _tileBjN(bi,bj)
biS = _tileBiS(bi,bj)
bjS = _tileBjS(bi,bj)
southCommMode = _tileCommModeS(bi,bj)
northCommMode = _tileCommModeN(bi,bj)
IF ( southCommMode .EQ. COMM_GET ) THEN
DO I=1,myd1
arrayN(I,bi,bj) = arrayS(I,biN,bjN)
ENDDO
ENDIF
IF ( southCommMode .EQ. COMM_GET ) THEN
DO I=1,myd1
arrayS(I,bi,bj) = arrayN(I,biS,bjS)
ENDDO
ENDIF
ENDDO
ENDDO
RETURN
END