1 |
C $Header: /u/gcmpack/models/MITgcmUV/eesupp/src/global_sum.F,v 1.8.2.1 2001/04/12 10:52:48 cnh Exp $ |
2 |
C $Name: $ |
3 |
|
4 |
C-- File global_sum.F: Routines that perform global sum on an array |
5 |
C of thread values. |
6 |
C Contents |
7 |
C o global_sum_r4 |
8 |
C o global_sum_r8 |
9 |
#include "CPP_EEOPTIONS.h" |
10 |
|
11 |
CStartOfInterface |
12 |
SUBROUTINE GLOBAL_SUM_R4( |
13 |
U sumPhi, |
14 |
I myThid ) |
15 |
C /==========================================================\ |
16 |
C | SUBROUTINE GLOBAL_SUM_R4 | |
17 |
C | o Handle sum for real*4 data. | |
18 |
C |==========================================================| |
19 |
C | Perform sum an array of one value per thread and then | |
20 |
C | sum result of all the processes. | |
21 |
C | Notes | |
22 |
C | ===== | |
23 |
C | Within a process only one thread does the sum, each | |
24 |
C | thread is assumed to have already summed its local data. | |
25 |
C | The same thread also does the inter-process sum for | |
26 |
C | example with MPI and then writes the result into a shared| |
27 |
C | location. All threads wait until the sum is avaiailable. | |
28 |
C \==========================================================/ |
29 |
IMPLICIT NONE |
30 |
|
31 |
C == Global data == |
32 |
#include "SIZE.h" |
33 |
#include "EEPARAMS.h" |
34 |
#include "EESUPPORT.h" |
35 |
#include "GLOBAL_SUM.h" |
36 |
|
37 |
C == Routine arguments == |
38 |
C sumPhi - Result of sum. |
39 |
C myThid - My thread id. |
40 |
Real*4 sumPhi |
41 |
INTEGER myThid |
42 |
CEndOfInterface |
43 |
|
44 |
C == Local variables == |
45 |
C I - Loop counters |
46 |
C mpiRC - MPI return code |
47 |
INTEGER I |
48 |
Real*4 tmp |
49 |
#ifdef ALLOW_USE_MPI |
50 |
INTEGER mpiRC |
51 |
#endif /* ALLOW_USE_MPI */ |
52 |
|
53 |
C-- write local sum into array |
54 |
CALL BAR2( myThid ) |
55 |
phiGSRS(1,myThid) = sumPhi |
56 |
|
57 |
C-- Can not start until everyone is ready |
58 |
CALL BAR2( myThid ) |
59 |
|
60 |
C-- Sum within the process first |
61 |
_BEGIN_MASTER( myThid ) |
62 |
tmp = 0. |
63 |
DO I=1,nThreads |
64 |
tmp = tmp + phiGSRS(1,I) |
65 |
ENDDO |
66 |
sumPhi = tmp |
67 |
#ifdef ALLOW_USE_MPI |
68 |
#ifndef ALWAYS_USE_MPI |
69 |
IF ( usingMPI ) THEN |
70 |
#endif |
71 |
CALL MPI_Allreduce(tmp,sumPhi,1,MPI_REAL,MPI_SUM, |
72 |
& MPI_COMM_WORLD,mpiRC) |
73 |
#ifndef ALWAYS_USE_MPI |
74 |
ENDIF |
75 |
#endif |
76 |
#endif /* ALLOW_USE_MPI */ |
77 |
C-- Write solution to place where all threads can see it |
78 |
phiGSRS(1,1) = sumPhi |
79 |
|
80 |
_END_MASTER( myThid ) |
81 |
C-- |
82 |
CALL BAR2( myThid ) |
83 |
|
84 |
C-- set result for every process |
85 |
sumPhi = phiGSRS(1,1) |
86 |
CALL BAR2( myThid ) |
87 |
|
88 |
RETURN |
89 |
END |
90 |
|
91 |
|
92 |
CStartOfInterface |
93 |
SUBROUTINE GLOBAL_SUM_R8( |
94 |
U sumPhi, |
95 |
I myThid ) |
96 |
C /==========================================================\ |
97 |
C | SUBROUTINE GLOBAL_SUM_R8 | |
98 |
C | o Handle sum for real*8 data. | |
99 |
C |==========================================================| |
100 |
C | Perform sum an array of one value per thread and then | |
101 |
C | sum result of all the processes. | |
102 |
C | Notes | |
103 |
C | ===== | |
104 |
C | Within a process only one thread does the sum, each | |
105 |
C | thread is assumed to have already summed its local data. | |
106 |
C | The same thread also does the inter-process sum for | |
107 |
C | example with MPI and then writes the result into a shared| |
108 |
C | location. All threads wait until the sum is avaiailable. | |
109 |
C \==========================================================/ |
110 |
IMPLICIT NONE |
111 |
|
112 |
C === Global data === |
113 |
#include "SIZE.h" |
114 |
#include "EEPARAMS.h" |
115 |
#include "EESUPPORT.h" |
116 |
#include "GLOBAL_SUM.h" |
117 |
|
118 |
C === Routine arguments === |
119 |
C sumPhi - Result of sum. |
120 |
C myThid - My thread id. |
121 |
Real*8 sumPhi |
122 |
INTEGER myThid |
123 |
CEndOfInterface |
124 |
|
125 |
C === Local variables === |
126 |
C I - Loop counters |
127 |
C mpiRC - MPI return code |
128 |
INTEGER I |
129 |
Real*8 tmp |
130 |
#ifdef ALLOW_USE_MPI |
131 |
INTEGER mpiRC |
132 |
#endif /* ALLOW_USE_MPI */ |
133 |
|
134 |
CALL BAR2( myThid ) |
135 |
C-- write local sum into array |
136 |
phiGSRL(1,myThid) = sumPhi |
137 |
|
138 |
C-- Can not start until everyone is ready |
139 |
C CALL FOOL_THE_COMPILER( phiGSRL ) |
140 |
C CALL MS |
141 |
CALL BAR2( myThid ) |
142 |
C _BARRIER |
143 |
C _BARRIER |
144 |
C CALL FOOL_THE_COMPILER( phiGSRL ) |
145 |
|
146 |
C-- Sum within the process first |
147 |
_BEGIN_MASTER( myThid ) |
148 |
tmp = 0. _d 0 |
149 |
DO I=1,nThreads |
150 |
tmp = tmp + phiGSRL(1,I) |
151 |
ENDDO |
152 |
sumPhi = tmp |
153 |
#ifdef ALLOW_USE_MPI |
154 |
#ifndef ALWAYS_USE_MPI |
155 |
IF ( usingMPI ) THEN |
156 |
#endif |
157 |
CALL MPI_Allreduce(tmp,sumPhi,1,MPI_DOUBLE_PRECISION,MPI_SUM, |
158 |
& MPI_COMM_WORLD,mpiRC) |
159 |
#ifndef ALWAYS_USE_MPI |
160 |
ENDIF |
161 |
#endif |
162 |
#endif /* ALLOW_USE_MPI */ |
163 |
C-- Write solution to place where all threads can see it |
164 |
phiGSRL(1,1) = sumPhi |
165 |
_END_MASTER( myThid ) |
166 |
|
167 |
C-- Do not leave until we are sure that the sum is done |
168 |
C CALL FOOL_THE_COMPILER( phiGSRL ) |
169 |
C CALL MS |
170 |
C _BARRIER |
171 |
C _BARRIER |
172 |
CALL BAR2( myThid ) |
173 |
C CALL FOOL_THE_COMPILER( phiGSRL ) |
174 |
|
175 |
C-- set result for every process |
176 |
sumPhi = phiGSRL(1,1) |
177 |
CALL BAR2( myThid ) |
178 |
|
179 |
RETURN |
180 |
END |
181 |
CStartOfInterface |
182 |
SUBROUTINE GLOBAL_SUM_INT( |
183 |
U sumPhi, |
184 |
I myThid ) |
185 |
C /==========================================================\ |
186 |
C | SUBROUTINE GLOBAL_SUM_INT | |
187 |
C | o Handle sum for integer data. | |
188 |
C |==========================================================| |
189 |
C | Perform sum an array of one value per thread and then | |
190 |
C | sum result of all the processes. | |
191 |
C | Notes | |
192 |
C | ===== | |
193 |
C | Within a process only one thread does the sum, each | |
194 |
C | thread is assumed to have already summed its local data. | |
195 |
C | The same thread also does the inter-process sum for | |
196 |
C | example with MPI and then writes the result into a shared| |
197 |
C | location. All threads wait until the sum is avaiailable. | |
198 |
C \==========================================================/ |
199 |
IMPLICIT NONE |
200 |
|
201 |
C === Global data === |
202 |
#include "SIZE.h" |
203 |
#include "EEPARAMS.h" |
204 |
#include "EESUPPORT.h" |
205 |
#include "GLOBAL_SUM.h" |
206 |
|
207 |
C === Routine arguments === |
208 |
C sumPhi - Result of sum. |
209 |
C myThid - My thread id. |
210 |
INTEGER sumPhi |
211 |
INTEGER myThid |
212 |
CEndOfInterface |
213 |
|
214 |
C === Local variables === |
215 |
C I - Loop counters |
216 |
C mpiRC - MPI return code |
217 |
INTEGER I |
218 |
INTEGER tmp |
219 |
#ifdef ALLOW_USE_MPI |
220 |
INTEGER mpiRC |
221 |
#endif /* ALLOW_USE_MPI */ |
222 |
|
223 |
_BARRIER |
224 |
C-- write local sum into array |
225 |
phiGSI(1,myThid) = sumPhi |
226 |
|
227 |
C-- Can not start until everyone is ready |
228 |
_BARRIER |
229 |
|
230 |
C-- Sum within the process first |
231 |
_BEGIN_MASTER( myThid ) |
232 |
tmp = 0. _d 0 |
233 |
DO I=1,nThreads |
234 |
tmp = tmp + phiGSI(1,I) |
235 |
ENDDO |
236 |
sumPhi = tmp |
237 |
#ifdef ALLOW_USE_MPI |
238 |
#ifndef ALWAYS_USE_MPI |
239 |
IF ( usingMPI ) THEN |
240 |
#endif |
241 |
CALL MPI_Allreduce(tmp,sumPhi,1,MPI_INTEGER,MPI_SUM, |
242 |
& MPI_COMM_WORLD,mpiRC) |
243 |
#ifndef ALWAYS_USE_MPI |
244 |
ENDIF |
245 |
#endif |
246 |
#endif /* ALLOW_USE_MPI */ |
247 |
C-- Write solution to place where all threads can see it |
248 |
phiGSI(1,1) = sumPhi |
249 |
_END_MASTER( myThid ) |
250 |
|
251 |
C-- Do not leave until we are sure that the sum is done |
252 |
_BARRIER |
253 |
|
254 |
C-- set result for every process |
255 |
sumPhi = phiGSI(1,1) |
256 |
_BARRIER |
257 |
|
258 |
RETURN |
259 |
END |