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
|
subroutine dsposp(op,ma,na,a,nela,inda,mb,nb,b,nelb,indb,
c Copyright INRIA
$ nelc,indc,ierr)
c!purpose
c compare the elements of two sparse matrices.
c!parameters
c op : integer flag which specifies the comparison to perform
c see routine dcompa for more precision
c ma,na: row and column dimension of the a matrix
c mb,nb: row and column dimension of the b matrix
c a,b : arrays.
c Contain non zero elements of first,second sparse matrices.
c nela :integer: number of non zero elements of a
c nelb :integer: number of non zero elements of b
c nelc :integer:
c on entry maximum number of non zero elements of c
c on return number of non zero elements of c
c inda : a matrix control data:
c inda(i) 1<=i<=nr contains the number of ith row non zero elements
c of a
c inda(m+i) 1<=i<=nela column index of each non zero element
c indb : b matrix control data:
c indb(i) 1<=i<=nr contains the number of ith row non zero elements
c of b
c indb(m+i) 1<=i<=nelb column index of each non zero element
c
c indc : on return contains c matrix control data:
c indc(i) 1<=i<=nr contains the number of ith row non zero elements
c of c
c indc(m+i) 1<=i<=nelb column index of each non zero element
c ierr : if non zero initial value of nelc is to small
c !
double precision a(*),b(*),t
integer op,nela,inda(*),nelb,indb(*),nelc,indc(*),ierr
c
integer jc,ka,kb,jb,i,ja,j1,j2
logical dcompa,z
external dcompa
c
nr=max(ma,mb)
nc=max(na,nb)
if (ma.eq.mb .and. na.eq.nb .and. .not.dcompa(0.d0,0.d0,op)) then
call cmpsp(op, ma, na, nela, a, inda, inda(nr+1),
$ mb, nb, nelb, b, indb, indb(nr+1),
$ nelc, indc, indc(nr+1), ierr)
return
endif
c
nelmx=nelc
ierr=0
c jc counts elements of c.
jc = 1
c ka,kb are numbers in first i rows of a,b.
ka = 1
kb = 1
kc = 1
c jb counts elements of b.
jb = 1
c i counts rows of a,b,c.
if(ma*na.eq.1.and.mb*nb.gt.1) then
c compare all element of b with scalar a
t=0.0d0
if(inda(1).eq.1) t=a(1)
z=dcompa(t,0.0d0,op)
do 10 i=1,nr
indc(i)=0
nirb=indb(i)
jb=kb
jc=kc
if(nirb.eq.0) then
do 03 j=1,nc
if (z) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
03 continue
else
j2=indb(nr+jb)
do 04 j=1,nc
if(j2.eq.j) then
if (dcompa(t,b(jb),op)) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
if(jb-kb+1.lt.nirb) jb=jb+1
j2=indb(nr+jb)
else
if (z) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
endif
04 continue
endif
kb=kb+indb(i)
indc(i)=jc-kc
kc=jc
10 continue
elseif(ma*na.gt.1.and.mb*nb.eq.1) then
c compare all elements of a with scalar b
t=0.0d0
if(indb(1).eq.1) t=b(1)
z=dcompa(0.0d0,t,op)
do 20 i=1,nr
indc(i)=0
nira=inda(i)
ja=ka
jc=kc
if(nira.eq.0) then
if(z) then
if(kc+nc.gt.nelmx) goto 99
indc(i)=nc
do 11 j=1,nc
indc(nr+kc-1+j)=j
11 continue
jc=kc+nc
endif
else
j1=inda(nr+ja)
do 12 j=1,nc
if(j1.eq.j) then
if (dcompa(a(ja),t,op)) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
if(ja-ka+1.lt.nira) ja=ja+1
j1=inda(nr+ja)
elseif(z) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
12 continue
endif
indc(i)=jc-kc
ka=ka+nira
kc=jc
20 continue
else
z=dcompa(0.0d0,0.0d0,op)
do 30 i=1,nr
indc(i)=0
nira=inda(i)
nirb=indb(i)
ja=ka
jb=kb
jc=kc
if(nira.eq.0) then
if(nirb.eq.0) then
if(z) then
if(kc+nc.gt.nelmx) goto 99
indc(i)=nc
do 21 j=1,nc
indc(nr+kc-1+j)=j
21 continue
jc=kc+nc
endif
else
j2=indb(nr+jb)
do 22 j=1,nc
if(j2.eq.j) then
if (dcompa(0.0d0,b(jb),op)) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
if(jb-kb+1.lt.nirb) jb=jb+1
j2=indb(nr+jb)
elseif(z) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
22 continue
endif
else
if(nirb.eq.0) then
j1=inda(nr+ja)
do 23 j=1,nc
if(j1.eq.j) then
if (dcompa(a(ja),0.0d0,op)) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
if(ja-ka+1.lt.nira) ja=ja+1
j1=inda(nr+ja)
elseif(z) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
23 continue
else
j1=inda(nr+ja)
j2=indb(nr+jb)
do 24 j=1,nc
if(j1.eq.j) then
if(j2.eq.j) then
if (dcompa(a(ja),b(jb),op)) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
if(ja-ka+1.lt.nira) ja=ja+1
if(jb-kb+1.lt.nirb) jb=jb+1
j1=inda(nr+ja)
j2=indb(nr+jb)
else
if (dcompa(a(ja),0.0d0,op)) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
if(ja-ka+1.lt.nira) ja=ja+1
j1=inda(nr+ja)
endif
else
if(j2.eq.j) then
if (dcompa(0.0d0,b(jb),op)) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
if(jb-kb+1.lt.nirb) jb=jb+1
j2=indb(nr+jb)
elseif(z) then
if(jc+1.gt.nelmx) goto 99
indc(nr+jc)=j
jc=jc+1
endif
endif
24 continue
endif
endif
ka=ka+inda(i)
kb=kb+indb(i)
indc(i)=jc-kc
kc=jc
30 continue
endif
nelc = jc-1
return
c error messages.
99 ierr=1
c no more place for c
return
end
subroutine cmp_and_update(x, y, op, C_mnel_i, C_icol, j, kC,
$ C_nelmax, ierr)
double precision x, y
integer op, C_mnel_i, C_icol(*), j, kC, C_nelmax, ierr
logical dcompa
external dcompa
if ( dcompa(x, y, op) ) then
kC = kC + 1
if ( kC .gt. C_nelmax ) then
ierr = 1
return
endif
C_icol(kC) = j
C_mnel_i = C_mnel_i + 1
endif
end
subroutine cmpsp(op, A_m, A_n, A_nel, A_R, A_mnel, A_icol,
$ B_m, B_n, B_nel, B_R, B_mnel, B_icol,
$ C_nelmax, C_mnel, C_icol, ierr)
implicit none
integer op, A_m, A_n, A_nel, A_mnel(*), A_icol(*),
$ B_m, B_n, B_nel, B_mnel(*), B_icol(*),
$ C_nelmax, C_mnel(*), C_icol(*), ierr
double precision A_R(*), B_R(*)
integer kA, kAf, kB, kBf, kC, i, jA, jB, k
kAf = 0
kBf = 0
kC = 0
ierr = 0
do i = 1, A_m
kA = kAf + 1
kB = kBf + 1
kAf = kAf + A_mnel(i)
kBf = kBf + B_mnel(i)
C_mnel(i) = 0
100 if ( kA .gt. kAf .or. kB .gt. kBf ) goto 300
jA = A_icol(kA)
jB = B_icol(kB)
200 continue
if ( jA .lt. jB ) then
call cmp_and_update(A_R(kA), 0.d0, op, C_mnel(i),
$ C_icol, jA, kC, C_nelmax, ierr)
if (ierr .eq. 1 ) return
kA = kA + 1
if ( kA .gt. kAf ) goto 300
jA = A_icol(kA)
goto 200
elseif ( jA .gt. jB ) then
call cmp_and_update(0.d0, B_R(kB), op, C_mnel(i),
$ C_icol, jB, kC, C_nelmax, ierr)
if (ierr .eq. 1 ) return
kB = kB + 1
if ( kB .gt. kBf ) goto 300
jB = B_icol(kB)
goto 200
else ! jA = jB
call cmp_and_update(A_R(kA), B_R(kB), op, C_mnel(i),
$ C_icol, jA, kC, C_nelmax, ierr)
if (ierr .eq. 1 ) return
kA = kA + 1
kB = kB + 1
goto 100
endif
300 continue
if ( kA .le. kAf) then
do k = kA, kAf
call cmp_and_update(A_R(k), 0.d0, op, C_mnel(i), C_icol,
$ A_icol(k), kC, C_nelmax, ierr)
if (ierr .eq. 1 ) return
enddo
else if ( kB .le. kBf ) then
do k = kB, kBf
call cmp_and_update(0.d0, B_R(k), op, C_mnel(i), C_icol,
$ B_icol(k), kC, C_nelmax, ierr)
if (ierr .eq. 1 ) return
enddo
endif
enddo
C_nelmax = kC
end
|
