1 SUBROUTINE DLATM5( PRTYPE, M, N, A, LDA, B, LDB, C, LDC, D, LDD,
2 $ E, LDE, F, LDF, R, LDR, L, LDL, ALPHA, QBLCKA,
3 $ QBLCKB )
4 *
5 * -- LAPACK test routine (version 3.1) --
6 * Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
7 * November 2006
8 *
9 * .. Scalar Arguments ..
10 INTEGER LDA, LDB, LDC, LDD, LDE, LDF, LDL, LDR, M, N,
11 $ PRTYPE, QBLCKA, QBLCKB
12 DOUBLE PRECISION ALPHA
13 * ..
14 * .. Array Arguments ..
15 DOUBLE PRECISION A( LDA, * ), B( LDB, * ), C( LDC, * ),
16 $ D( LDD, * ), E( LDE, * ), F( LDF, * ),
17 $ L( LDL, * ), R( LDR, * )
18 * ..
19 *
20 * Purpose
21 * =======
22 *
23 * DLATM5 generates matrices involved in the Generalized Sylvester
24 * equation:
25 *
26 * A * R - L * B = C
27 * D * R - L * E = F
28 *
29 * They also satisfy (the diagonalization condition)
30 *
31 * [ I -L ] ( [ A -C ], [ D -F ] ) [ I R ] = ( [ A ], [ D ] )
32 * [ I ] ( [ B ] [ E ] ) [ I ] ( [ B ] [ E ] )
33 *
34 *
35 * Arguments
36 * =========
37 *
38 * PRTYPE (input) INTEGER
39 * "Points" to a certian type of the matrices to generate
40 * (see futher details).
41 *
42 * M (input) INTEGER
43 * Specifies the order of A and D and the number of rows in
44 * C, F, R and L.
45 *
46 * N (input) INTEGER
47 * Specifies the order of B and E and the number of columns in
48 * C, F, R and L.
49 *
50 * A (output) DOUBLE PRECISION array, dimension (LDA, M).
51 * On exit A M-by-M is initialized according to PRTYPE.
52 *
53 * LDA (input) INTEGER
54 * The leading dimension of A.
55 *
56 * B (output) DOUBLE PRECISION array, dimension (LDB, N).
57 * On exit B N-by-N is initialized according to PRTYPE.
58 *
59 * LDB (input) INTEGER
60 * The leading dimension of B.
61 *
62 * C (output) DOUBLE PRECISION array, dimension (LDC, N).
63 * On exit C M-by-N is initialized according to PRTYPE.
64 *
65 * LDC (input) INTEGER
66 * The leading dimension of C.
67 *
68 * D (output) DOUBLE PRECISION array, dimension (LDD, M).
69 * On exit D M-by-M is initialized according to PRTYPE.
70 *
71 * LDD (input) INTEGER
72 * The leading dimension of D.
73 *
74 * E (output) DOUBLE PRECISION array, dimension (LDE, N).
75 * On exit E N-by-N is initialized according to PRTYPE.
76 *
77 * LDE (input) INTEGER
78 * The leading dimension of E.
79 *
80 * F (output) DOUBLE PRECISION array, dimension (LDF, N).
81 * On exit F M-by-N is initialized according to PRTYPE.
82 *
83 * LDF (input) INTEGER
84 * The leading dimension of F.
85 *
86 * R (output) DOUBLE PRECISION array, dimension (LDR, N).
87 * On exit R M-by-N is initialized according to PRTYPE.
88 *
89 * LDR (input) INTEGER
90 * The leading dimension of R.
91 *
92 * L (output) DOUBLE PRECISION array, dimension (LDL, N).
93 * On exit L M-by-N is initialized according to PRTYPE.
94 *
95 * LDL (input) INTEGER
96 * The leading dimension of L.
97 *
98 * ALPHA (input) DOUBLE PRECISION
99 * Parameter used in generating PRTYPE = 1 and 5 matrices.
100 *
101 * QBLCKA (input) INTEGER
102 * When PRTYPE = 3, specifies the distance between 2-by-2
103 * blocks on the diagonal in A. Otherwise, QBLCKA is not
104 * referenced. QBLCKA > 1.
105 *
106 * QBLCKB (input) INTEGER
107 * When PRTYPE = 3, specifies the distance between 2-by-2
108 * blocks on the diagonal in B. Otherwise, QBLCKB is not
109 * referenced. QBLCKB > 1.
110 *
111 *
112 * Further Details
113 * ===============
114 *
115 * PRTYPE = 1: A and B are Jordan blocks, D and E are identity matrices
116 *
117 * A : if (i == j) then A(i, j) = 1.0
118 * if (j == i + 1) then A(i, j) = -1.0
119 * else A(i, j) = 0.0, i, j = 1...M
120 *
121 * B : if (i == j) then B(i, j) = 1.0 - ALPHA
122 * if (j == i + 1) then B(i, j) = 1.0
123 * else B(i, j) = 0.0, i, j = 1...N
124 *
125 * D : if (i == j) then D(i, j) = 1.0
126 * else D(i, j) = 0.0, i, j = 1...M
127 *
128 * E : if (i == j) then E(i, j) = 1.0
129 * else E(i, j) = 0.0, i, j = 1...N
130 *
131 * L = R are chosen from [-10...10],
132 * which specifies the right hand sides (C, F).
133 *
134 * PRTYPE = 2 or 3: Triangular and/or quasi- triangular.
135 *
136 * A : if (i <= j) then A(i, j) = [-1...1]
137 * else A(i, j) = 0.0, i, j = 1...M
138 *
139 * if (PRTYPE = 3) then
140 * A(k + 1, k + 1) = A(k, k)
141 * A(k + 1, k) = [-1...1]
142 * sign(A(k, k + 1) = -(sin(A(k + 1, k))
143 * k = 1, M - 1, QBLCKA
144 *
145 * B : if (i <= j) then B(i, j) = [-1...1]
146 * else B(i, j) = 0.0, i, j = 1...N
147 *
148 * if (PRTYPE = 3) then
149 * B(k + 1, k + 1) = B(k, k)
150 * B(k + 1, k) = [-1...1]
151 * sign(B(k, k + 1) = -(sign(B(k + 1, k))
152 * k = 1, N - 1, QBLCKB
153 *
154 * D : if (i <= j) then D(i, j) = [-1...1].
155 * else D(i, j) = 0.0, i, j = 1...M
156 *
157 *
158 * E : if (i <= j) then D(i, j) = [-1...1]
159 * else E(i, j) = 0.0, i, j = 1...N
160 *
161 * L, R are chosen from [-10...10],
162 * which specifies the right hand sides (C, F).
163 *
164 * PRTYPE = 4 Full
165 * A(i, j) = [-10...10]
166 * D(i, j) = [-1...1] i,j = 1...M
167 * B(i, j) = [-10...10]
168 * E(i, j) = [-1...1] i,j = 1...N
169 * R(i, j) = [-10...10]
170 * L(i, j) = [-1...1] i = 1..M ,j = 1...N
171 *
172 * L, R specifies the right hand sides (C, F).
173 *
174 * PRTYPE = 5 special case common and/or close eigs.
175 *
176 * =====================================================================
177 *
178 * .. Parameters ..
179 DOUBLE PRECISION ONE, ZERO, TWENTY, HALF, TWO
180 PARAMETER ( ONE = 1.0D+0, ZERO = 0.0D+0, TWENTY = 2.0D+1,
181 $ HALF = 0.5D+0, TWO = 2.0D+0 )
182 * ..
183 * .. Local Scalars ..
184 INTEGER I, J, K
185 DOUBLE PRECISION IMEPS, REEPS
186 * ..
187 * .. Intrinsic Functions ..
188 INTRINSIC DBLE, MOD, SIN
189 * ..
190 * .. External Subroutines ..
191 EXTERNAL DGEMM
192 * ..
193 * .. Executable Statements ..
194 *
195 IF( PRTYPE.EQ.1 ) THEN
196 DO 20 I = 1, M
197 DO 10 J = 1, M
198 IF( I.EQ.J ) THEN
199 A( I, J ) = ONE
200 D( I, J ) = ONE
201 ELSE IF( I.EQ.J-1 ) THEN
202 A( I, J ) = -ONE
203 D( I, J ) = ZERO
204 ELSE
205 A( I, J ) = ZERO
206 D( I, J ) = ZERO
207 END IF
208 10 CONTINUE
209 20 CONTINUE
210 *
211 DO 40 I = 1, N
212 DO 30 J = 1, N
213 IF( I.EQ.J ) THEN
214 B( I, J ) = ONE - ALPHA
215 E( I, J ) = ONE
216 ELSE IF( I.EQ.J-1 ) THEN
217 B( I, J ) = ONE
218 E( I, J ) = ZERO
219 ELSE
220 B( I, J ) = ZERO
221 E( I, J ) = ZERO
222 END IF
223 30 CONTINUE
224 40 CONTINUE
225 *
226 DO 60 I = 1, M
227 DO 50 J = 1, N
228 R( I, J ) = ( HALF-SIN( DBLE( I / J ) ) )*TWENTY
229 L( I, J ) = R( I, J )
230 50 CONTINUE
231 60 CONTINUE
232 *
233 ELSE IF( PRTYPE.EQ.2 .OR. PRTYPE.EQ.3 ) THEN
234 DO 80 I = 1, M
235 DO 70 J = 1, M
236 IF( I.LE.J ) THEN
237 A( I, J ) = ( HALF-SIN( DBLE( I ) ) )*TWO
238 D( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWO
239 ELSE
240 A( I, J ) = ZERO
241 D( I, J ) = ZERO
242 END IF
243 70 CONTINUE
244 80 CONTINUE
245 *
246 DO 100 I = 1, N
247 DO 90 J = 1, N
248 IF( I.LE.J ) THEN
249 B( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWO
250 E( I, J ) = ( HALF-SIN( DBLE( J ) ) )*TWO
251 ELSE
252 B( I, J ) = ZERO
253 E( I, J ) = ZERO
254 END IF
255 90 CONTINUE
256 100 CONTINUE
257 *
258 DO 120 I = 1, M
259 DO 110 J = 1, N
260 R( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWENTY
261 L( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWENTY
262 110 CONTINUE
263 120 CONTINUE
264 *
265 IF( PRTYPE.EQ.3 ) THEN
266 IF( QBLCKA.LE.1 )
267 $ QBLCKA = 2
268 DO 130 K = 1, M - 1, QBLCKA
269 A( K+1, K+1 ) = A( K, K )
270 A( K+1, K ) = -SIN( A( K, K+1 ) )
271 130 CONTINUE
272 *
273 IF( QBLCKB.LE.1 )
274 $ QBLCKB = 2
275 DO 140 K = 1, N - 1, QBLCKB
276 B( K+1, K+1 ) = B( K, K )
277 B( K+1, K ) = -SIN( B( K, K+1 ) )
278 140 CONTINUE
279 END IF
280 *
281 ELSE IF( PRTYPE.EQ.4 ) THEN
282 DO 160 I = 1, M
283 DO 150 J = 1, M
284 A( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWENTY
285 D( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWO
286 150 CONTINUE
287 160 CONTINUE
288 *
289 DO 180 I = 1, N
290 DO 170 J = 1, N
291 B( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWENTY
292 E( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWO
293 170 CONTINUE
294 180 CONTINUE
295 *
296 DO 200 I = 1, M
297 DO 190 J = 1, N
298 R( I, J ) = ( HALF-SIN( DBLE( J / I ) ) )*TWENTY
299 L( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWO
300 190 CONTINUE
301 200 CONTINUE
302 *
303 ELSE IF( PRTYPE.GE.5 ) THEN
304 REEPS = HALF*TWO*TWENTY / ALPHA
305 IMEPS = ( HALF-TWO ) / ALPHA
306 DO 220 I = 1, M
307 DO 210 J = 1, N
308 R( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*ALPHA / TWENTY
309 L( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*ALPHA / TWENTY
310 210 CONTINUE
311 220 CONTINUE
312 *
313 DO 230 I = 1, M
314 D( I, I ) = ONE
315 230 CONTINUE
316 *
317 DO 240 I = 1, M
318 IF( I.LE.4 ) THEN
319 A( I, I ) = ONE
320 IF( I.GT.2 )
321 $ A( I, I ) = ONE + REEPS
322 IF( MOD( I, 2 ).NE.0 .AND. I.LT.M ) THEN
323 A( I, I+1 ) = IMEPS
324 ELSE IF( I.GT.1 ) THEN
325 A( I, I-1 ) = -IMEPS
326 END IF
327 ELSE IF( I.LE.8 ) THEN
328 IF( I.LE.6 ) THEN
329 A( I, I ) = REEPS
330 ELSE
331 A( I, I ) = -REEPS
332 END IF
333 IF( MOD( I, 2 ).NE.0 .AND. I.LT.M ) THEN
334 A( I, I+1 ) = ONE
335 ELSE IF( I.GT.1 ) THEN
336 A( I, I-1 ) = -ONE
337 END IF
338 ELSE
339 A( I, I ) = ONE
340 IF( MOD( I, 2 ).NE.0 .AND. I.LT.M ) THEN
341 A( I, I+1 ) = IMEPS*2
342 ELSE IF( I.GT.1 ) THEN
343 A( I, I-1 ) = -IMEPS*2
344 END IF
345 END IF
346 240 CONTINUE
347 *
348 DO 250 I = 1, N
349 E( I, I ) = ONE
350 IF( I.LE.4 ) THEN
351 B( I, I ) = -ONE
352 IF( I.GT.2 )
353 $ B( I, I ) = ONE - REEPS
354 IF( MOD( I, 2 ).NE.0 .AND. I.LT.N ) THEN
355 B( I, I+1 ) = IMEPS
356 ELSE IF( I.GT.1 ) THEN
357 B( I, I-1 ) = -IMEPS
358 END IF
359 ELSE IF( I.LE.8 ) THEN
360 IF( I.LE.6 ) THEN
361 B( I, I ) = REEPS
362 ELSE
363 B( I, I ) = -REEPS
364 END IF
365 IF( MOD( I, 2 ).NE.0 .AND. I.LT.N ) THEN
366 B( I, I+1 ) = ONE + IMEPS
367 ELSE IF( I.GT.1 ) THEN
368 B( I, I-1 ) = -ONE - IMEPS
369 END IF
370 ELSE
371 B( I, I ) = ONE - REEPS
372 IF( MOD( I, 2 ).NE.0 .AND. I.LT.N ) THEN
373 B( I, I+1 ) = IMEPS*2
374 ELSE IF( I.GT.1 ) THEN
375 B( I, I-1 ) = -IMEPS*2
376 END IF
377 END IF
378 250 CONTINUE
379 END IF
380 *
381 * Compute rhs (C, F)
382 *
383 CALL DGEMM( 'N', 'N', M, N, M, ONE, A, LDA, R, LDR, ZERO, C, LDC )
384 CALL DGEMM( 'N', 'N', M, N, N, -ONE, L, LDL, B, LDB, ONE, C, LDC )
385 CALL DGEMM( 'N', 'N', M, N, M, ONE, D, LDD, R, LDR, ZERO, F, LDF )
386 CALL DGEMM( 'N', 'N', M, N, N, -ONE, L, LDL, E, LDE, ONE, F, LDF )
387 *
388 * End of DLATM5
389 *
390 END
2 $ E, LDE, F, LDF, R, LDR, L, LDL, ALPHA, QBLCKA,
3 $ QBLCKB )
4 *
5 * -- LAPACK test routine (version 3.1) --
6 * Univ. of Tennessee, Univ. of California Berkeley and NAG Ltd..
7 * November 2006
8 *
9 * .. Scalar Arguments ..
10 INTEGER LDA, LDB, LDC, LDD, LDE, LDF, LDL, LDR, M, N,
11 $ PRTYPE, QBLCKA, QBLCKB
12 DOUBLE PRECISION ALPHA
13 * ..
14 * .. Array Arguments ..
15 DOUBLE PRECISION A( LDA, * ), B( LDB, * ), C( LDC, * ),
16 $ D( LDD, * ), E( LDE, * ), F( LDF, * ),
17 $ L( LDL, * ), R( LDR, * )
18 * ..
19 *
20 * Purpose
21 * =======
22 *
23 * DLATM5 generates matrices involved in the Generalized Sylvester
24 * equation:
25 *
26 * A * R - L * B = C
27 * D * R - L * E = F
28 *
29 * They also satisfy (the diagonalization condition)
30 *
31 * [ I -L ] ( [ A -C ], [ D -F ] ) [ I R ] = ( [ A ], [ D ] )
32 * [ I ] ( [ B ] [ E ] ) [ I ] ( [ B ] [ E ] )
33 *
34 *
35 * Arguments
36 * =========
37 *
38 * PRTYPE (input) INTEGER
39 * "Points" to a certian type of the matrices to generate
40 * (see futher details).
41 *
42 * M (input) INTEGER
43 * Specifies the order of A and D and the number of rows in
44 * C, F, R and L.
45 *
46 * N (input) INTEGER
47 * Specifies the order of B and E and the number of columns in
48 * C, F, R and L.
49 *
50 * A (output) DOUBLE PRECISION array, dimension (LDA, M).
51 * On exit A M-by-M is initialized according to PRTYPE.
52 *
53 * LDA (input) INTEGER
54 * The leading dimension of A.
55 *
56 * B (output) DOUBLE PRECISION array, dimension (LDB, N).
57 * On exit B N-by-N is initialized according to PRTYPE.
58 *
59 * LDB (input) INTEGER
60 * The leading dimension of B.
61 *
62 * C (output) DOUBLE PRECISION array, dimension (LDC, N).
63 * On exit C M-by-N is initialized according to PRTYPE.
64 *
65 * LDC (input) INTEGER
66 * The leading dimension of C.
67 *
68 * D (output) DOUBLE PRECISION array, dimension (LDD, M).
69 * On exit D M-by-M is initialized according to PRTYPE.
70 *
71 * LDD (input) INTEGER
72 * The leading dimension of D.
73 *
74 * E (output) DOUBLE PRECISION array, dimension (LDE, N).
75 * On exit E N-by-N is initialized according to PRTYPE.
76 *
77 * LDE (input) INTEGER
78 * The leading dimension of E.
79 *
80 * F (output) DOUBLE PRECISION array, dimension (LDF, N).
81 * On exit F M-by-N is initialized according to PRTYPE.
82 *
83 * LDF (input) INTEGER
84 * The leading dimension of F.
85 *
86 * R (output) DOUBLE PRECISION array, dimension (LDR, N).
87 * On exit R M-by-N is initialized according to PRTYPE.
88 *
89 * LDR (input) INTEGER
90 * The leading dimension of R.
91 *
92 * L (output) DOUBLE PRECISION array, dimension (LDL, N).
93 * On exit L M-by-N is initialized according to PRTYPE.
94 *
95 * LDL (input) INTEGER
96 * The leading dimension of L.
97 *
98 * ALPHA (input) DOUBLE PRECISION
99 * Parameter used in generating PRTYPE = 1 and 5 matrices.
100 *
101 * QBLCKA (input) INTEGER
102 * When PRTYPE = 3, specifies the distance between 2-by-2
103 * blocks on the diagonal in A. Otherwise, QBLCKA is not
104 * referenced. QBLCKA > 1.
105 *
106 * QBLCKB (input) INTEGER
107 * When PRTYPE = 3, specifies the distance between 2-by-2
108 * blocks on the diagonal in B. Otherwise, QBLCKB is not
109 * referenced. QBLCKB > 1.
110 *
111 *
112 * Further Details
113 * ===============
114 *
115 * PRTYPE = 1: A and B are Jordan blocks, D and E are identity matrices
116 *
117 * A : if (i == j) then A(i, j) = 1.0
118 * if (j == i + 1) then A(i, j) = -1.0
119 * else A(i, j) = 0.0, i, j = 1...M
120 *
121 * B : if (i == j) then B(i, j) = 1.0 - ALPHA
122 * if (j == i + 1) then B(i, j) = 1.0
123 * else B(i, j) = 0.0, i, j = 1...N
124 *
125 * D : if (i == j) then D(i, j) = 1.0
126 * else D(i, j) = 0.0, i, j = 1...M
127 *
128 * E : if (i == j) then E(i, j) = 1.0
129 * else E(i, j) = 0.0, i, j = 1...N
130 *
131 * L = R are chosen from [-10...10],
132 * which specifies the right hand sides (C, F).
133 *
134 * PRTYPE = 2 or 3: Triangular and/or quasi- triangular.
135 *
136 * A : if (i <= j) then A(i, j) = [-1...1]
137 * else A(i, j) = 0.0, i, j = 1...M
138 *
139 * if (PRTYPE = 3) then
140 * A(k + 1, k + 1) = A(k, k)
141 * A(k + 1, k) = [-1...1]
142 * sign(A(k, k + 1) = -(sin(A(k + 1, k))
143 * k = 1, M - 1, QBLCKA
144 *
145 * B : if (i <= j) then B(i, j) = [-1...1]
146 * else B(i, j) = 0.0, i, j = 1...N
147 *
148 * if (PRTYPE = 3) then
149 * B(k + 1, k + 1) = B(k, k)
150 * B(k + 1, k) = [-1...1]
151 * sign(B(k, k + 1) = -(sign(B(k + 1, k))
152 * k = 1, N - 1, QBLCKB
153 *
154 * D : if (i <= j) then D(i, j) = [-1...1].
155 * else D(i, j) = 0.0, i, j = 1...M
156 *
157 *
158 * E : if (i <= j) then D(i, j) = [-1...1]
159 * else E(i, j) = 0.0, i, j = 1...N
160 *
161 * L, R are chosen from [-10...10],
162 * which specifies the right hand sides (C, F).
163 *
164 * PRTYPE = 4 Full
165 * A(i, j) = [-10...10]
166 * D(i, j) = [-1...1] i,j = 1...M
167 * B(i, j) = [-10...10]
168 * E(i, j) = [-1...1] i,j = 1...N
169 * R(i, j) = [-10...10]
170 * L(i, j) = [-1...1] i = 1..M ,j = 1...N
171 *
172 * L, R specifies the right hand sides (C, F).
173 *
174 * PRTYPE = 5 special case common and/or close eigs.
175 *
176 * =====================================================================
177 *
178 * .. Parameters ..
179 DOUBLE PRECISION ONE, ZERO, TWENTY, HALF, TWO
180 PARAMETER ( ONE = 1.0D+0, ZERO = 0.0D+0, TWENTY = 2.0D+1,
181 $ HALF = 0.5D+0, TWO = 2.0D+0 )
182 * ..
183 * .. Local Scalars ..
184 INTEGER I, J, K
185 DOUBLE PRECISION IMEPS, REEPS
186 * ..
187 * .. Intrinsic Functions ..
188 INTRINSIC DBLE, MOD, SIN
189 * ..
190 * .. External Subroutines ..
191 EXTERNAL DGEMM
192 * ..
193 * .. Executable Statements ..
194 *
195 IF( PRTYPE.EQ.1 ) THEN
196 DO 20 I = 1, M
197 DO 10 J = 1, M
198 IF( I.EQ.J ) THEN
199 A( I, J ) = ONE
200 D( I, J ) = ONE
201 ELSE IF( I.EQ.J-1 ) THEN
202 A( I, J ) = -ONE
203 D( I, J ) = ZERO
204 ELSE
205 A( I, J ) = ZERO
206 D( I, J ) = ZERO
207 END IF
208 10 CONTINUE
209 20 CONTINUE
210 *
211 DO 40 I = 1, N
212 DO 30 J = 1, N
213 IF( I.EQ.J ) THEN
214 B( I, J ) = ONE - ALPHA
215 E( I, J ) = ONE
216 ELSE IF( I.EQ.J-1 ) THEN
217 B( I, J ) = ONE
218 E( I, J ) = ZERO
219 ELSE
220 B( I, J ) = ZERO
221 E( I, J ) = ZERO
222 END IF
223 30 CONTINUE
224 40 CONTINUE
225 *
226 DO 60 I = 1, M
227 DO 50 J = 1, N
228 R( I, J ) = ( HALF-SIN( DBLE( I / J ) ) )*TWENTY
229 L( I, J ) = R( I, J )
230 50 CONTINUE
231 60 CONTINUE
232 *
233 ELSE IF( PRTYPE.EQ.2 .OR. PRTYPE.EQ.3 ) THEN
234 DO 80 I = 1, M
235 DO 70 J = 1, M
236 IF( I.LE.J ) THEN
237 A( I, J ) = ( HALF-SIN( DBLE( I ) ) )*TWO
238 D( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWO
239 ELSE
240 A( I, J ) = ZERO
241 D( I, J ) = ZERO
242 END IF
243 70 CONTINUE
244 80 CONTINUE
245 *
246 DO 100 I = 1, N
247 DO 90 J = 1, N
248 IF( I.LE.J ) THEN
249 B( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWO
250 E( I, J ) = ( HALF-SIN( DBLE( J ) ) )*TWO
251 ELSE
252 B( I, J ) = ZERO
253 E( I, J ) = ZERO
254 END IF
255 90 CONTINUE
256 100 CONTINUE
257 *
258 DO 120 I = 1, M
259 DO 110 J = 1, N
260 R( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWENTY
261 L( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWENTY
262 110 CONTINUE
263 120 CONTINUE
264 *
265 IF( PRTYPE.EQ.3 ) THEN
266 IF( QBLCKA.LE.1 )
267 $ QBLCKA = 2
268 DO 130 K = 1, M - 1, QBLCKA
269 A( K+1, K+1 ) = A( K, K )
270 A( K+1, K ) = -SIN( A( K, K+1 ) )
271 130 CONTINUE
272 *
273 IF( QBLCKB.LE.1 )
274 $ QBLCKB = 2
275 DO 140 K = 1, N - 1, QBLCKB
276 B( K+1, K+1 ) = B( K, K )
277 B( K+1, K ) = -SIN( B( K, K+1 ) )
278 140 CONTINUE
279 END IF
280 *
281 ELSE IF( PRTYPE.EQ.4 ) THEN
282 DO 160 I = 1, M
283 DO 150 J = 1, M
284 A( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWENTY
285 D( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWO
286 150 CONTINUE
287 160 CONTINUE
288 *
289 DO 180 I = 1, N
290 DO 170 J = 1, N
291 B( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*TWENTY
292 E( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWO
293 170 CONTINUE
294 180 CONTINUE
295 *
296 DO 200 I = 1, M
297 DO 190 J = 1, N
298 R( I, J ) = ( HALF-SIN( DBLE( J / I ) ) )*TWENTY
299 L( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*TWO
300 190 CONTINUE
301 200 CONTINUE
302 *
303 ELSE IF( PRTYPE.GE.5 ) THEN
304 REEPS = HALF*TWO*TWENTY / ALPHA
305 IMEPS = ( HALF-TWO ) / ALPHA
306 DO 220 I = 1, M
307 DO 210 J = 1, N
308 R( I, J ) = ( HALF-SIN( DBLE( I*J ) ) )*ALPHA / TWENTY
309 L( I, J ) = ( HALF-SIN( DBLE( I+J ) ) )*ALPHA / TWENTY
310 210 CONTINUE
311 220 CONTINUE
312 *
313 DO 230 I = 1, M
314 D( I, I ) = ONE
315 230 CONTINUE
316 *
317 DO 240 I = 1, M
318 IF( I.LE.4 ) THEN
319 A( I, I ) = ONE
320 IF( I.GT.2 )
321 $ A( I, I ) = ONE + REEPS
322 IF( MOD( I, 2 ).NE.0 .AND. I.LT.M ) THEN
323 A( I, I+1 ) = IMEPS
324 ELSE IF( I.GT.1 ) THEN
325 A( I, I-1 ) = -IMEPS
326 END IF
327 ELSE IF( I.LE.8 ) THEN
328 IF( I.LE.6 ) THEN
329 A( I, I ) = REEPS
330 ELSE
331 A( I, I ) = -REEPS
332 END IF
333 IF( MOD( I, 2 ).NE.0 .AND. I.LT.M ) THEN
334 A( I, I+1 ) = ONE
335 ELSE IF( I.GT.1 ) THEN
336 A( I, I-1 ) = -ONE
337 END IF
338 ELSE
339 A( I, I ) = ONE
340 IF( MOD( I, 2 ).NE.0 .AND. I.LT.M ) THEN
341 A( I, I+1 ) = IMEPS*2
342 ELSE IF( I.GT.1 ) THEN
343 A( I, I-1 ) = -IMEPS*2
344 END IF
345 END IF
346 240 CONTINUE
347 *
348 DO 250 I = 1, N
349 E( I, I ) = ONE
350 IF( I.LE.4 ) THEN
351 B( I, I ) = -ONE
352 IF( I.GT.2 )
353 $ B( I, I ) = ONE - REEPS
354 IF( MOD( I, 2 ).NE.0 .AND. I.LT.N ) THEN
355 B( I, I+1 ) = IMEPS
356 ELSE IF( I.GT.1 ) THEN
357 B( I, I-1 ) = -IMEPS
358 END IF
359 ELSE IF( I.LE.8 ) THEN
360 IF( I.LE.6 ) THEN
361 B( I, I ) = REEPS
362 ELSE
363 B( I, I ) = -REEPS
364 END IF
365 IF( MOD( I, 2 ).NE.0 .AND. I.LT.N ) THEN
366 B( I, I+1 ) = ONE + IMEPS
367 ELSE IF( I.GT.1 ) THEN
368 B( I, I-1 ) = -ONE - IMEPS
369 END IF
370 ELSE
371 B( I, I ) = ONE - REEPS
372 IF( MOD( I, 2 ).NE.0 .AND. I.LT.N ) THEN
373 B( I, I+1 ) = IMEPS*2
374 ELSE IF( I.GT.1 ) THEN
375 B( I, I-1 ) = -IMEPS*2
376 END IF
377 END IF
378 250 CONTINUE
379 END IF
380 *
381 * Compute rhs (C, F)
382 *
383 CALL DGEMM( 'N', 'N', M, N, M, ONE, A, LDA, R, LDR, ZERO, C, LDC )
384 CALL DGEMM( 'N', 'N', M, N, N, -ONE, L, LDL, B, LDB, ONE, C, LDC )
385 CALL DGEMM( 'N', 'N', M, N, M, ONE, D, LDD, R, LDR, ZERO, F, LDF )
386 CALL DGEMM( 'N', 'N', M, N, N, -ONE, L, LDL, E, LDE, ONE, F, LDF )
387 *
388 * End of DLATM5
389 *
390 END