1*> \brief \b DLAMSWLQ 2* 3* Definition: 4* =========== 5* 6* SUBROUTINE DLAMSWLQ( SIDE, TRANS, M, N, K, MB, NB, A, LDA, T, 7* $ LDT, C, LDC, WORK, LWORK, INFO ) 8* 9* 10* .. Scalar Arguments .. 11* CHARACTER SIDE, TRANS 12* INTEGER INFO, LDA, M, N, K, MB, NB, LDT, LWORK, LDC 13* .. 14* .. Array Arguments .. 15* DOUBLE A( LDA, * ), WORK( * ), C(LDC, * ), 16* $ T( LDT, * ) 17*> \par Purpose: 18* ============= 19*> 20*> \verbatim 21*> 22*> DLAMQRTS overwrites the general real M-by-N matrix C with 23*> 24*> 25*> SIDE = 'L' SIDE = 'R' 26*> TRANS = 'N': Q * C C * Q 27*> TRANS = 'T': Q**T * C C * Q**T 28*> where Q is a real orthogonal matrix defined as the product of blocked 29*> elementary reflectors computed by short wide LQ 30*> factorization (DLASWLQ) 31*> \endverbatim 32* 33* Arguments: 34* ========== 35* 36*> \param[in] SIDE 37*> \verbatim 38*> SIDE is CHARACTER*1 39*> = 'L': apply Q or Q**T from the Left; 40*> = 'R': apply Q or Q**T from the Right. 41*> \endverbatim 42*> 43*> \param[in] TRANS 44*> \verbatim 45*> TRANS is CHARACTER*1 46*> = 'N': No transpose, apply Q; 47*> = 'T': Transpose, apply Q**T. 48*> \endverbatim 49*> 50*> \param[in] M 51*> \verbatim 52*> M is INTEGER 53*> The number of rows of the matrix C. M >=0. 54*> \endverbatim 55*> 56*> \param[in] N 57*> \verbatim 58*> N is INTEGER 59*> The number of columns of the matrix C. N >= M. 60*> \endverbatim 61*> 62*> \param[in] K 63*> \verbatim 64*> K is INTEGER 65*> The number of elementary reflectors whose product defines 66*> the matrix Q. 67*> M >= K >= 0; 68*> 69*> \endverbatim 70*> \param[in] MB 71*> \verbatim 72*> MB is INTEGER 73*> The row block size to be used in the blocked QR. 74*> M >= MB >= 1 75*> \endverbatim 76*> 77*> \param[in] NB 78*> \verbatim 79*> NB is INTEGER 80*> The column block size to be used in the blocked QR. 81*> NB > M. 82*> \endverbatim 83*> 84*> \param[in] A 85*> \verbatim 86*> A is DOUBLE PRECISION array, dimension 87*> (LDA,M) if SIDE = 'L', 88*> (LDA,N) if SIDE = 'R' 89*> The i-th row must contain the vector which defines the blocked 90*> elementary reflector H(i), for i = 1,2,...,k, as returned by 91*> DLASWLQ in the first k rows of its array argument A. 92*> \endverbatim 93*> 94*> \param[in] LDA 95*> \verbatim 96*> LDA is INTEGER 97*> The leading dimension of the array A. 98*> If SIDE = 'L', LDA >= max(1,M); 99*> if SIDE = 'R', LDA >= max(1,N). 100*> \endverbatim 101*> 102*> \param[in] T 103*> \verbatim 104*> T is DOUBLE PRECISION array, dimension 105*> ( M * Number of blocks(CEIL(N-K/NB-K)), 106*> The blocked upper triangular block reflectors stored in compact form 107*> as a sequence of upper triangular blocks. See below 108*> for further details. 109*> \endverbatim 110*> 111*> \param[in] LDT 112*> \verbatim 113*> LDT is INTEGER 114*> The leading dimension of the array T. LDT >= MB. 115*> \endverbatim 116*> 117*> \param[in,out] C 118*> \verbatim 119*> C is DOUBLE PRECISION array, dimension (LDC,N) 120*> On entry, the M-by-N matrix C. 121*> On exit, C is overwritten by Q*C or Q**T*C or C*Q**T or C*Q. 122*> \endverbatim 123*> 124*> \param[in] LDC 125*> \verbatim 126*> LDC is INTEGER 127*> The leading dimension of the array C. LDC >= max(1,M). 128*> \endverbatim 129*> 130*> \param[out] WORK 131*> \verbatim 132*> (workspace) DOUBLE PRECISION array, dimension (MAX(1,LWORK)) 133*> \endverbatim 134*> 135*> \param[in] LWORK 136*> \verbatim 137*> LWORK is INTEGER 138*> The dimension of the array WORK. 139*> If SIDE = 'L', LWORK >= max(1,NB) * MB; 140*> if SIDE = 'R', LWORK >= max(1,M) * MB. 141*> If LWORK = -1, then a workspace query is assumed; the routine 142*> only calculates the optimal size of the WORK array, returns 143*> this value as the first entry of the WORK array, and no error 144*> message related to LWORK is issued by XERBLA. 145*> \endverbatim 146*> 147*> \param[out] INFO 148*> \verbatim 149*> INFO is INTEGER 150*> = 0: successful exit 151*> < 0: if INFO = -i, the i-th argument had an illegal value 152*> \endverbatim 153* 154* Authors: 155* ======== 156* 157*> \author Univ. of Tennessee 158*> \author Univ. of California Berkeley 159*> \author Univ. of Colorado Denver 160*> \author NAG Ltd. 161* 162*> \par Further Details: 163* ===================== 164*> 165*> \verbatim 166*> Short-Wide LQ (SWLQ) performs LQ by a sequence of orthogonal transformations, 167*> representing Q as a product of other orthogonal matrices 168*> Q = Q(1) * Q(2) * . . . * Q(k) 169*> where each Q(i) zeros out upper diagonal entries of a block of NB rows of A: 170*> Q(1) zeros out the upper diagonal entries of rows 1:NB of A 171*> Q(2) zeros out the bottom MB-N rows of rows [1:M,NB+1:2*NB-M] of A 172*> Q(3) zeros out the bottom MB-N rows of rows [1:M,2*NB-M+1:3*NB-2*M] of A 173*> . . . 174*> 175*> Q(1) is computed by GELQT, which represents Q(1) by Householder vectors 176*> stored under the diagonal of rows 1:MB of A, and by upper triangular 177*> block reflectors, stored in array T(1:LDT,1:N). 178*> For more information see Further Details in GELQT. 179*> 180*> Q(i) for i>1 is computed by TPLQT, which represents Q(i) by Householder vectors 181*> stored in columns [(i-1)*(NB-M)+M+1:i*(NB-M)+M] of A, and by upper triangular 182*> block reflectors, stored in array T(1:LDT,(i-1)*M+1:i*M). 183*> The last Q(k) may use fewer rows. 184*> For more information see Further Details in TPQRT. 185*> 186*> For more details of the overall algorithm, see the description of 187*> Sequential TSQR in Section 2.2 of [1]. 188*> 189*> [1] “Communication-Optimal Parallel and Sequential QR and LU Factorizations,” 190*> J. Demmel, L. Grigori, M. Hoemmen, J. Langou, 191*> SIAM J. Sci. Comput, vol. 34, no. 1, 2012 192*> \endverbatim 193*> 194* ===================================================================== 195 SUBROUTINE DLAMSWLQ( SIDE, TRANS, M, N, K, MB, NB, A, LDA, T, 196 $ LDT, C, LDC, WORK, LWORK, INFO ) 197* 198* -- LAPACK computational routine -- 199* -- LAPACK is a software package provided by Univ. of Tennessee, -- 200* -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..-- 201* 202* .. Scalar Arguments .. 203 CHARACTER SIDE, TRANS 204 INTEGER INFO, LDA, M, N, K, MB, NB, LDT, LWORK, LDC 205* .. 206* .. Array Arguments .. 207 DOUBLE PRECISION A( LDA, * ), WORK( * ), C(LDC, * ), 208 $ T( LDT, * ) 209* .. 210* 211* ===================================================================== 212* 213* .. 214* .. Local Scalars .. 215 LOGICAL LEFT, RIGHT, TRAN, NOTRAN, LQUERY 216 INTEGER I, II, KK, CTR, LW 217* .. 218* .. External Functions .. 219 LOGICAL LSAME 220 EXTERNAL LSAME 221* .. External Subroutines .. 222 EXTERNAL DTPMLQT, DGEMLQT, XERBLA 223* .. 224* .. Executable Statements .. 225* 226* Test the input arguments 227* 228 LQUERY = LWORK.LT.0 229 NOTRAN = LSAME( TRANS, 'N' ) 230 TRAN = LSAME( TRANS, 'T' ) 231 LEFT = LSAME( SIDE, 'L' ) 232 RIGHT = LSAME( SIDE, 'R' ) 233 IF (LEFT) THEN 234 LW = N * MB 235 ELSE 236 LW = M * MB 237 END IF 238* 239 INFO = 0 240 IF( .NOT.LEFT .AND. .NOT.RIGHT ) THEN 241 INFO = -1 242 ELSE IF( .NOT.TRAN .AND. .NOT.NOTRAN ) THEN 243 INFO = -2 244 ELSE IF( M.LT.0 ) THEN 245 INFO = -3 246 ELSE IF( N.LT.0 ) THEN 247 INFO = -4 248 ELSE IF( K.LT.0 ) THEN 249 INFO = -5 250 ELSE IF( LDA.LT.MAX( 1, K ) ) THEN 251 INFO = -9 252 ELSE IF( LDT.LT.MAX( 1, MB) ) THEN 253 INFO = -11 254 ELSE IF( LDC.LT.MAX( 1, M ) ) THEN 255 INFO = -13 256 ELSE IF(( LWORK.LT.MAX(1,LW)).AND.(.NOT.LQUERY)) THEN 257 INFO = -15 258 END IF 259* 260 IF( INFO.NE.0 ) THEN 261 CALL XERBLA( 'DLAMSWLQ', -INFO ) 262 WORK(1) = LW 263 RETURN 264 ELSE IF (LQUERY) THEN 265 WORK(1) = LW 266 RETURN 267 END IF 268* 269* Quick return if possible 270* 271 IF( MIN(M,N,K).EQ.0 ) THEN 272 RETURN 273 END IF 274* 275 IF((NB.LE.K).OR.(NB.GE.MAX(M,N,K))) THEN 276 CALL DGEMLQT( SIDE, TRANS, M, N, K, MB, A, LDA, 277 $ T, LDT, C, LDC, WORK, INFO) 278 RETURN 279 END IF 280* 281 IF(LEFT.AND.TRAN) THEN 282* 283* Multiply Q to the last block of C 284* 285 KK = MOD((M-K),(NB-K)) 286 CTR = (M-K)/(NB-K) 287 IF (KK.GT.0) THEN 288 II=M-KK+1 289 CALL DTPMLQT('L','T',KK , N, K, 0, MB, A(1,II), LDA, 290 $ T(1,CTR*K+1), LDT, C(1,1), LDC, 291 $ C(II,1), LDC, WORK, INFO ) 292 ELSE 293 II=M+1 294 END IF 295* 296 DO I=II-(NB-K),NB+1,-(NB-K) 297* 298* Multiply Q to the current block of C (1:M,I:I+NB) 299* 300 CTR = CTR - 1 301 CALL DTPMLQT('L','T',NB-K , N, K, 0,MB, A(1,I), LDA, 302 $ T(1, CTR*K+1),LDT, C(1,1), LDC, 303 $ C(I,1), LDC, WORK, INFO ) 304 305 END DO 306* 307* Multiply Q to the first block of C (1:M,1:NB) 308* 309 CALL DGEMLQT('L','T',NB , N, K, MB, A(1,1), LDA, T 310 $ ,LDT ,C(1,1), LDC, WORK, INFO ) 311* 312 ELSE IF (LEFT.AND.NOTRAN) THEN 313* 314* Multiply Q to the first block of C 315* 316 KK = MOD((M-K),(NB-K)) 317 II=M-KK+1 318 CTR = 1 319 CALL DGEMLQT('L','N',NB , N, K, MB, A(1,1), LDA, T 320 $ ,LDT ,C(1,1), LDC, WORK, INFO ) 321* 322 DO I=NB+1,II-NB+K,(NB-K) 323* 324* Multiply Q to the current block of C (I:I+NB,1:N) 325* 326 CALL DTPMLQT('L','N',NB-K , N, K, 0,MB, A(1,I), LDA, 327 $ T(1,CTR*K+1), LDT, C(1,1), LDC, 328 $ C(I,1), LDC, WORK, INFO ) 329 CTR = CTR + 1 330* 331 END DO 332 IF(II.LE.M) THEN 333* 334* Multiply Q to the last block of C 335* 336 CALL DTPMLQT('L','N',KK , N, K, 0, MB, A(1,II), LDA, 337 $ T(1,CTR*K+1), LDT, C(1,1), LDC, 338 $ C(II,1), LDC, WORK, INFO ) 339* 340 END IF 341* 342 ELSE IF(RIGHT.AND.NOTRAN) THEN 343* 344* Multiply Q to the last block of C 345* 346 KK = MOD((N-K),(NB-K)) 347 CTR = (N-K)/(NB-K) 348 IF (KK.GT.0) THEN 349 II=N-KK+1 350 CALL DTPMLQT('R','N',M , KK, K, 0, MB, A(1, II), LDA, 351 $ T(1,CTR *K+1), LDT, C(1,1), LDC, 352 $ C(1,II), LDC, WORK, INFO ) 353 ELSE 354 II=N+1 355 END IF 356* 357 DO I=II-(NB-K),NB+1,-(NB-K) 358* 359* Multiply Q to the current block of C (1:M,I:I+MB) 360* 361 CTR = CTR - 1 362 CALL DTPMLQT('R','N', M, NB-K, K, 0, MB, A(1, I), LDA, 363 $ T(1,CTR*K+1), LDT, C(1,1), LDC, 364 $ C(1,I), LDC, WORK, INFO ) 365* 366 END DO 367* 368* Multiply Q to the first block of C (1:M,1:MB) 369* 370 CALL DGEMLQT('R','N',M , NB, K, MB, A(1,1), LDA, T 371 $ ,LDT ,C(1,1), LDC, WORK, INFO ) 372* 373 ELSE IF (RIGHT.AND.TRAN) THEN 374* 375* Multiply Q to the first block of C 376* 377 KK = MOD((N-K),(NB-K)) 378 CTR = 1 379 II=N-KK+1 380 CALL DGEMLQT('R','T',M , NB, K, MB, A(1,1), LDA, T 381 $ ,LDT ,C(1,1), LDC, WORK, INFO ) 382* 383 DO I=NB+1,II-NB+K,(NB-K) 384* 385* Multiply Q to the current block of C (1:M,I:I+MB) 386* 387 CALL DTPMLQT('R','T',M , NB-K, K, 0,MB, A(1,I), LDA, 388 $ T(1,CTR*K+1), LDT, C(1,1), LDC, 389 $ C(1,I), LDC, WORK, INFO ) 390 CTR = CTR + 1 391* 392 END DO 393 IF(II.LE.N) THEN 394* 395* Multiply Q to the last block of C 396* 397 CALL DTPMLQT('R','T',M , KK, K, 0,MB, A(1,II), LDA, 398 $ T(1,CTR*K+1),LDT, C(1,1), LDC, 399 $ C(1,II), LDC, WORK, INFO ) 400* 401 END IF 402* 403 END IF 404* 405 WORK(1) = LW 406 RETURN 407* 408* End of DLAMSWLQ 409* 410 END 411