1*> \brief \b SLASD3 finds all square roots of the roots of the secular equation, as defined by the values in D and Z, and then updates the singular vectors by matrix multiplication. Used by sbdsdc. 2* 3* =========== DOCUMENTATION =========== 4* 5* Online html documentation available at 6* http://www.netlib.org/lapack/explore-html/ 7* 8*> \htmlonly 9*> Download SLASD3 + dependencies 10*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/slasd3.f"> 11*> [TGZ]</a> 12*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/slasd3.f"> 13*> [ZIP]</a> 14*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/slasd3.f"> 15*> [TXT]</a> 16*> \endhtmlonly 17* 18* Definition: 19* =========== 20* 21* SUBROUTINE SLASD3( NL, NR, SQRE, K, D, Q, LDQ, DSIGMA, U, LDU, U2, 22* LDU2, VT, LDVT, VT2, LDVT2, IDXC, CTOT, Z, 23* INFO ) 24* 25* .. Scalar Arguments .. 26* INTEGER INFO, K, LDQ, LDU, LDU2, LDVT, LDVT2, NL, NR, 27* $ SQRE 28* .. 29* .. Array Arguments .. 30* INTEGER CTOT( * ), IDXC( * ) 31* REAL D( * ), DSIGMA( * ), Q( LDQ, * ), U( LDU, * ), 32* $ U2( LDU2, * ), VT( LDVT, * ), VT2( LDVT2, * ), 33* $ Z( * ) 34* .. 35* 36* 37*> \par Purpose: 38* ============= 39*> 40*> \verbatim 41*> 42*> SLASD3 finds all the square roots of the roots of the secular 43*> equation, as defined by the values in D and Z. It makes the 44*> appropriate calls to SLASD4 and then updates the singular 45*> vectors by matrix multiplication. 46*> 47*> This code makes very mild assumptions about floating point 48*> arithmetic. It will work on machines with a guard digit in 49*> add/subtract, or on those binary machines without guard digits 50*> which subtract like the Cray XMP, Cray YMP, Cray C 90, or Cray 2. 51*> It could conceivably fail on hexadecimal or decimal machines 52*> without guard digits, but we know of none. 53*> 54*> SLASD3 is called from SLASD1. 55*> \endverbatim 56* 57* Arguments: 58* ========== 59* 60*> \param[in] NL 61*> \verbatim 62*> NL is INTEGER 63*> The row dimension of the upper block. NL >= 1. 64*> \endverbatim 65*> 66*> \param[in] NR 67*> \verbatim 68*> NR is INTEGER 69*> The row dimension of the lower block. NR >= 1. 70*> \endverbatim 71*> 72*> \param[in] SQRE 73*> \verbatim 74*> SQRE is INTEGER 75*> = 0: the lower block is an NR-by-NR square matrix. 76*> = 1: the lower block is an NR-by-(NR+1) rectangular matrix. 77*> 78*> The bidiagonal matrix has N = NL + NR + 1 rows and 79*> M = N + SQRE >= N columns. 80*> \endverbatim 81*> 82*> \param[in] K 83*> \verbatim 84*> K is INTEGER 85*> The size of the secular equation, 1 =< K = < N. 86*> \endverbatim 87*> 88*> \param[out] D 89*> \verbatim 90*> D is REAL array, dimension(K) 91*> On exit the square roots of the roots of the secular equation, 92*> in ascending order. 93*> \endverbatim 94*> 95*> \param[out] Q 96*> \verbatim 97*> Q is REAL array, 98*> dimension at least (LDQ,K). 99*> \endverbatim 100*> 101*> \param[in] LDQ 102*> \verbatim 103*> LDQ is INTEGER 104*> The leading dimension of the array Q. LDQ >= K. 105*> \endverbatim 106*> 107*> \param[in,out] DSIGMA 108*> \verbatim 109*> DSIGMA is REAL array, dimension(K) 110*> The first K elements of this array contain the old roots 111*> of the deflated updating problem. These are the poles 112*> of the secular equation. 113*> \endverbatim 114*> 115*> \param[out] U 116*> \verbatim 117*> U is REAL array, dimension (LDU, N) 118*> The last N - K columns of this matrix contain the deflated 119*> left singular vectors. 120*> \endverbatim 121*> 122*> \param[in] LDU 123*> \verbatim 124*> LDU is INTEGER 125*> The leading dimension of the array U. LDU >= N. 126*> \endverbatim 127*> 128*> \param[in] U2 129*> \verbatim 130*> U2 is REAL array, dimension (LDU2, N) 131*> The first K columns of this matrix contain the non-deflated 132*> left singular vectors for the split problem. 133*> \endverbatim 134*> 135*> \param[in] LDU2 136*> \verbatim 137*> LDU2 is INTEGER 138*> The leading dimension of the array U2. LDU2 >= N. 139*> \endverbatim 140*> 141*> \param[out] VT 142*> \verbatim 143*> VT is REAL array, dimension (LDVT, M) 144*> The last M - K columns of VT**T contain the deflated 145*> right singular vectors. 146*> \endverbatim 147*> 148*> \param[in] LDVT 149*> \verbatim 150*> LDVT is INTEGER 151*> The leading dimension of the array VT. LDVT >= N. 152*> \endverbatim 153*> 154*> \param[in,out] VT2 155*> \verbatim 156*> VT2 is REAL array, dimension (LDVT2, N) 157*> The first K columns of VT2**T contain the non-deflated 158*> right singular vectors for the split problem. 159*> \endverbatim 160*> 161*> \param[in] LDVT2 162*> \verbatim 163*> LDVT2 is INTEGER 164*> The leading dimension of the array VT2. LDVT2 >= N. 165*> \endverbatim 166*> 167*> \param[in] IDXC 168*> \verbatim 169*> IDXC is INTEGER array, dimension (N) 170*> The permutation used to arrange the columns of U (and rows of 171*> VT) into three groups: the first group contains non-zero 172*> entries only at and above (or before) NL +1; the second 173*> contains non-zero entries only at and below (or after) NL+2; 174*> and the third is dense. The first column of U and the row of 175*> VT are treated separately, however. 176*> 177*> The rows of the singular vectors found by SLASD4 178*> must be likewise permuted before the matrix multiplies can 179*> take place. 180*> \endverbatim 181*> 182*> \param[in] CTOT 183*> \verbatim 184*> CTOT is INTEGER array, dimension (4) 185*> A count of the total number of the various types of columns 186*> in U (or rows in VT), as described in IDXC. The fourth column 187*> type is any column which has been deflated. 188*> \endverbatim 189*> 190*> \param[in,out] Z 191*> \verbatim 192*> Z is REAL array, dimension (K) 193*> The first K elements of this array contain the components 194*> of the deflation-adjusted updating row vector. 195*> \endverbatim 196*> 197*> \param[out] INFO 198*> \verbatim 199*> INFO is INTEGER 200*> = 0: successful exit. 201*> < 0: if INFO = -i, the i-th argument had an illegal value. 202*> > 0: if INFO = 1, a singular value did not converge 203*> \endverbatim 204* 205* Authors: 206* ======== 207* 208*> \author Univ. of Tennessee 209*> \author Univ. of California Berkeley 210*> \author Univ. of Colorado Denver 211*> \author NAG Ltd. 212* 213*> \date November 2015 214* 215*> \ingroup auxOTHERauxiliary 216* 217*> \par Contributors: 218* ================== 219*> 220*> Ming Gu and Huan Ren, Computer Science Division, University of 221*> California at Berkeley, USA 222*> 223* ===================================================================== 224 SUBROUTINE SLASD3( NL, NR, SQRE, K, D, Q, LDQ, DSIGMA, U, LDU, U2, 225 $ LDU2, VT, LDVT, VT2, LDVT2, IDXC, CTOT, Z, 226 $ INFO ) 227* 228* -- LAPACK auxiliary routine (version 3.6.0) -- 229* -- LAPACK is a software package provided by Univ. of Tennessee, -- 230* -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..-- 231* November 2015 232* 233* .. Scalar Arguments .. 234 INTEGER INFO, K, LDQ, LDU, LDU2, LDVT, LDVT2, NL, NR, 235 $ SQRE 236* .. 237* .. Array Arguments .. 238 INTEGER CTOT( * ), IDXC( * ) 239 REAL D( * ), DSIGMA( * ), Q( LDQ, * ), U( LDU, * ), 240 $ U2( LDU2, * ), VT( LDVT, * ), VT2( LDVT2, * ), 241 $ Z( * ) 242* .. 243* 244* ===================================================================== 245* 246* .. Parameters .. 247 REAL ONE, ZERO, NEGONE 248 PARAMETER ( ONE = 1.0E+0, ZERO = 0.0E+0, 249 $ NEGONE = -1.0E+0 ) 250* .. 251* .. Local Scalars .. 252 INTEGER CTEMP, I, J, JC, KTEMP, M, N, NLP1, NLP2, NRP1 253 REAL RHO, TEMP 254* .. 255* .. External Functions .. 256 REAL SLAMC3, SNRM2 257 EXTERNAL SLAMC3, SNRM2 258* .. 259* .. External Subroutines .. 260 EXTERNAL SCOPY, SGEMM, SLACPY, SLASCL, SLASD4, XERBLA 261* .. 262* .. Intrinsic Functions .. 263 INTRINSIC ABS, SIGN, SQRT 264* .. 265* .. Executable Statements .. 266* 267* Test the input parameters. 268* 269 INFO = 0 270* 271 IF( NL.LT.1 ) THEN 272 INFO = -1 273 ELSE IF( NR.LT.1 ) THEN 274 INFO = -2 275 ELSE IF( ( SQRE.NE.1 ) .AND. ( SQRE.NE.0 ) ) THEN 276 INFO = -3 277 END IF 278* 279 N = NL + NR + 1 280 M = N + SQRE 281 NLP1 = NL + 1 282 NLP2 = NL + 2 283* 284 IF( ( K.LT.1 ) .OR. ( K.GT.N ) ) THEN 285 INFO = -4 286 ELSE IF( LDQ.LT.K ) THEN 287 INFO = -7 288 ELSE IF( LDU.LT.N ) THEN 289 INFO = -10 290 ELSE IF( LDU2.LT.N ) THEN 291 INFO = -12 292 ELSE IF( LDVT.LT.M ) THEN 293 INFO = -14 294 ELSE IF( LDVT2.LT.M ) THEN 295 INFO = -16 296 END IF 297 IF( INFO.NE.0 ) THEN 298 CALL XERBLA( 'SLASD3', -INFO ) 299 RETURN 300 END IF 301* 302* Quick return if possible 303* 304 IF( K.EQ.1 ) THEN 305 D( 1 ) = ABS( Z( 1 ) ) 306 CALL SCOPY( M, VT2( 1, 1 ), LDVT2, VT( 1, 1 ), LDVT ) 307 IF( Z( 1 ).GT.ZERO ) THEN 308 CALL SCOPY( N, U2( 1, 1 ), 1, U( 1, 1 ), 1 ) 309 ELSE 310 DO 10 I = 1, N 311 U( I, 1 ) = -U2( I, 1 ) 312 10 CONTINUE 313 END IF 314 RETURN 315 END IF 316* 317* Modify values DSIGMA(i) to make sure all DSIGMA(i)-DSIGMA(j) can 318* be computed with high relative accuracy (barring over/underflow). 319* This is a problem on machines without a guard digit in 320* add/subtract (Cray XMP, Cray YMP, Cray C 90 and Cray 2). 321* The following code replaces DSIGMA(I) by 2*DSIGMA(I)-DSIGMA(I), 322* which on any of these machines zeros out the bottommost 323* bit of DSIGMA(I) if it is 1; this makes the subsequent 324* subtractions DSIGMA(I)-DSIGMA(J) unproblematic when cancellation 325* occurs. On binary machines with a guard digit (almost all 326* machines) it does not change DSIGMA(I) at all. On hexadecimal 327* and decimal machines with a guard digit, it slightly 328* changes the bottommost bits of DSIGMA(I). It does not account 329* for hexadecimal or decimal machines without guard digits 330* (we know of none). We use a subroutine call to compute 331* 2*DSIGMA(I) to prevent optimizing compilers from eliminating 332* this code. 333* 334 DO 20 I = 1, K 335 DSIGMA( I ) = SLAMC3( DSIGMA( I ), DSIGMA( I ) ) - DSIGMA( I ) 336 20 CONTINUE 337* 338* Keep a copy of Z. 339* 340 CALL SCOPY( K, Z, 1, Q, 1 ) 341* 342* Normalize Z. 343* 344 RHO = SNRM2( K, Z, 1 ) 345 CALL SLASCL( 'G', 0, 0, RHO, ONE, K, 1, Z, K, INFO ) 346 RHO = RHO*RHO 347* 348* Find the new singular values. 349* 350 DO 30 J = 1, K 351 CALL SLASD4( K, J, DSIGMA, Z, U( 1, J ), RHO, D( J ), 352 $ VT( 1, J ), INFO ) 353* 354* If the zero finder fails, report the convergence failure. 355* 356 IF( INFO.NE.0 ) THEN 357 RETURN 358 END IF 359 30 CONTINUE 360* 361* Compute updated Z. 362* 363 DO 60 I = 1, K 364 Z( I ) = U( I, K )*VT( I, K ) 365 DO 40 J = 1, I - 1 366 Z( I ) = Z( I )*( U( I, J )*VT( I, J ) / 367 $ ( DSIGMA( I )-DSIGMA( J ) ) / 368 $ ( DSIGMA( I )+DSIGMA( J ) ) ) 369 40 CONTINUE 370 DO 50 J = I, K - 1 371 Z( I ) = Z( I )*( U( I, J )*VT( I, J ) / 372 $ ( DSIGMA( I )-DSIGMA( J+1 ) ) / 373 $ ( DSIGMA( I )+DSIGMA( J+1 ) ) ) 374 50 CONTINUE 375 Z( I ) = SIGN( SQRT( ABS( Z( I ) ) ), Q( I, 1 ) ) 376 60 CONTINUE 377* 378* Compute left singular vectors of the modified diagonal matrix, 379* and store related information for the right singular vectors. 380* 381 DO 90 I = 1, K 382 VT( 1, I ) = Z( 1 ) / U( 1, I ) / VT( 1, I ) 383 U( 1, I ) = NEGONE 384 DO 70 J = 2, K 385 VT( J, I ) = Z( J ) / U( J, I ) / VT( J, I ) 386 U( J, I ) = DSIGMA( J )*VT( J, I ) 387 70 CONTINUE 388 TEMP = SNRM2( K, U( 1, I ), 1 ) 389 Q( 1, I ) = U( 1, I ) / TEMP 390 DO 80 J = 2, K 391 JC = IDXC( J ) 392 Q( J, I ) = U( JC, I ) / TEMP 393 80 CONTINUE 394 90 CONTINUE 395* 396* Update the left singular vector matrix. 397* 398 IF( K.EQ.2 ) THEN 399 CALL SGEMM( 'N', 'N', N, K, K, ONE, U2, LDU2, Q, LDQ, ZERO, U, 400 $ LDU ) 401 GO TO 100 402 END IF 403 IF( CTOT( 1 ).GT.0 ) THEN 404 CALL SGEMM( 'N', 'N', NL, K, CTOT( 1 ), ONE, U2( 1, 2 ), LDU2, 405 $ Q( 2, 1 ), LDQ, ZERO, U( 1, 1 ), LDU ) 406 IF( CTOT( 3 ).GT.0 ) THEN 407 KTEMP = 2 + CTOT( 1 ) + CTOT( 2 ) 408 CALL SGEMM( 'N', 'N', NL, K, CTOT( 3 ), ONE, U2( 1, KTEMP ), 409 $ LDU2, Q( KTEMP, 1 ), LDQ, ONE, U( 1, 1 ), LDU ) 410 END IF 411 ELSE IF( CTOT( 3 ).GT.0 ) THEN 412 KTEMP = 2 + CTOT( 1 ) + CTOT( 2 ) 413 CALL SGEMM( 'N', 'N', NL, K, CTOT( 3 ), ONE, U2( 1, KTEMP ), 414 $ LDU2, Q( KTEMP, 1 ), LDQ, ZERO, U( 1, 1 ), LDU ) 415 ELSE 416 CALL SLACPY( 'F', NL, K, U2, LDU2, U, LDU ) 417 END IF 418 CALL SCOPY( K, Q( 1, 1 ), LDQ, U( NLP1, 1 ), LDU ) 419 KTEMP = 2 + CTOT( 1 ) 420 CTEMP = CTOT( 2 ) + CTOT( 3 ) 421 CALL SGEMM( 'N', 'N', NR, K, CTEMP, ONE, U2( NLP2, KTEMP ), LDU2, 422 $ Q( KTEMP, 1 ), LDQ, ZERO, U( NLP2, 1 ), LDU ) 423* 424* Generate the right singular vectors. 425* 426 100 CONTINUE 427 DO 120 I = 1, K 428 TEMP = SNRM2( K, VT( 1, I ), 1 ) 429 Q( I, 1 ) = VT( 1, I ) / TEMP 430 DO 110 J = 2, K 431 JC = IDXC( J ) 432 Q( I, J ) = VT( JC, I ) / TEMP 433 110 CONTINUE 434 120 CONTINUE 435* 436* Update the right singular vector matrix. 437* 438 IF( K.EQ.2 ) THEN 439 CALL SGEMM( 'N', 'N', K, M, K, ONE, Q, LDQ, VT2, LDVT2, ZERO, 440 $ VT, LDVT ) 441 RETURN 442 END IF 443 KTEMP = 1 + CTOT( 1 ) 444 CALL SGEMM( 'N', 'N', K, NLP1, KTEMP, ONE, Q( 1, 1 ), LDQ, 445 $ VT2( 1, 1 ), LDVT2, ZERO, VT( 1, 1 ), LDVT ) 446 KTEMP = 2 + CTOT( 1 ) + CTOT( 2 ) 447 IF( KTEMP.LE.LDVT2 ) 448 $ CALL SGEMM( 'N', 'N', K, NLP1, CTOT( 3 ), ONE, Q( 1, KTEMP ), 449 $ LDQ, VT2( KTEMP, 1 ), LDVT2, ONE, VT( 1, 1 ), 450 $ LDVT ) 451* 452 KTEMP = CTOT( 1 ) + 1 453 NRP1 = NR + SQRE 454 IF( KTEMP.GT.1 ) THEN 455 DO 130 I = 1, K 456 Q( I, KTEMP ) = Q( I, 1 ) 457 130 CONTINUE 458 DO 140 I = NLP2, M 459 VT2( KTEMP, I ) = VT2( 1, I ) 460 140 CONTINUE 461 END IF 462 CTEMP = 1 + CTOT( 2 ) + CTOT( 3 ) 463 CALL SGEMM( 'N', 'N', K, NRP1, CTEMP, ONE, Q( 1, KTEMP ), LDQ, 464 $ VT2( KTEMP, NLP2 ), LDVT2, ZERO, VT( 1, NLP2 ), LDVT ) 465* 466 RETURN 467* 468* End of SLASD3 469* 470 END 471