| @@ -135,7 +135,7 @@ | |||
| *> \param[in,out] S | |||
| *> \verbatim | |||
| *> S is DOUBLE PRECISION array, dimension (N) | |||
| *> The row scale factors for A. If EQUED = 'Y', A is multiplied on | |||
| *> The scale factors for A. If EQUED = 'Y', A is multiplied on | |||
| *> the left and right by diag(S). S is an input argument if FACT = | |||
| *> 'F'; otherwise, S is an output argument. If FACT = 'F' and EQUED | |||
| *> = 'Y', each element of S must be positive. If S is output, each | |||
| @@ -263,7 +263,7 @@ | |||
| *> information as described below. There currently are up to three | |||
| *> pieces of information returned for each right-hand side. If | |||
| *> componentwise accuracy is not requested (PARAMS(3) = 0.0), then | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS .LT. 3, then at most | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS < 3, then at most | |||
| *> the first (:,N_ERR_BNDS) entries are returned. | |||
| *> | |||
| *> The first index in ERR_BNDS_COMP(i,:) corresponds to the ith | |||
| @@ -299,14 +299,14 @@ | |||
| *> \param[in] NPARAMS | |||
| *> \verbatim | |||
| *> NPARAMS is INTEGER | |||
| *> Specifies the number of parameters set in PARAMS. If .LE. 0, the | |||
| *> Specifies the number of parameters set in PARAMS. If <= 0, the | |||
| *> PARAMS array is never referenced and default values are used. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in,out] PARAMS | |||
| *> \verbatim | |||
| *> PARAMS is DOUBLE PRECISION array, dimension (NPARAMS) | |||
| *> Specifies algorithm parameters. If an entry is .LT. 0.0, then | |||
| *> Specifies algorithm parameters. If an entry is < 0.0, then | |||
| *> that entry will be filled with default value used for that | |||
| *> parameter. Only positions up to NPARAMS are accessed; defaults | |||
| *> are used for higher-numbered parameters. | |||
| @@ -314,9 +314,9 @@ | |||
| *> PARAMS(LA_LINRX_ITREF_I = 1) : Whether to perform iterative | |||
| *> refinement or not. | |||
| *> Default: 1.0D+0 | |||
| *> = 0.0 : No refinement is performed, and no error bounds are | |||
| *> = 0.0: No refinement is performed, and no error bounds are | |||
| *> computed. | |||
| *> = 1.0 : Use the double-precision refinement algorithm, | |||
| *> = 1.0: Use the double-precision refinement algorithm, | |||
| *> possibly with doubled-single computations if the | |||
| *> compilation environment does not support DOUBLE | |||
| *> PRECISION. | |||
| @@ -366,7 +366,7 @@ | |||
| *> information as described below. There currently are up to three | |||
| *> pieces of information returned for each right-hand side. If | |||
| *> componentwise accuracy is not requested (PARAMS(3) = 0.0), then | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS .LT. 3, then at most | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS < 3, then at most | |||
| *> the first (:,N_ERR_BNDS) entries are returned. | |||
| *> | |||
| *> The first index in ERR_BNDS_COMP(i,:) corresponds to the ith | |||
| @@ -402,14 +402,14 @@ | |||
| *> \param[in] NPARAMS | |||
| *> \verbatim | |||
| *> NPARAMS is INTEGER | |||
| *> Specifies the number of parameters set in PARAMS. If .LE. 0, the | |||
| *> Specifies the number of parameters set in PARAMS. If <= 0, the | |||
| *> PARAMS array is never referenced and default values are used. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in,out] PARAMS | |||
| *> \verbatim | |||
| *> PARAMS is DOUBLE PRECISION array, dimension NPARAMS | |||
| *> Specifies algorithm parameters. If an entry is .LT. 0.0, then | |||
| *> Specifies algorithm parameters. If an entry is < 0.0, then | |||
| *> that entry will be filled with default value used for that | |||
| *> parameter. Only positions up to NPARAMS are accessed; defaults | |||
| *> are used for higher-numbered parameters. | |||
| @@ -417,9 +417,9 @@ | |||
| *> PARAMS(LA_LINRX_ITREF_I = 1) : Whether to perform iterative | |||
| *> refinement or not. | |||
| *> Default: 1.0D+0 | |||
| *> = 0.0 : No refinement is performed, and no error bounds are | |||
| *> = 0.0: No refinement is performed, and no error bounds are | |||
| *> computed. | |||
| *> = 1.0 : Use the extra-precise refinement algorithm. | |||
| *> = 1.0: Use the extra-precise refinement algorithm. | |||
| *> (other values are reserved for future use) | |||
| *> | |||
| *> PARAMS(LA_LINRX_ITHRESH_I = 2) : Maximum number of residual | |||
| @@ -1,26 +1,26 @@ | |||
| *> \brief \b DSB2ST_KERNELS | |||
| * | |||
| * @generated from zhb2st_kernels.f, fortran z -> d, Wed Dec 7 08:22:39 2016 | |||
| * | |||
| * | |||
| * =========== DOCUMENTATION =========== | |||
| * | |||
| * Online html documentation available at | |||
| * http://www.netlib.org/lapack/explore-html/ | |||
| * Online html documentation available at | |||
| * http://www.netlib.org/lapack/explore-html/ | |||
| * | |||
| *> \htmlonly | |||
| *> Download DSB2ST_KERNELS + dependencies | |||
| *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dsb2st_kernels.f"> | |||
| *> [TGZ]</a> | |||
| *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dsb2st_kernels.f"> | |||
| *> [ZIP]</a> | |||
| *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dsb2st_kernels.f"> | |||
| *> Download DSB2ST_KERNELS + dependencies | |||
| *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dsb2st_kernels.f"> | |||
| *> [TGZ]</a> | |||
| *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dsb2st_kernels.f"> | |||
| *> [ZIP]</a> | |||
| *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dsb2st_kernels.f"> | |||
| *> [TXT]</a> | |||
| *> \endhtmlonly | |||
| *> \endhtmlonly | |||
| * | |||
| * Definition: | |||
| * =========== | |||
| * | |||
| * SUBROUTINE DSB2ST_KERNELS( UPLO, WANTZ, TTYPE, | |||
| * SUBROUTINE DSB2ST_KERNELS( UPLO, WANTZ, TTYPE, | |||
| * ST, ED, SWEEP, N, NB, IB, | |||
| * A, LDA, V, TAU, LDVT, WORK) | |||
| * | |||
| @@ -32,9 +32,9 @@ | |||
| * INTEGER TTYPE, ST, ED, SWEEP, N, NB, IB, LDA, LDVT | |||
| * .. | |||
| * .. Array Arguments .. | |||
| * DOUBLE PRECISION A( LDA, * ), V( * ), | |||
| * DOUBLE PRECISION A( LDA, * ), V( * ), | |||
| * TAU( * ), WORK( * ) | |||
| * | |||
| * | |||
| *> \par Purpose: | |||
| * ============= | |||
| *> | |||
| @@ -124,7 +124,7 @@ | |||
| *> LDVT is INTEGER. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in] WORK | |||
| *> \param[out] WORK | |||
| *> \verbatim | |||
| *> WORK is DOUBLE PRECISION array. Workspace of size nb. | |||
| *> \endverbatim | |||
| @@ -147,7 +147,7 @@ | |||
| *> http://doi.acm.org/10.1145/2063384.2063394 | |||
| *> | |||
| *> A. Haidar, J. Kurzak, P. Luszczek, 2013. | |||
| *> An improved parallel singular value algorithm and its implementation | |||
| *> An improved parallel singular value algorithm and its implementation | |||
| *> for multicore hardware, In Proceedings of 2013 International Conference | |||
| *> for High Performance Computing, Networking, Storage and Analysis (SC '13). | |||
| *> Denver, Colorado, USA, 2013. | |||
| @@ -155,16 +155,16 @@ | |||
| *> http://doi.acm.org/10.1145/2503210.2503292 | |||
| *> | |||
| *> A. Haidar, R. Solca, S. Tomov, T. Schulthess and J. Dongarra. | |||
| *> A novel hybrid CPU-GPU generalized eigensolver for electronic structure | |||
| *> A novel hybrid CPU-GPU generalized eigensolver for electronic structure | |||
| *> calculations based on fine-grained memory aware tasks. | |||
| *> International Journal of High Performance Computing Applications. | |||
| *> Volume 28 Issue 2, Pages 196-209, May 2014. | |||
| *> http://hpc.sagepub.com/content/28/2/196 | |||
| *> http://hpc.sagepub.com/content/28/2/196 | |||
| *> | |||
| *> \endverbatim | |||
| *> | |||
| * ===================================================================== | |||
| SUBROUTINE DSB2ST_KERNELS( UPLO, WANTZ, TTYPE, | |||
| SUBROUTINE DSB2ST_KERNELS( UPLO, WANTZ, TTYPE, | |||
| $ ST, ED, SWEEP, N, NB, IB, | |||
| $ A, LDA, V, TAU, LDVT, WORK) | |||
| * | |||
| @@ -181,7 +181,7 @@ | |||
| INTEGER TTYPE, ST, ED, SWEEP, N, NB, IB, LDA, LDVT | |||
| * .. | |||
| * .. Array Arguments .. | |||
| DOUBLE PRECISION A( LDA, * ), V( * ), | |||
| DOUBLE PRECISION A( LDA, * ), V( * ), | |||
| $ TAU( * ), WORK( * ) | |||
| * .. | |||
| * | |||
| @@ -195,8 +195,8 @@ | |||
| * .. Local Scalars .. | |||
| LOGICAL UPPER | |||
| INTEGER I, J1, J2, LM, LN, VPOS, TAUPOS, | |||
| $ DPOS, OFDPOS, AJETER | |||
| DOUBLE PRECISION CTMP | |||
| $ DPOS, OFDPOS, AJETER | |||
| DOUBLE PRECISION CTMP | |||
| * .. | |||
| * .. External Subroutines .. | |||
| EXTERNAL DLARFG, DLARFX, DLARFY | |||
| @@ -209,7 +209,7 @@ | |||
| * .. | |||
| * .. | |||
| * .. Executable Statements .. | |||
| * | |||
| * | |||
| AJETER = IB + LDVT | |||
| UPPER = LSAME( UPLO, 'U' ) | |||
| @@ -240,10 +240,10 @@ | |||
| V( VPOS ) = ONE | |||
| DO 10 I = 1, LM-1 | |||
| V( VPOS+I ) = ( A( OFDPOS-I, ST+I ) ) | |||
| A( OFDPOS-I, ST+I ) = ZERO | |||
| A( OFDPOS-I, ST+I ) = ZERO | |||
| 10 CONTINUE | |||
| CTMP = ( A( OFDPOS, ST ) ) | |||
| CALL DLARFG( LM, CTMP, V( VPOS+1 ), 1, | |||
| CALL DLARFG( LM, CTMP, V( VPOS+1 ), 1, | |||
| $ TAU( TAUPOS ) ) | |||
| A( OFDPOS, ST ) = CTMP | |||
| * | |||
| @@ -281,14 +281,14 @@ | |||
| * | |||
| V( VPOS ) = ONE | |||
| DO 30 I = 1, LM-1 | |||
| V( VPOS+I ) = | |||
| V( VPOS+I ) = | |||
| $ ( A( DPOS-NB-I, J1+I ) ) | |||
| A( DPOS-NB-I, J1+I ) = ZERO | |||
| 30 CONTINUE | |||
| CTMP = ( A( DPOS-NB, J1 ) ) | |||
| CALL DLARFG( LM, CTMP, V( VPOS+1 ), 1, TAU( TAUPOS ) ) | |||
| A( DPOS-NB, J1 ) = CTMP | |||
| * | |||
| * | |||
| CALL DLARFX( 'Right', LN-1, LM, V( VPOS ), | |||
| $ TAU( TAUPOS ), | |||
| $ A( DPOS-NB+1, J1 ), LDA-1, WORK) | |||
| @@ -296,9 +296,9 @@ | |||
| ENDIF | |||
| * | |||
| * Lower case | |||
| * | |||
| * | |||
| ELSE | |||
| * | |||
| * | |||
| IF( WANTZ ) THEN | |||
| VPOS = MOD( SWEEP-1, 2 ) * N + ST | |||
| TAUPOS = MOD( SWEEP-1, 2 ) * N + ST | |||
| @@ -313,9 +313,9 @@ | |||
| V( VPOS ) = ONE | |||
| DO 20 I = 1, LM-1 | |||
| V( VPOS+I ) = A( OFDPOS+I, ST-1 ) | |||
| A( OFDPOS+I, ST-1 ) = ZERO | |||
| A( OFDPOS+I, ST-1 ) = ZERO | |||
| 20 CONTINUE | |||
| CALL DLARFG( LM, A( OFDPOS, ST-1 ), V( VPOS+1 ), 1, | |||
| CALL DLARFG( LM, A( OFDPOS, ST-1 ), V( VPOS+1 ), 1, | |||
| $ TAU( TAUPOS ) ) | |||
| * | |||
| LM = ED - ST + 1 | |||
| @@ -342,7 +342,7 @@ | |||
| LM = J2-J1+1 | |||
| * | |||
| IF( LM.GT.0) THEN | |||
| CALL DLARFX( 'Right', LM, LN, V( VPOS ), | |||
| CALL DLARFX( 'Right', LM, LN, V( VPOS ), | |||
| $ TAU( TAUPOS ), A( DPOS+NB, ST ), | |||
| $ LDA-1, WORK) | |||
| * | |||
| @@ -359,13 +359,13 @@ | |||
| V( VPOS+I ) = A( DPOS+NB+I, ST ) | |||
| A( DPOS+NB+I, ST ) = ZERO | |||
| 40 CONTINUE | |||
| CALL DLARFG( LM, A( DPOS+NB, ST ), V( VPOS+1 ), 1, | |||
| CALL DLARFG( LM, A( DPOS+NB, ST ), V( VPOS+1 ), 1, | |||
| $ TAU( TAUPOS ) ) | |||
| * | |||
| CALL DLARFX( 'Left', LM, LN-1, V( VPOS ), | |||
| CALL DLARFX( 'Left', LM, LN-1, V( VPOS ), | |||
| $ ( TAU( TAUPOS ) ), | |||
| $ A( DPOS+NB-1, ST+1 ), LDA-1, WORK) | |||
| ENDIF | |||
| ENDIF | |||
| ENDIF | |||
| @@ -374,4 +374,4 @@ | |||
| * | |||
| * END OF DSB2ST_KERNELS | |||
| * | |||
| END | |||
| END | |||
| @@ -261,11 +261,11 @@ | |||
| *> \param[out] INFO | |||
| *> \verbatim | |||
| *> INFO is INTEGER | |||
| *> = 0 : successful exit | |||
| *> < 0 : if INFO = -i, the i-th argument had an illegal value | |||
| *> = 0: successful exit | |||
| *> < 0: if INFO = -i, the i-th argument had an illegal value | |||
| *> <= N: if INFO = i, then i eigenvectors failed to converge. | |||
| *> Their indices are stored in IFAIL. | |||
| *> > N : DPBSTF returned an error code; i.e., | |||
| *> > N: DPBSTF returned an error code; i.e., | |||
| *> if INFO = N + i, for 1 <= i <= N, then the leading | |||
| *> minor of order i of B is not positive definite. | |||
| *> The factorization of B could not be completed and | |||
| @@ -92,9 +92,9 @@ | |||
| *> dimension (LDA,N) | |||
| *> On entry, the N-by-N coefficient matrix A. | |||
| *> On exit, if iterative refinement has been successfully used | |||
| *> (INFO.EQ.0 and ITER.GE.0, see description below), then A is | |||
| *> (INFO = 0 and ITER >= 0, see description below), then A is | |||
| *> unchanged, if double precision factorization has been used | |||
| *> (INFO.EQ.0 and ITER.LT.0, see description below), then the | |||
| *> (INFO = 0 and ITER < 0, see description below), then the | |||
| *> array A contains the factors L and U from the factorization | |||
| *> A = P*L*U; the unit diagonal elements of L are not stored. | |||
| *> \endverbatim | |||
| @@ -111,8 +111,8 @@ | |||
| *> The pivot indices that define the permutation matrix P; | |||
| *> row i of the matrix was interchanged with row IPIV(i). | |||
| *> Corresponds either to the single precision factorization | |||
| *> (if INFO.EQ.0 and ITER.GE.0) or the double precision | |||
| *> factorization (if INFO.EQ.0 and ITER.LT.0). | |||
| *> (if INFO = 0 and ITER >= 0) or the double precision | |||
| *> factorization (if INFO = 0 and ITER < 0). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in] B | |||
| @@ -406,7 +406,7 @@ | |||
| 30 CONTINUE | |||
| * | |||
| * If we are at this place of the code, this is because we have | |||
| * performed ITER=ITERMAX iterations and never satisified the | |||
| * performed ITER=ITERMAX iterations and never satisfied the | |||
| * stopping criterion, set up the ITER flag accordingly and follow up | |||
| * on double precision routine. | |||
| * | |||
| @@ -106,9 +106,9 @@ | |||
| *> triangular part of the matrix A, and the strictly upper | |||
| *> triangular part of A is not referenced. | |||
| *> On exit, if iterative refinement has been successfully used | |||
| *> (INFO.EQ.0 and ITER.GE.0, see description below), then A is | |||
| *> (INFO = 0 and ITER >= 0, see description below), then A is | |||
| *> unchanged, if double precision factorization has been used | |||
| *> (INFO.EQ.0 and ITER.LT.0, see description below), then the | |||
| *> (INFO = 0 and ITER < 0, see description below), then the | |||
| *> array A contains the factor U or L from the Cholesky | |||
| *> factorization A = U**T*U or A = L*L**T. | |||
| *> \endverbatim | |||
| @@ -413,7 +413,7 @@ | |||
| 30 CONTINUE | |||
| * | |||
| * If we are at this place of the code, this is because we have | |||
| * performed ITER=ITERMAX iterations and never satisified the | |||
| * performed ITER=ITERMAX iterations and never satisfied the | |||
| * stopping criterion, set up the ITER flag accordingly and follow | |||
| * up on double precision routine. | |||
| * | |||
| @@ -233,13 +233,13 @@ | |||
| *> \param[in,out] TRYRAC | |||
| *> \verbatim | |||
| *> TRYRAC is LOGICAL | |||
| *> If TRYRAC.EQ..TRUE., indicates that the code should check whether | |||
| *> If TRYRAC = .TRUE., indicates that the code should check whether | |||
| *> the tridiagonal matrix defines its eigenvalues to high relative | |||
| *> accuracy. If so, the code uses relative-accuracy preserving | |||
| *> algorithms that might be (a bit) slower depending on the matrix. | |||
| *> If the matrix does not define its eigenvalues to high relative | |||
| *> accuracy, the code can uses possibly faster algorithms. | |||
| *> If TRYRAC.EQ..FALSE., the code is not required to guarantee | |||
| *> If TRYRAC = .FALSE., the code is not required to guarantee | |||
| *> relatively accurate eigenvalues and can use the fastest possible | |||
| *> techniques. | |||
| *> On exit, a .TRUE. TRYRAC will be set to .FALSE. if the matrix | |||
| @@ -291,7 +291,7 @@ | |||
| * | |||
| * Convert PERMUTATIONS and IPIV | |||
| * | |||
| * Apply permutaions to submatrices of upper part of A | |||
| * Apply permutations to submatrices of upper part of A | |||
| * in factorization order where i decreases from N to 1 | |||
| * | |||
| I = N | |||
| @@ -344,7 +344,7 @@ | |||
| * | |||
| * Revert PERMUTATIONS and IPIV | |||
| * | |||
| * Apply permutaions to submatrices of upper part of A | |||
| * Apply permutations to submatrices of upper part of A | |||
| * in reverse factorization order where i increases from 1 to N | |||
| * | |||
| I = 1 | |||
| @@ -435,7 +435,7 @@ | |||
| * | |||
| * Convert PERMUTATIONS and IPIV | |||
| * | |||
| * Apply permutaions to submatrices of lower part of A | |||
| * Apply permutations to submatrices of lower part of A | |||
| * in factorization order where k increases from 1 to N | |||
| * | |||
| I = 1 | |||
| @@ -488,7 +488,7 @@ | |||
| * | |||
| * Revert PERMUTATIONS and IPIV | |||
| * | |||
| * Apply permutaions to submatrices of lower part of A | |||
| * Apply permutations to submatrices of lower part of A | |||
| * in reverse factorization order where i decreases from N to 1 | |||
| * | |||
| I = N | |||
| @@ -282,7 +282,7 @@ | |||
| * | |||
| * Convert PERMUTATIONS | |||
| * | |||
| * Apply permutaions to submatrices of upper part of A | |||
| * Apply permutations to submatrices of upper part of A | |||
| * in factorization order where i decreases from N to 1 | |||
| * | |||
| I = N | |||
| @@ -333,7 +333,7 @@ | |||
| * | |||
| * Revert PERMUTATIONS | |||
| * | |||
| * Apply permutaions to submatrices of upper part of A | |||
| * Apply permutations to submatrices of upper part of A | |||
| * in reverse factorization order where i increases from 1 to N | |||
| * | |||
| I = 1 | |||
| @@ -423,7 +423,7 @@ | |||
| * | |||
| * Convert PERMUTATIONS | |||
| * | |||
| * Apply permutaions to submatrices of lower part of A | |||
| * Apply permutations to submatrices of lower part of A | |||
| * in factorization order where i increases from 1 to N | |||
| * | |||
| I = 1 | |||
| @@ -474,7 +474,7 @@ | |||
| * | |||
| * Revert PERMUTATIONS | |||
| * | |||
| * Apply permutaions to submatrices of lower part of A | |||
| * Apply permutations to submatrices of lower part of A | |||
| * in reverse factorization order where i decreases from N to 1 | |||
| * | |||
| I = N | |||
| @@ -317,7 +317,7 @@ | |||
| IF( .NOT.WANTZ ) THEN | |||
| CALL DSTERF( N, W, WORK( INDE ), INFO ) | |||
| ELSE | |||
| * Not available in this release, and agrument checking should not | |||
| * Not available in this release, and argument checking should not | |||
| * let it getting here | |||
| RETURN | |||
| CALL DORGTR( UPLO, N, A, LDA, WORK( INDTAU ), WORK( INDWRK ), | |||
| @@ -385,7 +385,7 @@ | |||
| IF( .NOT.WANTZ ) THEN | |||
| CALL DSTERF( N, W, WORK( INDE ), INFO ) | |||
| ELSE | |||
| * Not available in this release, and agrument checking should not | |||
| * Not available in this release, and argument checking should not | |||
| * let it getting here | |||
| RETURN | |||
| CALL DSTEDC( 'I', N, W, WORK( INDE ), WORK( INDWRK ), N, | |||
| @@ -271,7 +271,7 @@ | |||
| *> information as described below. There currently are up to three | |||
| *> pieces of information returned for each right-hand side. If | |||
| *> componentwise accuracy is not requested (PARAMS(3) = 0.0), then | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS .LT. 3, then at most | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS < 3, then at most | |||
| *> the first (:,N_ERR_BNDS) entries are returned. | |||
| *> | |||
| *> The first index in ERR_BNDS_COMP(i,:) corresponds to the ith | |||
| @@ -307,14 +307,14 @@ | |||
| *> \param[in] NPARAMS | |||
| *> \verbatim | |||
| *> NPARAMS is INTEGER | |||
| *> Specifies the number of parameters set in PARAMS. If .LE. 0, the | |||
| *> Specifies the number of parameters set in PARAMS. If <= 0, the | |||
| *> PARAMS array is never referenced and default values are used. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in,out] PARAMS | |||
| *> \verbatim | |||
| *> PARAMS is DOUBLE PRECISION array, dimension (NPARAMS) | |||
| *> Specifies algorithm parameters. If an entry is .LT. 0.0, then | |||
| *> Specifies algorithm parameters. If an entry is < 0.0, then | |||
| *> that entry will be filled with default value used for that | |||
| *> parameter. Only positions up to NPARAMS are accessed; defaults | |||
| *> are used for higher-numbered parameters. | |||
| @@ -322,9 +322,9 @@ | |||
| *> PARAMS(LA_LINRX_ITREF_I = 1) : Whether to perform iterative | |||
| *> refinement or not. | |||
| *> Default: 1.0D+0 | |||
| *> = 0.0 : No refinement is performed, and no error bounds are | |||
| *> = 0.0: No refinement is performed, and no error bounds are | |||
| *> computed. | |||
| *> = 1.0 : Use the double-precision refinement algorithm, | |||
| *> = 1.0: Use the double-precision refinement algorithm, | |||
| *> possibly with doubled-single computations if the | |||
| *> compilation environment does not support DOUBLE | |||
| *> PRECISION. | |||
| @@ -42,7 +42,7 @@ | |||
| *> matrices. | |||
| *> | |||
| *> Aasen's algorithm is used to factor A as | |||
| *> A = U * T * U**T, if UPLO = 'U', or | |||
| *> A = U**T * T * U, if UPLO = 'U', or | |||
| *> A = L * T * L**T, if UPLO = 'L', | |||
| *> where U (or L) is a product of permutation and unit upper (lower) | |||
| *> triangular matrices, and T is symmetric tridiagonal. The factored | |||
| @@ -86,7 +86,7 @@ | |||
| *> | |||
| *> On exit, if INFO = 0, the tridiagonal matrix T and the | |||
| *> multipliers used to obtain the factor U or L from the | |||
| *> factorization A = U*T*U**T or A = L*T*L**T as computed by | |||
| *> factorization A = U**T*T*U or A = L*T*L**T as computed by | |||
| *> DSYTRF. | |||
| *> \endverbatim | |||
| *> | |||
| @@ -230,7 +230,7 @@ | |||
| RETURN | |||
| END IF | |||
| * | |||
| * Compute the factorization A = U*T*U**T or A = L*T*L**T. | |||
| * Compute the factorization A = U**T*T*U or A = L*T*L**T. | |||
| * | |||
| CALL DSYTRF_AA( UPLO, N, A, LDA, IPIV, WORK, LWORK, INFO ) | |||
| IF( INFO.EQ.0 ) THEN | |||
| @@ -45,7 +45,7 @@ | |||
| *> matrices. | |||
| *> | |||
| *> Aasen's 2-stage algorithm is used to factor A as | |||
| *> A = U * T * U**T, if UPLO = 'U', or | |||
| *> A = U**T * T * U, if UPLO = 'U', or | |||
| *> A = L * T * L**T, if UPLO = 'L', | |||
| *> where U (or L) is a product of permutation and unit upper (lower) | |||
| *> triangular matrices, and T is symmetric and band. The matrix T is | |||
| @@ -259,7 +259,7 @@ | |||
| END IF | |||
| * | |||
| * | |||
| * Compute the factorization A = U*T*U**T or A = L*T*L**T. | |||
| * Compute the factorization A = U**T*T*U or A = L*T*L**T. | |||
| * | |||
| CALL DSYTRF_AA_2STAGE( UPLO, N, A, LDA, TB, LTB, IPIV, IPIV2, | |||
| $ WORK, LWORK, INFO ) | |||
| @@ -377,7 +377,7 @@ | |||
| *> information as described below. There currently are up to three | |||
| *> pieces of information returned for each right-hand side. If | |||
| *> componentwise accuracy is not requested (PARAMS(3) = 0.0), then | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS .LT. 3, then at most | |||
| *> ERR_BNDS_COMP is not accessed. If N_ERR_BNDS < 3, then at most | |||
| *> the first (:,N_ERR_BNDS) entries are returned. | |||
| *> | |||
| *> The first index in ERR_BNDS_COMP(i,:) corresponds to the ith | |||
| @@ -413,14 +413,14 @@ | |||
| *> \param[in] NPARAMS | |||
| *> \verbatim | |||
| *> NPARAMS is INTEGER | |||
| *> Specifies the number of parameters set in PARAMS. If .LE. 0, the | |||
| *> Specifies the number of parameters set in PARAMS. If <= 0, the | |||
| *> PARAMS array is never referenced and default values are used. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in,out] PARAMS | |||
| *> \verbatim | |||
| *> PARAMS is DOUBLE PRECISION array, dimension (NPARAMS) | |||
| *> Specifies algorithm parameters. If an entry is .LT. 0.0, then | |||
| *> Specifies algorithm parameters. If an entry is < 0.0, then | |||
| *> that entry will be filled with default value used for that | |||
| *> parameter. Only positions up to NPARAMS are accessed; defaults | |||
| *> are used for higher-numbered parameters. | |||
| @@ -428,9 +428,9 @@ | |||
| *> PARAMS(LA_LINRX_ITREF_I = 1) : Whether to perform iterative | |||
| *> refinement or not. | |||
| *> Default: 1.0D+0 | |||
| *> = 0.0 : No refinement is performed, and no error bounds are | |||
| *> = 0.0: No refinement is performed, and no error bounds are | |||
| *> computed. | |||
| *> = 1.0 : Use the extra-precise refinement algorithm. | |||
| *> = 1.0: Use the extra-precise refinement algorithm. | |||
| *> (other values are reserved for future use) | |||
| *> | |||
| *> PARAMS(LA_LINRX_ITHRESH_I = 2) : Maximum number of residual | |||
| @@ -312,7 +312,7 @@ | |||
| * | |||
| * Factorize A as U*D*U**T using the upper triangle of A | |||
| * | |||
| * Initilize the first entry of array E, where superdiagonal | |||
| * Initialize the first entry of array E, where superdiagonal | |||
| * elements of D are stored | |||
| * | |||
| E( 1 ) = ZERO | |||
| @@ -623,7 +623,7 @@ | |||
| * | |||
| * Factorize A as L*D*L**T using the lower triangle of A | |||
| * | |||
| * Initilize the unused last entry of the subdiagonal array E. | |||
| * Initialize the unused last entry of the subdiagonal array E. | |||
| * | |||
| E( N ) = ZERO | |||
| * | |||
| @@ -123,23 +123,22 @@ | |||
| *> | |||
| *> \param[out] HOUS2 | |||
| *> \verbatim | |||
| *> HOUS2 is DOUBLE PRECISION array, dimension LHOUS2, that | |||
| *> store the Householder representation of the stage2 | |||
| *> HOUS2 is DOUBLE PRECISION array, dimension (LHOUS2) | |||
| *> Stores the Householder representation of the stage2 | |||
| *> band to tridiagonal. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in] LHOUS2 | |||
| *> \verbatim | |||
| *> LHOUS2 is INTEGER | |||
| *> The dimension of the array HOUS2. LHOUS2 = MAX(1, dimension) | |||
| *> If LWORK = -1, or LHOUS2=-1, | |||
| *> The dimension of the array HOUS2. | |||
| *> If LWORK = -1, or LHOUS2 = -1, | |||
| *> then a query is assumed; the routine | |||
| *> only calculates the optimal size of the HOUS2 array, returns | |||
| *> this value as the first entry of the HOUS2 array, and no error | |||
| *> message related to LHOUS2 is issued by XERBLA. | |||
| *> LHOUS2 = MAX(1, dimension) where | |||
| *> dimension = 4*N if VECT='N' | |||
| *> not available now if VECT='H' | |||
| *> If VECT='N', LHOUS2 = max(1, 4*n); | |||
| *> if VECT='V', option not yet available. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] WORK | |||
| @@ -50,9 +50,9 @@ | |||
| * Arguments: | |||
| * ========== | |||
| * | |||
| *> \param[in] STAGE | |||
| *> \param[in] STAGE1 | |||
| *> \verbatim | |||
| *> STAGE is CHARACTER*1 | |||
| *> STAGE1 is CHARACTER*1 | |||
| *> = 'N': "No": to mention that the stage 1 of the reduction | |||
| *> from dense to band using the dsytrd_sy2sb routine | |||
| *> was not called before this routine to reproduce AB. | |||
| @@ -481,7 +481,7 @@ | |||
| * | |||
| * Call the kernel | |||
| * | |||
| #if defined(_OPENMP) && _OPENMP >= 201307 | |||
| #if defined(_OPENMP) | |||
| IF( TTYPE.NE.1 ) THEN | |||
| !$OMP TASK DEPEND(in:WORK(MYID+SHIFT-1)) | |||
| !$OMP$ DEPEND(in:WORK(MYID-1)) | |||
| @@ -363,7 +363,7 @@ | |||
| * | |||
| * | |||
| * Set the workspace of the triangular matrix T to zero once such a | |||
| * way everytime T is generated the upper/lower portion will be always zero | |||
| * way every time T is generated the upper/lower portion will be always zero | |||
| * | |||
| CALL DLASET( "A", LDT, KD, ZERO, ZERO, WORK( TPOS ), LDT ) | |||
| * | |||
| @@ -39,7 +39,7 @@ | |||
| *> the Bunch-Kaufman diagonal pivoting method. The form of the | |||
| *> factorization is | |||
| *> | |||
| *> A = U*D*U**T or A = L*D*L**T | |||
| *> A = U**T*D*U or A = L*D*L**T | |||
| *> | |||
| *> where U (or L) is a product of permutation and unit upper (lower) | |||
| *> triangular matrices, and D is symmetric and block diagonal with | |||
| @@ -144,7 +144,7 @@ | |||
| *> | |||
| *> \verbatim | |||
| *> | |||
| *> If UPLO = 'U', then A = U*D*U**T, where | |||
| *> If UPLO = 'U', then A = U**T*D*U, where | |||
| *> U = P(n)*U(n)* ... *P(k)U(k)* ..., | |||
| *> i.e., U is a product of terms P(k)*U(k), where k decreases from n to | |||
| *> 1 in steps of 1 or 2, and D is a block diagonal matrix with 1-by-1 | |||
| @@ -262,7 +262,7 @@ | |||
| * | |||
| IF( UPPER ) THEN | |||
| * | |||
| * Factorize A as U*D*U**T using the upper triangle of A | |||
| * Factorize A as U**T*D*U using the upper triangle of A | |||
| * | |||
| * K is the main loop index, decreasing from N to 1 in steps of | |||
| * KB, where KB is the number of columns factorized by DLASYF; | |||
| @@ -37,7 +37,7 @@ | |||
| *> DSYTRF_AA computes the factorization of a real symmetric matrix A | |||
| *> using the Aasen's algorithm. The form of the factorization is | |||
| *> | |||
| *> A = U*T*U**T or A = L*T*L**T | |||
| *> A = U**T*T*U or A = L*T*L**T | |||
| *> | |||
| *> where U (or L) is a product of permutation and unit upper (lower) | |||
| *> triangular matrices, and T is a symmetric tridiagonal matrix. | |||
| @@ -223,7 +223,7 @@ | |||
| IF( UPPER ) THEN | |||
| * | |||
| * ..................................................... | |||
| * Factorize A as L*D*L**T using the upper triangle of A | |||
| * Factorize A as U**T*D*U using the upper triangle of A | |||
| * ..................................................... | |||
| * | |||
| * Copy first row A(1, 1:N) into H(1:n) (stored in WORK(1:N)) | |||
| @@ -256,7 +256,7 @@ | |||
| $ A( MAX(1, J), J+1 ), LDA, | |||
| $ IPIV( J+1 ), WORK, N, WORK( N*NB+1 ) ) | |||
| * | |||
| * Ajust IPIV and apply it back (J-th step picks (J+1)-th pivot) | |||
| * Adjust IPIV and apply it back (J-th step picks (J+1)-th pivot) | |||
| * | |||
| DO J2 = J+2, MIN(N, J+JB+1) | |||
| IPIV( J2 ) = IPIV( J2 ) + J | |||
| @@ -375,7 +375,7 @@ | |||
| $ A( J+1, MAX(1, J) ), LDA, | |||
| $ IPIV( J+1 ), WORK, N, WORK( N*NB+1 ) ) | |||
| * | |||
| * Ajust IPIV and apply it back (J-th step picks (J+1)-th pivot) | |||
| * Adjust IPIV and apply it back (J-th step picks (J+1)-th pivot) | |||
| * | |||
| DO J2 = J+2, MIN(N, J+JB+1) | |||
| IPIV( J2 ) = IPIV( J2 ) + J | |||
| @@ -38,7 +38,7 @@ | |||
| *> DSYTRF_AA_2STAGE computes the factorization of a real symmetric matrix A | |||
| *> using the Aasen's algorithm. The form of the factorization is | |||
| *> | |||
| *> A = U*T*U**T or A = L*T*L**T | |||
| *> A = U**T*T*U or A = L*T*L**T | |||
| *> | |||
| *> where U (or L) is a product of permutation and unit upper (lower) | |||
| *> triangular matrices, and T is a symmetric band matrix with the | |||
| @@ -103,6 +103,22 @@ | |||
| *> no error message related to LTB is issued by XERBLA. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] IPIV | |||
| *> \verbatim | |||
| *> IPIV is INTEGER array, dimension (N) | |||
| *> On exit, it contains the details of the interchanges, i.e., | |||
| *> the row and column k of A were interchanged with the | |||
| *> row and column IPIV(k). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] IPIV2 | |||
| *> \verbatim | |||
| *> IPIV2 is INTEGER array, dimension (N) | |||
| *> On exit, it contains the details of the interchanges, i.e., | |||
| *> the row and column k of T were interchanged with the | |||
| *> row and column IPIV2(k). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] WORK | |||
| *> \verbatim | |||
| *> WORK is DOUBLE PRECISION workspace of size LWORK | |||
| @@ -120,22 +136,6 @@ | |||
| *> no error message related to LWORK is issued by XERBLA. | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] IPIV | |||
| *> \verbatim | |||
| *> IPIV is INTEGER array, dimension (N) | |||
| *> On exit, it contains the details of the interchanges, i.e., | |||
| *> the row and column k of A were interchanged with the | |||
| *> row and column IPIV(k). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] IPIV2 | |||
| *> \verbatim | |||
| *> IPIV2 is INTEGER array, dimension (N) | |||
| *> On exit, it contains the details of the interchanges, i.e., | |||
| *> the row and column k of T were interchanged with the | |||
| *> row and column IPIV2(k). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] INFO | |||
| *> \verbatim | |||
| *> INFO is INTEGER | |||
| @@ -275,7 +275,7 @@ | |||
| IF( UPPER ) THEN | |||
| * | |||
| * ..................................................... | |||
| * Factorize A as L*D*L**T using the upper triangle of A | |||
| * Factorize A as U**T*D*U using the upper triangle of A | |||
| * ..................................................... | |||
| * | |||
| DO J = 0, NT-1 | |||
| @@ -442,12 +442,14 @@ c END IF | |||
| * > Apply pivots to previous columns of L | |||
| CALL DSWAP( K-1, A( (J+1)*NB+1, I1 ), 1, | |||
| $ A( (J+1)*NB+1, I2 ), 1 ) | |||
| * > Swap A(I1+1:M, I1) with A(I2, I1+1:M) | |||
| CALL DSWAP( I2-I1-1, A( I1, I1+1 ), LDA, | |||
| $ A( I1+1, I2 ), 1 ) | |||
| * > Swap A(I1+1:M, I1) with A(I2, I1+1:M) | |||
| IF( I2.GT.(I1+1) ) | |||
| $ CALL DSWAP( I2-I1-1, A( I1, I1+1 ), LDA, | |||
| $ A( I1+1, I2 ), 1 ) | |||
| * > Swap A(I2+1:M, I1) with A(I2+1:M, I2) | |||
| CALL DSWAP( N-I2, A( I1, I2+1 ), LDA, | |||
| $ A( I2, I2+1 ), LDA ) | |||
| IF( I2.LT.N ) | |||
| $ CALL DSWAP( N-I2, A( I1, I2+1 ), LDA, | |||
| $ A( I2, I2+1 ), LDA ) | |||
| * > Swap A(I1, I1) with A(I2, I2) | |||
| PIV = A( I1, I1 ) | |||
| A( I1, I1 ) = A( I2, I2 ) | |||
| @@ -616,11 +618,13 @@ c END IF | |||
| CALL DSWAP( K-1, A( I1, (J+1)*NB+1 ), LDA, | |||
| $ A( I2, (J+1)*NB+1 ), LDA ) | |||
| * > Swap A(I1+1:M, I1) with A(I2, I1+1:M) | |||
| CALL DSWAP( I2-I1-1, A( I1+1, I1 ), 1, | |||
| $ A( I2, I1+1 ), LDA ) | |||
| IF( I2.GT.(I1+1) ) | |||
| $ CALL DSWAP( I2-I1-1, A( I1+1, I1 ), 1, | |||
| $ A( I2, I1+1 ), LDA ) | |||
| * > Swap A(I2+1:M, I1) with A(I2+1:M, I2) | |||
| CALL DSWAP( N-I2, A( I2+1, I1 ), 1, | |||
| $ A( I2+1, I2 ), 1 ) | |||
| IF( I2.LT.N ) | |||
| $ CALL DSWAP( N-I2, A( I2+1, I1 ), 1, | |||
| $ A( I2+1, I2 ), 1 ) | |||
| * > Swap A(I1, I1) with A(I2, I2) | |||
| PIV = A( I1, I1 ) | |||
| A( I1, I1 ) = A( I2, I2 ) | |||
| @@ -62,7 +62,7 @@ | |||
| *> \param[in,out] A | |||
| *> \verbatim | |||
| *> A is DOUBLE PRECISION array, dimension (LDA,N) | |||
| *> On entry, the NB diagonal matrix D and the multipliers | |||
| *> On entry, the block diagonal matrix D and the multipliers | |||
| *> used to obtain the factor U or L as computed by DSYTRF. | |||
| *> | |||
| *> On exit, if INFO = 0, the (symmetric) inverse of the original | |||
| @@ -82,7 +82,7 @@ | |||
| *> \param[in] IPIV | |||
| *> \verbatim | |||
| *> IPIV is INTEGER array, dimension (N) | |||
| *> Details of the interchanges and the NB structure of D | |||
| *> Details of the interchanges and the block structure of D | |||
| *> as determined by DSYTRF. | |||
| *> \endverbatim | |||
| *> | |||
| @@ -37,7 +37,7 @@ | |||
| *> \verbatim | |||
| *> | |||
| *> DSYTRS_AA solves a system of linear equations A*X = B with a real | |||
| *> symmetric matrix A using the factorization A = U*T*U**T or | |||
| *> symmetric matrix A using the factorization A = U**T*T*U or | |||
| *> A = L*T*L**T computed by DSYTRF_AA. | |||
| *> \endverbatim | |||
| * | |||
| @@ -49,7 +49,7 @@ | |||
| *> UPLO is CHARACTER*1 | |||
| *> Specifies whether the details of the factorization are stored | |||
| *> as an upper or lower triangular matrix. | |||
| *> = 'U': Upper triangular, form is A = U*T*U**T; | |||
| *> = 'U': Upper triangular, form is A = U**T*T*U; | |||
| *> = 'L': Lower triangular, form is A = L*T*L**T. | |||
| *> \endverbatim | |||
| *> | |||
| @@ -97,14 +97,16 @@ | |||
| *> The leading dimension of the array B. LDB >= max(1,N). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in] WORK | |||
| *> \param[out] WORK | |||
| *> \verbatim | |||
| *> WORK is DOUBLE array, dimension (MAX(1,LWORK)) | |||
| *> WORK is DOUBLE PRECISION array, dimension (MAX(1,LWORK)) | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in] LWORK | |||
| *> \verbatim | |||
| *> LWORK is INTEGER, LWORK >= MAX(1,3*N-2). | |||
| *> LWORK is INTEGER | |||
| *> The dimension of the array WORK. LWORK >= max(1,3*N-2). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[out] INFO | |||
| *> \verbatim | |||
| @@ -198,22 +200,29 @@ | |||
| * | |||
| IF( UPPER ) THEN | |||
| * | |||
| * Solve A*X = B, where A = U*T*U**T. | |||
| * Solve A*X = B, where A = U**T*T*U. | |||
| * | |||
| * 1) Forward substitution with U**T | |||
| * | |||
| IF( N.GT.1 ) THEN | |||
| * | |||
| * Pivot, P**T * B -> B | |||
| * | |||
| * Pivot, P**T * B | |||
| DO K = 1, N | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| * | |||
| DO K = 1, N | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| * Compute U**T \ B -> B [ (U**T \P**T * B) ] | |||
| * | |||
| * Compute (U \P**T * B) -> B [ (U \P**T * B) ] | |||
| CALL DTRSM('L', 'U', 'T', 'U', N-1, NRHS, ONE, A( 1, 2 ), | |||
| $ LDA, B( 2, 1 ), LDB) | |||
| END IF | |||
| * | |||
| CALL DTRSM('L', 'U', 'T', 'U', N-1, NRHS, ONE, A( 1, 2 ), LDA, | |||
| $ B( 2, 1 ), LDB) | |||
| * 2) Solve with triangular matrix T | |||
| * | |||
| * Compute T \ B -> B [ T \ (U \P**T * B) ] | |||
| * Compute T \ B -> B [ T \ (U**T \P**T * B) ] | |||
| * | |||
| CALL DLACPY( 'F', 1, N, A( 1, 1 ), LDA+1, WORK( N ), 1) | |||
| IF( N.GT.1 ) THEN | |||
| @@ -223,35 +232,47 @@ | |||
| CALL DGTSV( N, NRHS, WORK( 1 ), WORK( N ), WORK( 2*N ), B, LDB, | |||
| $ INFO ) | |||
| * | |||
| * Compute (U**T \ B) -> B [ U**T \ (T \ (U \P**T * B) ) ] | |||
| * 3) Backward substitution with U | |||
| * | |||
| IF( N.GT.1 ) THEN | |||
| * | |||
| CALL DTRSM( 'L', 'U', 'N', 'U', N-1, NRHS, ONE, A( 1, 2 ), LDA, | |||
| $ B( 2, 1 ), LDB) | |||
| * Compute U \ B -> B [ U \ (T \ (U**T \P**T * B) ) ] | |||
| * | |||
| * Pivot, P * B [ P * (U**T \ (T \ (U \P**T * B) )) ] | |||
| CALL DTRSM( 'L', 'U', 'N', 'U', N-1, NRHS, ONE, A( 1, 2 ), | |||
| $ LDA, B( 2, 1 ), LDB) | |||
| * | |||
| DO K = N, 1, -1 | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| * Pivot, P * B -> B [ P * (U \ (T \ (U**T \P**T * B) )) ] | |||
| * | |||
| DO K = N, 1, -1 | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| END IF | |||
| * | |||
| ELSE | |||
| * | |||
| * Solve A*X = B, where A = L*T*L**T. | |||
| * | |||
| * Pivot, P**T * B | |||
| * 1) Forward substitution with L | |||
| * | |||
| DO K = 1, N | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| IF( N.GT.1 ) THEN | |||
| * | |||
| * Pivot, P**T * B -> B | |||
| * | |||
| DO K = 1, N | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| * | |||
| * Compute (L \P**T * B) -> B [ (L \P**T * B) ] | |||
| * Compute L \ B -> B [ (L \P**T * B) ] | |||
| * | |||
| CALL DTRSM( 'L', 'L', 'N', 'U', N-1, NRHS, ONE, A( 2, 1 ), | |||
| $ LDA, B( 2, 1 ), LDB) | |||
| END IF | |||
| * | |||
| CALL DTRSM( 'L', 'L', 'N', 'U', N-1, NRHS, ONE, A( 2, 1 ), LDA, | |||
| $ B( 2, 1 ), LDB) | |||
| * 2) Solve with triangular matrix T | |||
| * | |||
| * Compute T \ B -> B [ T \ (L \P**T * B) ] | |||
| * | |||
| @@ -263,18 +284,23 @@ | |||
| CALL DGTSV( N, NRHS, WORK( 1 ), WORK(N), WORK( 2*N ), B, LDB, | |||
| $ INFO) | |||
| * | |||
| * Compute (L**T \ B) -> B [ L**T \ (T \ (L \P**T * B) ) ] | |||
| * 3) Backward substitution with L**T | |||
| * | |||
| CALL DTRSM( 'L', 'L', 'T', 'U', N-1, NRHS, ONE, A( 2, 1 ), LDA, | |||
| $ B( 2, 1 ), LDB) | |||
| IF( N.GT.1 ) THEN | |||
| * | |||
| * Compute (L**T \ B) -> B [ L**T \ (T \ (L \P**T * B) ) ] | |||
| * | |||
| * Pivot, P * B [ P * (L**T \ (T \ (L \P**T * B) )) ] | |||
| CALL DTRSM( 'L', 'L', 'T', 'U', N-1, NRHS, ONE, A( 2, 1 ), | |||
| $ LDA, B( 2, 1 ), LDB) | |||
| * | |||
| DO K = N, 1, -1 | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| * Pivot, P * B -> B [ P * (L**T \ (T \ (L \P**T * B) )) ] | |||
| * | |||
| DO K = N, 1, -1 | |||
| KP = IPIV( K ) | |||
| IF( KP.NE.K ) | |||
| $ CALL DSWAP( NRHS, B( K, 1 ), LDB, B( KP, 1 ), LDB ) | |||
| END DO | |||
| END IF | |||
| * | |||
| END IF | |||
| * | |||
| @@ -36,7 +36,7 @@ | |||
| *> \verbatim | |||
| *> | |||
| *> DSYTRS_AA_2STAGE solves a system of linear equations A*X = B with a real | |||
| *> symmetric matrix A using the factorization A = U*T*U**T or | |||
| *> symmetric matrix A using the factorization A = U**T*T*U or | |||
| *> A = L*T*L**T computed by DSYTRF_AA_2STAGE. | |||
| *> \endverbatim | |||
| * | |||
| @@ -48,7 +48,7 @@ | |||
| *> UPLO is CHARACTER*1 | |||
| *> Specifies whether the details of the factorization are stored | |||
| *> as an upper or lower triangular matrix. | |||
| *> = 'U': Upper triangular, form is A = U*T*U**T; | |||
| *> = 'U': Upper triangular, form is A = U**T*T*U; | |||
| *> = 'L': Lower triangular, form is A = L*T*L**T. | |||
| *> \endverbatim | |||
| *> | |||
| @@ -208,15 +208,15 @@ | |||
| * | |||
| IF( UPPER ) THEN | |||
| * | |||
| * Solve A*X = B, where A = U*T*U**T. | |||
| * Solve A*X = B, where A = U**T*T*U. | |||
| * | |||
| IF( N.GT.NB ) THEN | |||
| * | |||
| * Pivot, P**T * B | |||
| * Pivot, P**T * B -> B | |||
| * | |||
| CALL DLASWP( NRHS, B, LDB, NB+1, N, IPIV, 1 ) | |||
| * | |||
| * Compute (U**T \P**T * B) -> B [ (U**T \P**T * B) ] | |||
| * Compute (U**T \ B) -> B [ (U**T \P**T * B) ] | |||
| * | |||
| CALL DTRSM( 'L', 'U', 'T', 'U', N-NB, NRHS, ONE, A(1, NB+1), | |||
| $ LDA, B(NB+1, 1), LDB) | |||
| @@ -234,7 +234,7 @@ | |||
| CALL DTRSM( 'L', 'U', 'N', 'U', N-NB, NRHS, ONE, A(1, NB+1), | |||
| $ LDA, B(NB+1, 1), LDB) | |||
| * | |||
| * Pivot, P * B [ P * (U \ (T \ (U**T \P**T * B) )) ] | |||
| * Pivot, P * B -> B [ P * (U \ (T \ (U**T \P**T * B) )) ] | |||
| * | |||
| CALL DLASWP( NRHS, B, LDB, NB+1, N, IPIV, -1 ) | |||
| * | |||
| @@ -246,11 +246,11 @@ | |||
| * | |||
| IF( N.GT.NB ) THEN | |||
| * | |||
| * Pivot, P**T * B | |||
| * Pivot, P**T * B -> B | |||
| * | |||
| CALL DLASWP( NRHS, B, LDB, NB+1, N, IPIV, 1 ) | |||
| * | |||
| * Compute (L \P**T * B) -> B [ (L \P**T * B) ] | |||
| * Compute (L \ B) -> B [ (L \P**T * B) ] | |||
| * | |||
| CALL DTRSM( 'L', 'L', 'N', 'U', N-NB, NRHS, ONE, A(NB+1, 1), | |||
| $ LDA, B(NB+1, 1), LDB) | |||
| @@ -268,7 +268,7 @@ | |||
| CALL DTRSM( 'L', 'L', 'T', 'U', N-NB, NRHS, ONE, A(NB+1, 1), | |||
| $ LDA, B(NB+1, 1), LDB) | |||
| * | |||
| * Pivot, P * B [ P * (L**T \ (T \ (L \P**T * B) )) ] | |||
| * Pivot, P * B -> B [ P * (L**T \ (T \ (L \P**T * B) )) ] | |||
| * | |||
| CALL DLASWP( NRHS, B, LDB, NB+1, N, IPIV, -1 ) | |||
| * | |||
| @@ -71,7 +71,7 @@ | |||
| *> R * B**T + L * E**T = scale * -F | |||
| *> | |||
| *> This case is used to compute an estimate of Dif[(A, D), (B, E)] = | |||
| *> sigma_min(Z) using reverse communicaton with DLACON. | |||
| *> sigma_min(Z) using reverse communication with DLACON. | |||
| *> | |||
| *> DTGSY2 also (IJOB >= 1) contributes to the computation in DTGSYL | |||
| *> of an upper bound on the separation between to matrix pairs. Then | |||
| @@ -85,7 +85,7 @@ | |||
| *> \param[in] TRANS | |||
| *> \verbatim | |||
| *> TRANS is CHARACTER*1 | |||
| *> = 'N', solve the generalized Sylvester equation (1). | |||
| *> = 'N': solve the generalized Sylvester equation (1). | |||
| *> = 'T': solve the 'transposed' system (3). | |||
| *> \endverbatim | |||
| *> | |||
| @@ -88,20 +88,20 @@ | |||
| *> \param[in] TRANS | |||
| *> \verbatim | |||
| *> TRANS is CHARACTER*1 | |||
| *> = 'N', solve the generalized Sylvester equation (1). | |||
| *> = 'T', solve the 'transposed' system (3). | |||
| *> = 'N': solve the generalized Sylvester equation (1). | |||
| *> = 'T': solve the 'transposed' system (3). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in] IJOB | |||
| *> \verbatim | |||
| *> IJOB is INTEGER | |||
| *> Specifies what kind of functionality to be performed. | |||
| *> =0: solve (1) only. | |||
| *> =1: The functionality of 0 and 3. | |||
| *> =2: The functionality of 0 and 4. | |||
| *> =3: Only an estimate of Dif[(A,D), (B,E)] is computed. | |||
| *> = 0: solve (1) only. | |||
| *> = 1: The functionality of 0 and 3. | |||
| *> = 2: The functionality of 0 and 4. | |||
| *> = 3: Only an estimate of Dif[(A,D), (B,E)] is computed. | |||
| *> (look ahead strategy IJOB = 1 is used). | |||
| *> =4: Only an estimate of Dif[(A,D), (B,E)] is computed. | |||
| *> = 4: Only an estimate of Dif[(A,D), (B,E)] is computed. | |||
| *> ( DGECON on sub-systems is used ). | |||
| *> Not referenced if TRANS = 'T'. | |||
| *> \endverbatim | |||
| @@ -94,7 +94,7 @@ | |||
| *> | |||
| *> \param[in] V | |||
| *> \verbatim | |||
| *> V is DOUBLE PRECISION array, dimension (LDA,K) | |||
| *> V is DOUBLE PRECISION array, dimension (LDV,K) | |||
| *> The i-th row must contain the vector which defines the | |||
| *> elementary reflector H(i), for i = 1,2,...,k, as returned by | |||
| *> DTPLQT in B. See Further Details. | |||
| @@ -94,7 +94,7 @@ | |||
| *> | |||
| *> \param[in] V | |||
| *> \verbatim | |||
| *> V is DOUBLE PRECISION array, dimension (LDA,K) | |||
| *> V is DOUBLE PRECISION array, dimension (LDV,K) | |||
| *> The i-th column must contain the vector which defines the | |||
| *> elementary reflector H(i), for i = 1,2,...,k, as returned by | |||
| *> CTPQRT in B. See Further Details. | |||
| @@ -152,8 +152,8 @@ | |||
| *> \verbatim | |||
| *> LDA is INTEGER | |||
| *> The leading dimension of the array A. | |||
| *> If SIDE = 'L', LDC >= max(1,K); | |||
| *> If SIDE = 'R', LDC >= max(1,M). | |||
| *> If SIDE = 'L', LDA >= max(1,K); | |||
| *> If SIDE = 'R', LDA >= max(1,M). | |||
| *> \endverbatim | |||
| *> | |||
| *> \param[in,out] B | |||