hetrf_rk(3)
π NAME π
hetrf_rk - {he,sy}trf_rk: triangular factor
okay bestie, welcome to the Library Functions Manual page for hetrf_rk(3) β this is the {he,sy}trf_rk family, and theyβre all about that triangular factor life. π
β¨ SYNOPSIS β¨
Functions
okay so hereβs the whole routine squad, no cap β every signature is exactly what you type, and every description keeps the real facts:
subroutine chetrf_rk (uplo, n, a, lda, e, ipiv, work, lwork, info)βCHETRF_RKis the complex-Hermitian-indefinite one: it computes the factorization of a complex Hermitian indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm). π¦subroutine csytrf_rk (uplo, n, a, lda, e, ipiv, work, lwork, info)βCSYTRF_RKhandles the complex-symmetric-indefinite case: it computes the factorization of a complex symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm). β¨subroutine dsytrf_rk (uplo, n, a, lda, e, ipiv, work, lwork, info)βDSYTRF_RKis the real-symmetric-indefinite girlie: it computes the factorization of a real symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm). πsubroutine ssytrf_rk (uplo, n, a, lda, e, ipiv, work, lwork, info)βSSYTRF_RKalso does the real-symmetric-indefinite thing: it computes the factorization of a real symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm). πΈsubroutine zhetrf_rk (uplo, n, a, lda, e, ipiv, work, lwork, info)βZHETRF_RKis the complex-Hermitian-indefinite icon: it computes the factorization of a complex Hermitian indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm). πsubroutine zsytrf_rk (uplo, n, a, lda, e, ipiv, work, lwork, info)βZSYTRF_RKrounds out the set for complex-symmetric-indefinite matrices: it computes the factorization of a complex symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm). π₯
πΈ Detailed Description πΈ
okay so this heading exists and then we immediately roll into the function docs β thatβs the tea. β
π¦ Function Documentation π¦
π subroutine chetrf_rk π
subroutine chetrf_rk (character uplo, integer n, complex, dimension( lda, *
) a, integer lda, complex, dimension( * ) e, integer, dimension( * )
ipiv, complex, dimension( * ) work, integer lwork, integer info)
CHETRF_RK computes the factorization of a complex Hermitian indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm). π
π Purpose: π
bestie, CHETRF_RK computes the factorization of a complex Hermitian matrix A using the bounded Bunch-Kaufman (rook) diagonal pivoting method:
A = P*U*D*(U**H)*(P**T) or A = P*L*D*(L**H)*(P**T)
where U (or L) is unit upper (or lower) triangular matrix, U**H (or L**H) is the conjugate of U (or L), P is a permutation matrix, P**T is the transpose of P, and D is Hermitian and block diagonal with 1-by-1 and 2-by-2 diagonal blocks.
This is the blocked version of the algorithm, calling Level 3 BLAS. For more information see Further Details section. β¨
π Parameters π
UPLOβUPLOisCHARACTER*1. It specifies whether the upper or lower triangular part of the Hermitian matrixAis stored:= 'U': Upper triangular= 'L': Lower triangular
NβNisINTEGER. The order of the matrixA.N >= 0.AβAisCOMPLEXarray, dimension(LDA,N). On entry, the Hermitian matrixA.- If
UPLO = 'U': the leadingN-by-Nupper triangular part ofAcontains the upper triangular part of the matrixA, and the strictly lower triangular part ofAis not referenced. - If
UPLO = 'L': the leadingN-by-Nlower triangular part ofAcontains the lower triangular part of the matrixA, and the strictly upper triangular part ofAis not referenced. - On exit, contains:
- a) ONLY diagonal elements of the Hermitian block diagonal matrix
Don the diagonal ofA, i.e.D(k,k) = A(k,k); (superdiagonal (or subdiagonal) elements ofDare stored on exit in arrayE), and - b) If
UPLO = 'U': factorUin the superdiagonal part ofA. IfUPLO = 'L': factorLin the subdiagonal part ofA.
- a) ONLY diagonal elements of the Hermitian block diagonal matrix
- If
LDAβLDAisINTEGER. The leading dimension of the arrayA.LDA >= max(1,N).EβEisCOMPLEXarray, dimension(N). On exit, contains the superdiagonal (or subdiagonal) elements of the Hermitian block diagonal matrixDwith 1-by-1 or 2-by-2 diagonal blocks, where:- If
UPLO = 'U':E(i) = D(i-1,i),i=2:N,E(1)is set to 0; - If
UPLO = 'L':E(i) = D(i+1,i),i=1:N-1,E(N)is set to 0. - NOTE: For 1-by-1 diagonal block
D(k), where1 <= k <= N, the elementE(k)is set to 0 in bothUPLO = 'U'orUPLO = 'L'cases.
- If
IPIVβIPIVisINTEGERarray, dimension(N).IPIVdescribes the permutation matrixPin the factorization of matrixAas follows. The absolute value ofIPIV(k)represents the index of row and column that were interchanged with thek-th row and column. The value ofUPLOdescribes the order in which the interchanges were applied. Also, the sign ofIPIVrepresents the block structure of the Hermitian block diagonal matrixDwith 1-by-1 or 2-by-2 diagonal blocks which correspond to 1 or 2 interchanges at each factorization step. For more info see Further Details section.
If UPLO = 'U', ( in factorization order, k decreases from N to 1 ):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N); - If
IPIV(k) = k, no interchange occurred.
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k-1) < 0means:D(k-1:k,k-1:k)is a 2-by-2 diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs). - 1) If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - 2) If
-IPIV(k-1) != k-1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k-1) = k-1, no interchange occurred.
- c) In both cases a) and b), always
ABS( IPIV(k) ) <= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
If UPLO = 'L', ( in factorization order, k increases from 1 to N ):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N). - If
IPIV(k) = k, no interchange occurred.
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k+1) < 0means:D(k:k+1,k:k+1)is a 2-by-2 diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs). - 1) If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - 2) If
-IPIV(k+1) != k+1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k+1) = k+1, no interchange occurred.
- c) In both cases a) and b), always
ABS( IPIV(k) ) >= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
WORKβWORKisCOMPLEXarray, dimension(MAX(1,LWORK)). On exit, ifINFO = 0,WORK(1)returns the optimalLWORK.LWORKβLWORKisINTEGER. The length ofWORK.LWORK >= 1. For best performanceLWORK >= N*NB, whereNBis the block size returned byILAENV.- If
LWORK = -1, then a workspace query is assumed; the routine only calculates the optimal size of theWORKarray, returns this value as the first entry of theWORKarray, and no error message related toLWORKis issued byXERBLA.
- If
INFOβINFOisINTEGER.= 0: successful exit< 0: IfINFO = -k, thek-th argument had an illegal value> 0: IfINFO = k, the matrixAis singular, because:- If
UPLO = 'U': columnkin the upper triangular part ofAcontains all zeros. - If
UPLO = 'L': columnkin the lower triangular part ofAcontains all zeros. - Therefore
D(k,k)is exactly zero, and superdiagonal elements of columnkofU(or subdiagonal elements of columnkofL) are all zeros. The factorization has been completed, but the block diagonal matrixDis exactly singular, and division by zero will occur if it is used to solve a system of equations. - NOTE:
INFOonly stores the first occurrence of a singularity, any subsequent occurrence of singularity is not stored inINFOeven though the factorization always completes.
- If
π Author π
- Univ. of Tennessee
- Univ. of California Berkeley
- Univ. of Colorado Denver
- NAG Ltd.
π§ Further Details: π§
TODO: put correct description
π«Ά Contributors: π«Ά
and the contributor list for chetrf_rk keeps the same energy:
- December 2016, Igor Kozachenko, Computer Science Division, University of California, Berkeley
- September 2007, Sven Hammarling, Nicholas J. Higham, Craig Lucas, School of Mathematics, University of Manchester
Definition at line 257 of file chetrf_rk.f.
π csytrf_rk β complex symmetric, rook pivoting, blocked era
subroutine csytrf_rk (character uplo, integer n, complex, dimension( lda, * ) a, integer lda, complex, dimension( * ) e, integer, dimension( * ) ipiv, complex, dimension( * ) work, integer lwork, integer info)
CSYTRF_RK computes the factorization of a complex symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method β the BLAS3 blocked algorithm version, bestie. Itβs giving serious linear algebra main character energy. π
π― Purpose
okay so hereβs the whole vibe: CSYTRF_RK computes the factorization of a complex symmetric matrix A using the bounded Bunch-Kaufman (rook) diagonal pivoting method:
A = P*U*D*(U**T)*(P**T) or A = P*L*D*(L**T)*(P**T),
where U (or L) is unit upper (or lower) triangular matrix, U**T (or L**T) is the transpose of U (or L), P is a permutation matrix, P**T is the transpose of P, and D is symmetric and block diagonal with 1-by-1 and 2-by-2 diagonal blocks.
This is the blocked version of the algorithm, calling Level 3 BLAS. For more information see Further Details section. iykyk π
π Parameters
UPLOβ isCHARACTER*1. Specifies whether the upper or lower triangular part of the symmetric matrixAis stored:= 'U': Upper triangular= 'L': Lower triangular
Nβ isINTEGER. The order of the matrixA.N >= 0.Aβ isCOMPLEX array, dimension (LDA,N). On entry, the symmetric matrixA.- If
UPLO = 'U': the leadingN-by-Nupper triangular part ofAcontains the upper triangular part of the matrixA, and the strictly lower triangular part ofAis not referenced. - If
UPLO = 'L': the leadingN-by-Nlower triangular part ofAcontains the lower triangular part of the matrixA, and the strictly upper triangular part ofAis not referenced.
- If
On exit, contains:
- a) ONLY diagonal elements of the symmetric block diagonal matrix
Don the diagonal ofA, i.e.D(k,k) = A(k,k); (superdiagonal (or subdiagonal) elements ofDare stored on exit in arrayE), and - b) If
UPLO = 'U': factorUin the superdiagonal part ofA. IfUPLO = 'L': factorLin the subdiagonal part ofA.
LDAβ isINTEGER. The leading dimension of the arrayA.LDA >= max(1,N).Eβ isCOMPLEX array, dimension (N). On exit, contains the superdiagonal (or subdiagonal) elements of the symmetric block diagonal matrixDwith 1-by-1 or 2-by-2 diagonal blocks, where- If
UPLO = 'U':E(i) = D(i-1,i),i=2:N,E(1)is set to 0; - If
UPLO = 'L':E(i) = D(i+1,i),i=1:N-1,E(N)is set to 0.
- If
NOTE: For 1-by-1 diagonal blockD(k), where1 <= k <= N, the elementE(k)is set to 0 in bothUPLO = 'U'orUPLO = 'L'cases. donβt forget it π
IPIVβ isINTEGER array, dimension (N).IPIVdescribes the permutation matrixPin the factorization of matrixAas follows. The absolute value ofIPIV(k)represents the index of row and column that were interchanged with thek-th row and column. The value ofUPLOdescribes the order in which the interchanges were applied. Also, the sign ofIPIVrepresents the block structure of the symmetric block diagonal matrixDwith 1-by-1 or 2-by-2 diagonal blocks which correspond to 1 or 2 interchanges at each factorization step. For more info see Further Details section.- If
UPLO = 'U', ( in factorization order,kdecreases fromNto 1 ):- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block. IfIPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N); IfIPIV(k) = k, no interchange occurred. - b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k-1) < 0means:D(k-1:k,k-1:k)is a 2-by-2 diagonal block. (NOTE: negative entries inIPIVappear ONLY in pairs).- If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - If
-IPIV(k-1) != k-1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k-1) = k-1, no interchange occurred.
- If
- c) In both cases a) and b), always
ABS( IPIV(k) ) <= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
- a) A single positive entry
- If
UPLO = 'L', ( in factorization order,kincreases from 1 toN):- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block. IfIPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N). IfIPIV(k) = k, no interchange occurred. - b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k+1) < 0means:D(k:k+1,k:k+1)is a 2-by-2 diagonal block. (NOTE: negative entries inIPIVappear ONLY in pairs).- If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - If
-IPIV(k+1) != k+1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k+1) = k+1, no interchange occurred.
- If
- c) In both cases a) and b), always
ABS( IPIV(k) ) >= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
- a) A single positive entry
- If
WORKβ isCOMPLEX array, dimension ( MAX(1,LWORK) ). On exit, ifINFO = 0,WORK(1)returns the optimalLWORK.LWORKβ isINTEGER. The length ofWORK.LWORK >=1. For best performanceLWORK >= N*NB, whereNBis the block size returned byILAENV.
If LWORK = -1, then a workspace query is assumed; the routine only calculates the optimal size of the WORK array, returns this value as the first entry of the WORK array, and no error message related to LWORK is issued by XERBLA. <butterfly>π¦</butterfly>
INFOβ isINTEGER.= 0: successful exit< 0: IfINFO = -k, thek-th argument had an illegal value> 0: IfINFO = k, the matrixAis singular, because:- If
UPLO = 'U': columnkin the upper triangular part ofAcontains all zeros. - If
UPLO = 'L': columnkin the lower triangular part ofAcontains all zeros.
- If
Therefore D(k,k) is exactly zero, and superdiagonal elements of column k of U (or subdiagonal elements of column k of L ) are all zeros. The factorization has been completed, but the block diagonal matrix D is exactly singular, and division by zero will occur if it is used to solve a system of equations.
NOTE:INFOonly stores the first occurrence of a singularity, any subsequent occurrence of singularity is not stored inINFOeven though the factorization always completes. this part is important, bestie β division by zero is not a vibe π
π Author
- Univ. of Tennessee
- Univ. of California Berkeley
- Univ. of Colorado Denver
- NAG Ltd.
π Further Details
!> TODO: put correct description
yes, the details section is literally a TODO. iconic behavior, honestly. π
β¨ Contributors
- December 2016, Igor Kozachenko, Computer Science Division, University of California, Berkeley
- September 2007, Sven Hammarling, Nicholas J. Higham, Craig Lucas, School of Mathematics, University of Manchester
Definition at line 257 of file csytrf_rk.f.
π¦ dsytrf_rk β real symmetric, rook pivoting, blocked era
subroutine dsytrf_rk (character uplo, integer n, double precision, dimension( lda, * ) a, integer lda, double precision, dimension( * ) e, integer, dimension( * ) ipiv, double precision, dimension( * ) work, integer lwork, integer info)
DSYTRF_RK computes the factorization of a real symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method β the BLAS3 blocked algorithm version. same energy, real arithmetic, still slaying β¨
π― Purpose
π Parameters π
bestie. the numbers are the moment. here's what DSYTRF_RK actually means by all of it: it computes the factorization of a real symmetric matrix A using the bounded Bunch-Kaufman (rook) diagonal pivoting method, and the whole thing is:
A = P*U*D*(U**T)*(P**T) or A = P*L*D*(L**T)*(P**T),
where U (or L) is unit upper (or lower) triangular matrix, UT (or LT) is the transpose of U (or L), P is a permutation matrix, P**T is the transpose of P, and D is symmetric and block diagonal with 1-by-1 and 2-by-2 diagonal blocks.
this is the blocked version of the algorithm, calling Level 3 BLAS. For more information see Further Details section. iykyk π₯
UPLOβCHARACTER*1. Specifies whether the upper or lower triangular part of the symmetric matrix A is stored:= 'U': Upper triangular= 'L': Lower triangular
NβINTEGER. The order of the matrix A.N >= 0.AβDOUBLE PRECISIONarray, dimension(LDA,N). On entry, the symmetric matrix A.- If
UPLO = 'U': the leading N-by-N upper triangular part of A contains the upper triangular part of the matrix A, and the strictly lower triangular part of A is not referenced. - If
UPLO = 'L': the leading N-by-N lower triangular part of A contains the lower triangular part of the matrix A, and the strictly upper triangular part of A is not referenced.
- If
On exit, contains:
- a) ONLY diagonal elements of the symmetric block diagonal matrix D on the diagonal of A, i.e.
D(k,k) = A(k,k); (superdiagonal (or subdiagonal) elements of D are stored on exit in array E), and - b) If
UPLO = 'U': factor U in the superdiagonal part of A. IfUPLO = 'L': factor L in the subdiagonal part of A.
LDAβINTEGER. The leading dimension of the array A.LDA >= max(1,N).EβDOUBLE PRECISIONarray, dimension(N). On exit, contains the superdiagonal (or subdiagonal) elements of the symmetric block diagonal matrix D with 1-by-1 or 2-by-2 diagonal blocks, where- If
UPLO = 'U':E(i) = D(i-1,i),i=2:N,E(1)is set to 0; - If
UPLO = 'L':E(i) = D(i+1,i),i=1:N-1,E(N)is set to 0.
- If
NOTE: For 1-by-1 diagonal blockD(k), where1 <= k <= N, the elementE(k)is set to 0 in bothUPLO = 'U'orUPLO = 'L'cases.
IPIVβINTEGERarray, dimension(N).IPIVdescribes the permutation matrix P in the factorization of matrix A as follows. The absolute value ofIPIV(k)represents the index of row and column that were interchanged with the k-th row and column. The value ofUPLOdescribes the order in which the interchanges were applied. Also, the sign ofIPIVrepresents the block structure of the symmetric block diagonal matrix D with 1-by-1 or 2-by-2 diagonal blocks which correspond to 1 or 2 interchanges at each factorization step. For more info see Further Details section. π
If UPLO = 'U', ( in factorization order, k decreases from N to 1 ):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block. IfIPIV(k) != k, rows and columns k andIPIV(k)were interchanged in the matrixA(1:N,1:N); IfIPIV(k) = k, no interchange occurred. - b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k-1) < 0means:D(k-1:k,k-1:k)is a 2-by-2 diagonal block. (NOTE: negative entries inIPIVappear ONLY in pairs).- If
-IPIV(k) != k, rows and columns k and-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - If
-IPIV(k-1) != k-1, rows and columns k-1 and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k-1) = k-1, no interchange occurred.
- If
- c) In both cases a) and b), always
ABS( IPIV(k) ) <= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
If UPLO = 'L', ( in factorization order, k increases from 1 to N ):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block. IfIPIV(k) != k, rows and columns k andIPIV(k)were interchanged in the matrixA(1:N,1:N). IfIPIV(k) = k, no interchange occurred. - b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k+1) < 0means:D(k:k+1,k:k+1)is a 2-by-2 diagonal block. (NOTE: negative entries inIPIVappear ONLY in pairs).- If
-IPIV(k) != k, rows and columns k and-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - If
-IPIV(k+1) != k+1, rows and columns k-1 and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k+1) = k+1, no interchange occurred.
- If
- c) In both cases a) and b), always
ABS( IPIV(k) ) >= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
WORKβDOUBLE PRECISIONarray, dimension(MAX(1,LWORK)). On exit, ifINFO = 0,WORK(1)returns the optimalLWORK.LWORKβINTEGER. The length ofWORK.LWORK >= 1. For best performanceLWORK >= N*NB, where NB is the block size returned byILAENV.
IfLWORK = -1, then a workspace query is assumed; the routine only calculates the optimal size of the WORK array, returns this value as the first entry of the WORK array, and no error message related toLWORKis issued byXERBLA. no cap, that's the shortcut π
INFOβINTEGER.= 0: successful exit< 0: IfINFO = -k, the k-th argument had an illegal value> 0: IfINFO = k, the matrix A is singular, because: IfUPLO = 'U': column k in the upper triangular part of A contains all zeros. IfUPLO = 'L': column k in the lower triangular part of A contains all zeros.
Therefore D(k,k) is exactly zero, and superdiagonal elements of column k of U (or subdiagonal elements of column k of L ) are all zeros. The factorization has been completed, but the block diagonal matrix D is exactly singular, and division by zero will occur if it is used to solve a system of equations. π literally do not do that
NOTE:INFOonly stores the first occurrence of a singularity, any subsequent occurrence of singularity is not stored inINFOeven though the factorization always completes. it's giving "we only log the first one" energy
π Author π
Univ. of Tennessee
Univ. of California Berkeley
Univ. of Colorado Denver
NAG Ltd.
πΈ Further Details: πΈ
!> TODO: put correct description
!>
not the TODO in the official docs π iconic honestly. anyway, the page says exactly that, so that's what you get.
π¦ Contributors: π¦
December 2016, Igor Kozachenko, Computer Science Division, University of California, Berkeley
September 2007, Sven Hammarling, Nicholas J. Higham, Craig Lucas, School of Mathematics, University of Manchester
Definition at line 257 of file dsytrf_rk.f.
subroutine ssytrf_rk (character uplo, integer n, real, dimension( lda, * )
a, integer lda, real, dimension( * ) e, integer, dimension( * ) ipiv,
real, dimension( * ) work, integer lwork, integer info)
SSYTRF_RK computes the factorization of a real symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm).
β Purpose: β
so SSYTRF_RK computes the factorization of a real symmetric matrix A using the bounded Bunch-Kaufman (rook) diagonal pivoting method:
A = P*U*D*(U**T)*(P**T) or A = P*L*D*(L**T)*(P**T),
where U (or L) is unit upper (or lower) triangular matrix, UT (or LT) is the transpose of U (or L), P is a permutation matrix, P**T is the transpose of P, and D is symmetric and block diagonal with 1-by-1 and 2-by-2 diagonal blocks.
this is the blocked version of the algorithm, calling Level 3 BLAS. For more information see Further Details section. same vibe, single precision edition β¨
π Parameters π
UPLOβCHARACTER*1. Specifies whether the upper or lower triangular part of the symmetric matrix A is stored:= 'U': Upper triangular= 'L': Lower triangular
NβINTEGER. The order of the matrix A.N >= 0.Aβ
okay so A is the REAL array, dimension (LDA,N).
On entry, the symmetric matrix A.
- If
UPLO = 'U': the leading N-by-N upper triangular part ofAcontains the upper triangular part of the matrixA, and the strictly lower triangular part ofAis not referenced. - If
UPLO = 'L': the leading N-by-N lower triangular part ofAcontains the lower triangular part of the matrixA, and the strictly upper triangular part ofAis not referenced.
On exit, contains:
- a) ONLY diagonal elements of the symmetric block diagonal matrix
Don the diagonal ofA, i.e.D(k,k) = A(k,k);(superdiagonal (or subdiagonal) elements ofDare stored on exit in arrayE), and - b) If
UPLO = 'U': factorUin the superdiagonal part ofA. IfUPLO = 'L': factorLin the subdiagonal part ofA.
LDA π
LDA is INTEGER. The leading dimension of the array A. LDA >= max(1,N).
E π¦
E is REAL array, dimension (N).
On exit, contains the superdiagonal (or subdiagonal) elements of the symmetric block diagonal matrix D with 1-by-1 or 2-by-2 diagonal blocks, where
- If
UPLO = 'U':E(i) = D(i-1,i),i=2:N,E(1)is set to 0; - If
UPLO = 'L':E(i) = D(i+1,i),i=1:N-1,E(N)is set to 0.
NOTE: For 1-by-1 diagonal blockD(k), where1 <= k <= N, the elementE(k)is set to 0 in bothUPLO = 'U'orUPLO = 'L'cases.
IPIV π
IPIV is INTEGER array, dimension (N).
IPIV describes the permutation matrix P in the factorization of matrix A as follows. The absolute value of IPIV(k) represents the index of row and column that were interchanged with the k-th row and column. The value of UPLO describes the order in which the interchanges were applied. Also, the sign of IPIV represents the block structure of the symmetric block diagonal matrix D with 1-by-1 or 2-by-2 diagonal blocks which correspond to 1 or 2 interchanges at each factorization step. For more info see Further Details section.
If UPLO = 'U', ( in factorization order, k decreases from N to 1 ):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N); - If
IPIV(k) = k, no interchange occurred.
- If
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k-1) < 0means:D(k-1:k,k-1:k)is a 2-by-2 diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs). - 1) If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - 2) If
-IPIV(k-1) != k-1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k-1) = k-1, no interchange occurred.
- (NOTE: negative entries in
- c) In both cases a) and b), always
ABS( IPIV(k) ) <= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
If UPLO = 'L', ( in factorization order, k increases from 1 to N ):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a 1-by-1 diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N). - If
IPIV(k) = k, no interchange occurred.
- If
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k+1) < 0means:D(k:k+1,k:k+1)is a 2-by-2 diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs). - 1) If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - 2) If
-IPIV(k+1) != k+1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k+1) = k+1, no interchange occurred.
- (NOTE: negative entries in
- c) In both cases a) and b), always
ABS( IPIV(k) ) >= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
WORK π§
WORK is REAL array, dimension (MAX(1,LWORK)).
On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
LWORK π
LWORK is INTEGER. The length of WORK. LWORK >= 1. For best performance LWORK >= N*NB, where NB is the block size returned by ILAENV.
If LWORK = -1, then a workspace query is assumed; the routine only calculates the optimal size of the WORK array, returns this value as the first entry of the WORK array, and no error message related to LWORK is issued by XERBLA.
INFO β
INFO is INTEGER.
= 0: successful exit< 0: IfINFO = -k, the k-th argument had an illegal value> 0: IfINFO = k, the matrixAis singular, because:- If
UPLO = 'U': columnkin the upper triangular part ofAcontains all zeros. - If
UPLO = 'L': columnkin the lower triangular part ofAcontains all zeros.
- If
Therefore D(k,k) is exactly zero, and superdiagonal elements of column k of U (or subdiagonal elements of column k of L ) are all zeros. The factorization has been completed, but the block diagonal matrix D is exactly singular, and division by zero will occur if it is used to solve a system of equations.
NOTE:INFOonly stores the first occurrence of a singularity, any subsequent occurrence of singularity is not stored inINFOeven though the factorization always completes.
Author β¨
- Univ. of Tennessee
- Univ. of California Berkeley
- Univ. of Colorado Denver
- NAG Ltd.
Further Details: π
TODO: put correct description
Contributors: π«Ά
- December 2016, Igor Kozachenko, Computer Science Division, University of California, Berkeley
- September 2007, Sven Hammarling, Nicholas J. Higham, Craig Lucas, School of Mathematics, University of Manchester
Definition at line 257 of file ssytrf_rk.f.
zhetrf_rk π
subroutine zhetrf_rk (character uplo, integer n, complex*16, dimension( lda,
* ) a, integer lda, complex*16, dimension( * ) e, integer, dimension( *
) ipiv, complex*16, dimension( * ) work, integer lwork, integer info)
ZHETRF_RK computes the factorization of a complex Hermitian indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm).
Purpose π«
ZHETRF_RK computes the factorization of a complex Hermitian matrix A using the bounded Bunch-Kaufman (rook) diagonal pivoting method:
A = P*U*D*(U**H)*(P**T) or A = P*L*D*(L**H)*(P**T),
where U (or L) is unit upper (or lower) triangular matrix, U**H (or L**H) is the conjugate of U (or L), P is a permutation matrix, P**T is the transpose of P, and D is Hermitian and block diagonal with 1-by-1 and 2-by-2 diagonal blocks.
This is the blocked version of the algorithm, calling Level 3 BLAS. For more information see Further Details section.
Parameters π
UPLOβUPLOisCHARACTER*1. Specifies whether the upper or lower triangular part of the Hermitian matrixAis stored:= 'U': Upper triangular= 'L': Lower triangular
NβNis INTEGER. The order of the matrixA.N >= 0.AβAisCOMPLEX*16array, dimension(LDA,N). On entry, the Hermitian matrixA.- If
UPLO = 'U': the leading N-by-N upper triangular part ofAcontains the upper triangular part of the matrixA, and the strictly lower triangular part ofAis not referenced. - If
UPLO = 'L': the leading N-by-N lower triangular part ofAcontains the lower triangular part of the matrixA, and the strictly upper triangular part ofAis not referenced.
- If
On exit, contains:
- a) ONLY diagonal elements of the Hermitian block diagonal matrix
Don the diagonal ofA, i.e.D(k,k) = A(k,k);(superdiagonal (or subdiagonal) elements ofDare stored on exit in arrayE), and - b) If
UPLO = 'U': factorUin the superdiagonal part ofA. IfUPLO = 'L': factorLin the subdiagonal part ofA.
LDAβLDAis INTEGER. The leading dimension of the arrayA.LDA >= max(1,N).E
E β okay so E is a COMPLEX*16 array, dimension (N). On exit itβs carrying the superdiagonal (or subdiagonal) elements of the Hermitian block diagonal matrix D, and D has 1-by-1 or 2-by-2 diagonal blocks, where:
- If
UPLO = 'U':E(i) = D(i-1,i),i=2:N, andE(1)is set to0. - If
UPLO = 'L':E(i) = D(i+1,i),i=1:N-1, andE(N)is set to0.
π NOTE: For a1-by-1diagonal blockD(k), where1 <= k <= N, the elementE(k)is set to0in both theUPLO = 'U'andUPLO = 'L'cases.
IPIV β IPIV is an INTEGER array, dimension (N). IPIV describes the permutation matrix P in the factorization of matrix A like this: the absolute value of IPIV(k) represents the index of the row and column that were interchanged with the k-th row and column. The value of UPLO describes the order in which the interchanges were applied. Also, the sign of IPIV represents the block structure of the Hermitian block diagonal matrix D with 1-by-1 or 2-by-2 diagonal blocks, which correspond to 1 or 2 interchanges at each factorization step. For more info see the Further Details section.
- If
UPLO = 'U'β in factorization order,kdecreases fromNto1:- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a1-by-1diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N). - If
IPIV(k) = k, no interchange occurred.
- If
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k-1) < 0means:D(k-1:k,k-1:k)is a2-by-2diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs). - 1) If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - 2) If
-IPIV(k-1) != k-1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k-1) = k-1, no interchange occurred.
- (NOTE: negative entries in
- c) In both cases a) and b), always
ABS( IPIV(k) ) <= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
- a) A single positive entry
- If
UPLO = 'L'β in factorization order,kincreases from1toN:- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a1-by-1diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N). - If
IPIV(k) = k, no interchange occurred.
- If
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k+1) < 0means:D(k:k+1,k:k+1)is a2-by-2diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs). - 1) If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - 2) If
-IPIV(k+1) != k+1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k+1) = k+1, no interchange occurred.
- (NOTE: negative entries in
- c) In both cases a) and b), always
ABS( IPIV(k) ) >= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
- a) A single positive entry
WORK β WORK is a COMPLEX*16 array, dimension (MAX(1,LWORK)). On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
LWORK β LWORK is INTEGER. The length of WORK. LWORK >= 1. For best performance, LWORK >= N*NB, where NB is the block size returned by ILAENV.
If LWORK = -1, then a workspace query is assumed; the routine only calculates the optimal size of the WORK array, returns this value as the first entry of the WORK array, and no error message related to LWORK is issued by XERBLA.
INFO β INFO is INTEGER.
= 0: successful exit.< 0: IfINFO = -k, the k-th argument had an illegal value.> 0: IfINFO = k, the matrixAis singular, because:- If
UPLO = 'U': columnkin the upper triangular part ofAcontains all zeros. - If
UPLO = 'L': columnkin the lower triangular part ofAcontains all zeros.
- If
Therefore D(k,k) is exactly zero, and superdiagonal elements of column k of U (or subdiagonal elements of column k of L ) are all zeros. The factorization has been completed, but the block diagonal matrix D is exactly singular, and division by zero will occur if it is used to solve a system of equations.
π¨ NOTE:INFOonly stores the first occurrence of a singularity, any subsequent occurrence of singularity is not stored inINFOeven though the factorization always completes.
π«Ά Author
- Univ. of Tennessee
- Univ. of California Berkeley
- Univ. of Colorado Denver
- NAG Ltd.
πΈ Further Details
!> TODO: put correct description
!>
β¨ Contributors
- December 2016, Igor Kozachenko, Computer Science Division, University of California, Berkeley
- September 2007, Sven Hammarling, Nicholas J. Higham, Craig Lucas, School of Mathematics, University of Manchester
Definition at line 257 of file zhetrf_rk.f.
subroutine zsytrf_rk (character uplo, integer n, complex*16, dimension( lda,
* ) a, integer lda, complex*16, dimension( * ) e, integer, dimension( *
) ipiv, complex*16, dimension( * ) work, integer lwork, integer info)
ZSYTRF_RK computes the factorization of a complex symmetric indefinite matrix using the bounded Bunch-Kaufman (rook) diagonal pivoting method (BLAS3 blocked algorithm).
π― Purpose
ZSYTRF_RK computes the factorization of a complex symmetric matrix A using the bounded Bunch-Kaufman (rook) diagonal pivoting method:
A = P*U*D*(U**T)*(P**T) or A = P*L*D*(L**T)*(P**T)
where U (or L) is unit upper (or lower) triangular matrix, U**T (or L**T) is the transpose of U (or L), P is a permutation matrix, P**T is the transpose of P, and D is symmetric and block diagonal with 1-by-1 and 2-by-2 diagonal blocks.
This is the blocked version of the algorithm, calling Level 3 BLAS. For more information see Further Details section.
π Parameters
UPLO β UPLO is CHARACTER*1. Specifies whether the upper or lower triangular part of the symmetric matrix A is stored:
= 'U': Upper triangular= 'L': Lower triangular
N β N is INTEGER. The order of the matrix A. N >= 0.
A β A is COMPLEX*16 array, dimension (LDA,N). On entry, the symmetric matrix A.
- If
UPLO = 'U': the leadingN-by-Nupper triangular part ofAcontains the upper triangular part of the matrixA, and the strictly lower triangular part ofAis not referenced. - If
UPLO = 'L': the leadingN-by-Nlower triangular part ofAcontains the lower triangular part of the matrixA, and the strictly upper triangular part ofAis not referenced.
On exit, contains:
- a) ONLY diagonal elements of the symmetric block diagonal matrix
Don the diagonal ofA, i.e.D(k,k) = A(k,k); (superdiagonal (or subdiagonal) elements ofDare stored on exit in arrayE), and - b) If
UPLO = 'U': factorUin the superdiagonal part ofA. IfUPLO = 'L': factorLin the subdiagonal part ofA.
LDA β LDA is INTEGER. The leading dimension of the array A. LDA >= max(1,N).
E β E is COMPLEX*16 array, dimension (N). On exit, contains the superdiagonal (or subdiagonal) elements of the symmetric block diagonal matrix D with 1-by-1 or 2-by-2 diagonal blocks, where:
- If
UPLO = 'U':E(i) = D(i-1,i),i=2:N,E(1)is set to0; - If
UPLO = 'L':E(i) = D(i+1,i),i=1:N-1,E(N)is set to0.
π NOTE: For a1-by-1diagonal blockD(k), where1 <= k <= N, the elementE(k)is set to0in both theUPLO = 'U'orUPLO = 'L'cases.
IPIV
okay so IPIV is an INTEGER array, dimension (N), and it's literally the receipts for the permutation matrix P in the factorization of matrix A. Here's the tea:
- the absolute value of
IPIV(k)is the index of the row and column that were interchanged with thek-th row and column. - the value of
UPLOdescribes the order in which those interchanges were applied. - the sign of
IPIVrepresents the block structure of the symmetric block diagonal matrixDwith1-by-1or2-by-2diagonal blocks which correspond to1or2interchanges at each factorization step. For more info see the Further Details section, bestie.
If UPLO = 'U' (in factorization order, k decreases from N to 1):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a1-by-1diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N). - If
IPIV(k) = k, no interchange occurred.
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k-1) < 0means:D(k-1:k,k-1:k)is a2-by-2diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs).
- If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - If
-IPIV(k-1) != k-1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k-1) = k-1, no interchange occurred.
- c) In both cases a) and b), always
ABS( IPIV(k) ) <= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
If UPLO = 'L' (in factorization order, k increases from 1 to N):
- a) A single positive entry
IPIV(k) > 0means:D(k,k)is a1-by-1diagonal block.- If
IPIV(k) != k, rows and columnskandIPIV(k)were interchanged in the matrixA(1:N,1:N). - If
IPIV(k) = k, no interchange occurred.
- b) A pair of consecutive negative entries
IPIV(k) < 0andIPIV(k+1) < 0means:D(k:k+1,k:k+1)is a2-by-2diagonal block.- (NOTE: negative entries in
IPIVappear ONLY in pairs).
- If
-IPIV(k) != k, rows and columnskand-IPIV(k)were interchanged in the matrixA(1:N,1:N). If-IPIV(k) = k, no interchange occurred. - If
-IPIV(k+1) != k+1, rows and columnsk-1and-IPIV(k-1)were interchanged in the matrixA(1:N,1:N). If-IPIV(k+1) = k+1, no interchange occurred. (yes, the doc literally saysk-1and-IPIV(k-1)in this sub-bullet β keep it exact, no cap.)
- c) In both cases a) and b), always
ABS( IPIV(k) ) >= k. - d) NOTE: Any entry
IPIV(k)is always NONZERO on output.
π§° WORK
WORK is a COMPLEX*16 array, dimension ( MAX(1,LWORK) ).
On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
π LWORK
LWORK is INTEGER.
It's the length of WORK. LWORK >=1. For best performance, LWORK >= N*NB, where NB is the block size returned by ILAENV.
If LWORK = -1, then a workspace query is assumed; the routine only calculates the optimal size of the WORK array, returns this value as the first entry of the WORK array, and no error message related to LWORK is issued by XERBLA.
π¨ INFO
INFO is INTEGER.
= 0: successful exit. slay.< 0: IfINFO = -k, thek-th argument had an illegal value.> 0: IfINFO = k, the matrixAis singular, because:- If
UPLO = 'U': columnkin the upper triangular part ofAcontains all zeros. - If
UPLO = 'L': columnkin the lower triangular part ofAcontains all zeros. - Therefore
D(k,k)is exactly zero, and superdiagonal elements of columnkofU(or subdiagonal elements of columnkofL )are all zeros. The factorization has been completed, but the block diagonal matrixDis exactly singular, and division by zero will occur if it is used to solve a system of equations. - NOTE:
INFOonly stores the first occurrence of a singularity, any subsequent occurrence of singularity is not stored inINFOeven though the factorization always completes.
- If
π Author
- Univ. of Tennessee
- Univ. of California Berkeley
- Univ. of Colorado Denver
- NAG Ltd.
π Further Details:
!> TODO: put correct description
!>
π Contributors:
- December 2016, Igor Kozachenko, Computer Science Division, University of California, Berkeley
- September 2007, Sven Hammarling, Nicholas J. Higham, Craig Lucas, School of Mathematics, University of Manchester
π Definition at line 257 of file zsytrf_rk.f.
π€ Author
Generated automatically by Doxygen for LAPACK from the source code.
π that's the whole page, bestie! π Footer vibe check: this is from LAPACK Version 3.12.0, page hetrf_rk(3), generated from the source code β the footer doesn't print a separate date, so the version is the timestamp we've got. mwah π
π¦ where this page comes from
- arch rolling: lapack-doc 3.12.1-2 Β· custom
- π girlified by deepseek-ai/DeepSeek-V4.1-Flash-1017 on : a language model's rewrite of the original, under its terms