TESTING/EIG/sbdt01.f(3)
π NAME
TESTING/EIG/sbdt01.f β okay bestie, this page lives in the Library Functions Manual under TESTING/EIG/sbdt01.f(3), and itβs all about the SBDT01 subroutine. Sheβs a testing routine, and sheβs doing her little reconstruction vibe check. π
β¨ SYNOPSIS β¨
π«Ά Functions/Subroutines
subroutine sbdt01 (m, n, kd, a, lda, q, ldq, d, e, pt, ldpt, work,
resid)
SBDT01
π§ Function/Subroutine Documentation
subroutine sbdt01 (integer m, integer n, integer kd, real, dimension( lda, *
) a, integer lda, real, dimension( ldq, * ) q, integer ldq, real,
dimension( * ) d, real, dimension( * ) e, real, dimension( ldpt, * ) pt,
integer ldpt, real, dimension( * ) work, real resid)
SBDT01
π¦ Purpose
So, SBDT01 reconstructs a general matrix A from its bidiagonal form:
A = Q * B * P**T
where Q (m by min(m,n)) and P**T (min(m,n) by n) are orthogonal matrices and B is bidiagonal. Literally the glow-up from bidiagonal back to general, no cap. β¨
The test ratio that checks the reduction is:
RESID = norm(A - Q * B * P**T) / ( n * norm(A) * EPS )
where EPS is the machine precision. π
πΈ Parameters
MβMisINTEGER. The number of rows of the matricesAandQ.NβNisINTEGER. The number of columns of the matricesAandP**T.KDβKDisINTEGER. IfKD = 0,Bis diagonal and the arrayEis not referenced. IfKD = 1, the reduction was performed byxGEBRD;Bis upper bidiagonal ifM >= N, and lower bidiagonal ifM < N. IfKD = -1, the reduction was performed byxGBBRD;Bis always upper bidiagonal.AβAisREAL array, dimension (LDA,N). ThembynmatrixA.LDAβLDAisINTEGER. The leading dimension of the arrayA.LDA >= max(1,M).QβQisREAL array, dimension (LDQ,N). Thembymin(m,n)orthogonal matrixQin the reductionA = Q * B * P**T.LDQβLDQisINTEGER. The leading dimension of the arrayQ.LDQ >= max(1,M).DβDisREAL array, dimension (min(M,N)). The diagonal elements of the bidiagonal matrixB.EβEisREAL array, dimension (min(M,N)-1). The superdiagonal elements of the bidiagonal matrixBifm >= n, or the subdiagonal elements ofBifm < n.PTβPTisREAL array, dimension (LDPT,N). Themin(m,n)bynorthogonal matrixP**Tin the reductionA = Q * B * P**T.LDPTβLDPTisINTEGER. The leading dimension of the arrayPT.LDPT >= max(1,min(M,N)).WORKβWORKisREAL array, dimension (M+N).RESIDβRESIDisREAL. The test ratio:norm(A - Q * B * P**T) / ( n * norm(A) * EPS ).
π Author
- Univ. of Tennessee
- Univ. of California Berkeley
- Univ. of Colorado Denver
- NAG Ltd.
Definition at line 139 of file sbdt01.f.
π Author
Generated automatically by Doxygen for LAPACK from the source code.
mwah π β source/version/date tea: this page is from LAPACK Version 3.12.0, the footer line is for TESTING/EIG/sbdt01.f(3), and thereβs literally no date printed on that footer line. Periodt. 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