SDBTRF(3)
π SDBTRF(3) β bestie edition π
okay so. you're here because you need an LU factorization of a real m-by-n band matrix A, and you specifically do not want partial pivoting or row interchanges happening behind your back. no cap. SDBTRF has got you. π«Ά
β¨ NAME β¨
SDBTRF β compute an LU factorization of a real m-by-n band matrix A without using partial pivoting or row interchanges.
πΈ SYNOPSIS πΈ
SUBROUTINE SDBTRF( M, N, KL, KU, AB, LDAB, INFO )
INTEGER INFO, KL, KU, LDAB, M, N
REAL AB( LDAB, * )
π¦ PURPOSE π¦
Sdbtrf computes an LU factorization of a real m-by-n band matrix A without using partial pivoting or row interchanges. that's the whole job, and she does it. π
π DESCRIPTION π
This is the blocked version of the algorithm, calling Level 3 BLAS. so like, the fast one. the one that actually respects your cache. icon behaviour fr fr. π₯
π ARGUMENTS π
M(input)INTEGERβ the number of rows of the matrixA.M >= 0.N(input)INTEGERβ the number of columns of the matrixA.N >= 0.KL(input)INTEGERβ the number of subdiagonals within the band ofA.KL >= 0.KU(input)INTEGERβ the number of superdiagonals within the band ofA.KU >= 0.AB(input/output)REALarray, dimension(LDAB,N)β okay this one's the main character, let's break it down.
On entry, this holds the matrix A in band storage, sitting in rows KL+1 to 2*KL+KU+1. rows 1 to KL of the array? they need not be set. do not stress about them. π
The j-th column of A is stored in the j-th column of the array AB like this:
AB(kl+ku+1+i-j,j) = A(i,j) for max(1,j-ku)<=i<=min(m,j+kl)
On exit, you get the details of the factorization, and here's the tea β: U is stored as an upper triangular band matrix with KL+KU superdiagonals in rows 1 to KL+KU+1, and the multipliers used during the factorization are stored in rows KL+KU+2 to 2*KL+KU+1. see below for further details, bestie, I'm not gatekeeping.
LDAB(input)INTEGERβ the leading dimension of the arrayAB.LDAB >= 2*KL+KU+1. she said what she said. πINFO(output)INTEGERβ the vibe report. three possible moods:= 0: successful exit. slay. β¨< 0: ifINFO = -i, the i-th argument had an illegal value. somebody typed something wrong, and it was not me. π> 0: ifINFO = +i,U(i,i)is exactly zero. The factorization has been completed, but the factorUis exactly singular, and division by zero will occur if it is used to solve a system of equations. so like... she finished, but she's singular, and dividing by zero is a whole situation. π
π THE BAND STORAGE SCHEME π
here's the picture, when M = N = 6, KL = 2, KU = 1:
On entry: On exit:
* a12 a23 a34 a45 a56 * u12 u23 u34 u45 u56
a11 a22 a33 a44 a55 a66 u11 u22 u33 u44 u55 u66
a21 a32 a43 a54 a65 * m21 m32 m43 m54 m65 *
a31 a42 a53 a64 * * m31 m42 m53 m64 * *
Array elements marked * are not used by the routine. ignore them, they're just standing there. π§
and that's the whole page, periodt. π this one comes out of the library's manual set, with the version and date tucked down at the very bottom of the original like every good manpage keeps them β so mwah, go factor something banded, and remember: no pivoting, no row swaps, no drama. πΈπ π¦
π¦ where this page comes from
- ubuntu jammy: scalapack-doc 1.5-11
- π girlified by deepseek-ai/DeepSeek-V4.1-Flash on : a language model's rewrite of the original, under its terms