HP UX Performance Tools manual 545

Jagged diagonal format triangular solve SJADSM/DJADSM/CJADSM/ZJADSM

Name SJADSM/DJADSM/CJADSM/ZJADSM

Jagged diagonal format triangular solve

Purpose Jagged diagonal format triangular solve. Given a scalar α, an upper- or lower-triangular sparse matrix A, and a m-by-nmatrix B, these subprograms compute either of the matrix solutions αA–1B, or αDA–1B, or αA–1DB, where D

is a diagonal matrix. The size of A is m-by-m. Optionally, A–1may be replaced by A–T, or by A–*. Here, A–Tis the transpose-inverse and A–* is the conjugate-transpose-inverse of A. The solution matrix may be stored in the result matrix C or optionally may be added to or subtracted from it. This is handled in a convenient, but general way by two scalar arguments, α and β, which are used as multipliers of the solution matrix and the result matrix. Specifically, these subprograms compute matrix solutions of the form

C ← αA–1B + βC	C ← αA–TB + βC	C ← αA–*B + βC
C ← αDA–1B + βC C ← αDA–TB + βC		C ← αDA–*B + βC
C ← αA–1DB + βC	C ← αA–TDB + βC	C ← αA–*DB + βC

Usage	VECLIB:
	SUBROUTINE	SJADSM
	INTEGER*4	transa, m, n, unitd, maxnz, ldb, ldc, lwork
	INTEGER*4	descra(*), indx(*), pntr(*), iperm(*)
	REAL*4	alpha, beta
	REAL*4	val(*), b(ldb,*), c(ldc,*), work(*)

CALL SJADSM (transa, m, n, unitd, dv, alpha, descra, val, pntr, iperm, indx, maxnz, b, ldb, beta, c, ldc, work, lwork)

SUBROUTINE	DJADSM
INTEGER*4	transa, m, n, unitd, maxnz, ldb, ldc, lwork
INTEGER*4	descra(*), indx(*), pntr(*), iperm(*)
REAL*8	alpha, beta
REAL*8	val(*), b(ldb,*), c(ldc,*), work(*)

CALL DJADSM (transa, m, n, unitd, dv, alpha, descra, val, pntr, iperm, indx, maxnz, b, ldb, beta, c, ldc, work, lwork)

SUBROUTINE	CJADSM
INTEGER*4	transa, m, n, unitd, maxnz, ldb, ldc, lwork
INTEGER*4	descra(*), indx(*), pntr(*), iperm(*)
COMPLEX*8	alpha, beta
COMPLEX*8	val(*), b(ldb,*), c(ldc,*), work(*)

CALL CJADSM (transa, m, n, unitd, dv, alpha, descra, val, pntr, iperm, indx, maxnz, b, ldb, beta, c, ldc, work, lwork)

Chapter 4 Sparse BLAS Operations 517