Purpose
To compute either the upper or lower triangular part of one of the
matrix formulas
_
R = alpha*R + beta*op( H )*B, (1)
_
R = alpha*R + beta*B*op( H ), (2)
_
where alpha and beta are scalars, H, B, R, and R are m-by-m
matrices, H is an upper Hessenberg matrix, and op( H ) is one of
op( H ) = H or op( H ) = H', the transpose of H.
The result is overwritten on R.
Specification
SUBROUTINE MB01RY( SIDE, UPLO, TRANS, M, ALPHA, BETA, R, LDR, H,
$ LDH, B, LDB, DWORK, INFO )
C .. Scalar Arguments ..
CHARACTER SIDE, TRANS, UPLO
INTEGER INFO, LDB, LDH, LDR, M
DOUBLE PRECISION ALPHA, BETA
C .. Array Arguments ..
DOUBLE PRECISION B(LDB,*), DWORK(*), H(LDH,*), R(LDR,*)
Arguments
Mode Parameters
SIDE CHARACTER*1
Specifies whether the Hessenberg matrix H appears on the
left or right in the matrix product as follows:
_
= 'L': R = alpha*R + beta*op( H )*B;
_
= 'R': R = alpha*R + beta*B*op( H ).
UPLO CHARACTER*1 _
Specifies which triangles of the matrices R and R are
computed and given, respectively, as follows:
= 'U': the upper triangular part;
= 'L': the lower triangular part.
TRANS CHARACTER*1
Specifies the form of op( H ) to be used in the matrix
multiplication as follows:
= 'N': op( H ) = H;
= 'T': op( H ) = H';
= 'C': op( H ) = H'.
Input/Output Parameters
M (input) INTEGER _
The order of the matrices R, R, H and B. M >= 0.
ALPHA (input) DOUBLE PRECISION
The scalar alpha. When alpha is zero then R need not be
set before entry.
BETA (input) DOUBLE PRECISION
The scalar beta. When beta is zero then H and B are not
referenced.
R (input/output) DOUBLE PRECISION array, dimension (LDR,M)
On entry with UPLO = 'U', the leading M-by-M upper
triangular part of this array must contain the upper
triangular part of the matrix R; the strictly lower
triangular part of the array is not referenced.
On entry with UPLO = 'L', the leading M-by-M lower
triangular part of this array must contain the lower
triangular part of the matrix R; the strictly upper
triangular part of the array is not referenced.
On exit, the leading M-by-M upper triangular part (if
UPLO = 'U'), or lower triangular part (if UPLO = 'L') of
this array contains the corresponding triangular part of
_
the computed matrix R.
LDR INTEGER
The leading dimension of array R. LDR >= MAX(1,M).
H (input) DOUBLE PRECISION array, dimension (LDH,M)
On entry, the leading M-by-M upper Hessenberg part of
this array must contain the upper Hessenberg part of the
matrix H.
The elements below the subdiagonal are not referenced,
except possibly for those in the first column, which
could be overwritten, but are restored on exit.
LDH INTEGER
The leading dimension of array H. LDH >= MAX(1,M).
B (input) DOUBLE PRECISION array, dimension (LDB,M)
On entry, the leading M-by-M part of this array must
contain the matrix B.
LDB INTEGER
The leading dimension of array B. LDB >= MAX(1,M).
Workspace
DWORK DOUBLE PRECISION array, dimension (LDWORK)
LDWORK >= M, if beta <> 0 and SIDE = 'L';
LDWORK >= 0, if beta = 0 or SIDE = 'R'.
This array is not referenced when beta = 0 or SIDE = 'R'.
Error Indicator
INFO INTEGER
= 0: successful exit;
< 0: if INFO = -i, the i-th argument had an illegal
value.
Method
The matrix expression is efficiently evaluated taking the Hessenberg/triangular structure into account. BLAS 2 operations are used. A block algorithm can be constructed; it can use BLAS 3 GEMM operations for most computations, and calls of this BLAS 2 algorithm for computing the triangles.Further Comments
The main application of this routine is when the result should be a symmetric matrix, e.g., when B = X*op( H )', for (1), or B = op( H )'*X, for (2), where B is already available and X = X'.Example
Program Text
NoneProgram Data
NoneProgram Results
None