SLAHILB generates an N by N scaled Hilbert matrix in A along with


 NRHS right-hand sides in B and solutions in X such that A*X=B.


 The Hilbert matrix is scaled by M = LCM(1, 2, ..., 2*N-1) so that all


 entries are integers.  The right-hand sides are the first NRHS


 columns of M * the identity matrix, and the solutions are the


 first NRHS columns of the inverse Hilbert matrix.


 The condition number of the Hilbert matrix grows exponentially with


 its size, roughly as O(e ** (3.5*N)).  Additionally, the inverse


 Hilbert matrices beyond a relatively small dimension cannot be


 generated exactly without extra precision.  Precision is exhausted


 when the largest entry in the inverse Hilbert matrix is greater than


 2 to the power of the number of bits in the fraction of the data type


 used plus one, which is 24 for single precision.


 In single, the generated solution is exact for N <= 6 and has


 small componentwise error for 7 <= N <= 11.

N

          N is INTEGER


          The dimension of the matrix A.

NRHS

          NRHS is INTEGER


          The requested number of right-hand sides.

A

          A is REAL array, dimension (LDA, N)


          The generated scaled Hilbert matrix.

LDA

          LDA is INTEGER


          The leading dimension of the array A.  LDA >= N.

X

          X is REAL array, dimension (LDX, NRHS)


          The generated exact solutions.  Currently, the first NRHS


          columns of the inverse Hilbert matrix.

LDX

          LDX is INTEGER


          The leading dimension of the array X.  LDX >= N.

B

          B is REAL array, dimension (LDB, NRHS)


          The generated right-hand sides.  Currently, the first NRHS


          columns of LCM(1, 2, ..., 2*N-1) * the identity matrix.

LDB

          LDB is INTEGER


          The leading dimension of the array B.  LDB >= N.

WORK

          WORK is REAL array, dimension (N)

INFO

          INFO is INTEGER


          = 0: successful exit


          = 1: N is too large; the data is still generated but may not


               be not exact.


          < 0: if INFO = -i, the i-th argument had an illegal value

Univ. of Tennessee

Univ. of California Berkeley

Univ. of Colorado Denver

NAG Ltd.