DSTT22  checks a set of M eigenvalues and eigenvectors,


     A U = U S


 where A is symmetric tridiagonal, the columns of U are orthogonal,


 and S is diagonal (if KBAND=0) or symmetric tridiagonal (if KBAND=1).


 Two tests are performed:


    RESULT(1) = | U' A U - S | / ( |A| m ulp )


    RESULT(2) = | I - U'U | / ( m ulp )

N

          N is INTEGER


          The size of the matrix.  If it is zero, DSTT22 does nothing.


          It must be at least zero.

M

          M is INTEGER


          The number of eigenpairs to check.  If it is zero, DSTT22


          does nothing.  It must be at least zero.

KBAND

          KBAND is INTEGER


          The bandwidth of the matrix S.  It may only be zero or one.


          If zero, then S is diagonal, and SE is not referenced.  If


          one, then S is symmetric tri-diagonal.

AD

          AD is DOUBLE PRECISION array, dimension (N)


          The diagonal of the original (unfactored) matrix A.  A is


          assumed to be symmetric tridiagonal.

AE

          AE is DOUBLE PRECISION array, dimension (N)


          The off-diagonal of the original (unfactored) matrix A.  A


          is assumed to be symmetric tridiagonal.  AE(1) is ignored,


          AE(2) is the (1,2) and (2,1) element, etc.

SD

          SD is DOUBLE PRECISION array, dimension (N)


          The diagonal of the (symmetric tri-) diagonal matrix S.

SE

          SE is DOUBLE PRECISION array, dimension (N)


          The off-diagonal of the (symmetric tri-) diagonal matrix S.


          Not referenced if KBSND=0.  If KBAND=1, then AE(1) is


          ignored, SE(2) is the (1,2) and (2,1) element, etc.

U

          U is DOUBLE PRECISION array, dimension (LDU, N)


          The orthogonal matrix in the decomposition.

LDU

          LDU is INTEGER


          The leading dimension of U.  LDU must be at least N.

WORK

          WORK is DOUBLE PRECISION array, dimension (LDWORK, M+1)

LDWORK

          LDWORK is INTEGER


          The leading dimension of WORK.  LDWORK must be at least


          max(1,M).

RESULT

          RESULT is DOUBLE PRECISION array, dimension (2)


          The values computed by the two tests described above.  The


          values are currently limited to 1/ulp, to avoid overflow.

Univ. of Tennessee

Univ. of California Berkeley

Univ. of Colorado Denver

NAG Ltd.