Re: Interfacing Ada with C

[Ada novice <posts@gmx.us>]

I have made some simple tests on some special matrices which can pose
some problems. I limited myself to matrices with real values. I have
tested:

1. non-symmetric real matrices with repeated values.
   No problem here

2. a real orthogonal matrix. Such a matrix gives all eigenvalues of
unit modulus. I tested the case from http://en.wikipedia.org/wiki/Orthogonal_matrix.
In Matlab,

   eig([0, -0.8, -0.6; 0.8, -0.36, 0.48; 0.6, 0.48, -0.64])

   gives the eigenvalues as 0 + 1i, 0 - 1i and -1.

   Using your Ada code gives:

   -1.78814E-08  1.00000E+00
   -8.72890E-09 -1.00000E+00
   -1.00000E+00  0.00000E+00


Does the Blas code work with Fortran double-precision? In the above,
Matlab with double (15 digits precision). I don't see you using the
type Long_Float in the Ada code. I don't know how many digits are
being passed from the Fortran result to Ada. Maybe this is causing the
discrepancy between the expected results (all eigenvalues of modulus
one) and the output in Ada.


3. a real skew-symmetric matrix. This is also interesting as all
eigenvalues are imaginary. I tested the case from
http://en.wikipedia.org/wiki/Skew-symmetric_matrix. In Matlab,

   eig([0, 2, -1; -2, 0, -4; 1, 4, 0])

   gives the eigenvalues as   0 ;  0 + 4.582575694955841i and 0 -
4.582575694955841i

   Using the Ada code gives:

    9.12635E-08  4.58258E+00
   -5.31220E-08  1.92730E-07
    3.95880E-08 -4.58258E+00

    Here also exist some slight discrepancies. I believe as mentioned
above that if we change use more precision digits, then we'll get
"better" answers.


Thanks.

YC