User`s guide

Upgrading from an Earlier Release
2-27
Source Control
If you use Microsoft Visual SourceSafe with MATLAB’s source control features,
you now need to specify the login information for SourceSafe using preferences.
Select
File -> Preferences -> General -> Source Control from the desktop.
Specify the
Username, Password, and Database.
Mathematics Issues
Finding Smallest Magnitude Eigenvalues
eigs(A,k,sigma) and eigs(A,B,k,sigma) return k eigenvalues based on
sigma. For sigma = 'sm', eigs returns the smallest magnitude eigenvalues.
In MATLAB 6.0,
eigs was reimplemented to use the ARPACK library of
routines. Unfortunately, the smallest magnitude case,
sigma = 'sm' and
sigma = 0, chose the wrong algorithm. For MATLAB 6.1, the correct ARPACK
algorithm is used and convergence is much quicker.
This bug fix introduces a backwards incompatibility. When
A is a function Afun
and
sigma = 'sm', Afun must now return Y=A\x. Prior to MATLAB 6.1, eigs
required
Afun to return y = A*x for this case.
Possible Changes in Results Returned by Matrix Functions
Starting in MATLAB 6.0 (R12.0), matrix computations are based on LAPACK,
a large, multiauthor Fortran subroutine library for numerical linear algebra.
While this change has many benefits (see the Release Notes for Release 12) and
matrix functions continue to operate in the same way in MATLAB 6.0, the
results returned by matrix functions may differ. Changes in roundoff errors
can be seen in most matrix computations. In cases where quantities are not
uniquely determined mathematically, results may differ in order and in
normalization.
For example:
•Eigenvalues may be returned in a different order.
•Eigenvectors may be normalized differently.
•The signs of columns of orthogonal matrices may differ.
•
rcond is a better estimate of the reciprocal condition.
•
lu can now be used to factor rectangular full matrices.