In my code, I am doing some vector subtractions in the form of
t = t - alpha * v
where t and v are vectors and alpha is a scalar. I am using daxpy for this however for some problems the norm of the resulting vector t, differs significantly than the value of the norm computed by MATLAB. Well, I tried almost everything I know to overcome the problem, but there is no success so far. What could be the reason of this problem?
Any pointers are appreciated.