[Help-gsl] gsl_blas question

[A J]

Hi All:

I am doing the exact same computation using gsl_blas_dgemv in C++ and in
MATLAB. However, I am getting somewhat different numbers. gsl_blas_dgemv
reports the resultant vector as:

-13.3633
-4.1205
4.3350
2.0995
-0.5040
-0.7772e-15
0.7216e-15
0.1943e-15
0.3331e-15
-0.1110e-15

MATLAB at the same time reports:

-13.3633
-4.1205
4.3350
2.0995
-0.5040
-0.4720e-15
0.2600e-15
-0.6153e-15
0.2305e-15
-0.2282e-15

Though probably the difference is due to numerical errors and possible
difference in implementation of the numerical schemes in the two platforms,
it becomes important because these values are used in subsequent
calculations that could change the final answers considerably.

Can someone help me figure out what we could do in this situation? Any help
would be greatly appreciated.
Thanks
A