Solved: Different results in SVD by using Igor and LabVIEW

[D-Kid]

Hi NI Community,

I try to solve a specific issue. However, I am out of ideas and I am convinced that I find here the necessary help 🙂

At the moment I am going to translate an algorithm from IGOR to LabVIEW for technical reasons.

In the IGOR source code there is a function called "MatrixSVD". This function is also available in LabVIEW.
Regretly, after putting a six by six matrix into these functions I receive different values for the U, VT vector as well as the S vector in IGOR and LabVIEW respectivly (see attachment)

Just the first column of the U matrix and the first row of the VT vector are the same as in IGOR.
Did you had some ideas why it happens to such a behaviour?

Thanks in advance

Regards

D-Kid

[D-Kid · Accepted solution]

Problem solved:

the main problem was that the input matrix has to be transposed. A fact which I did not thought about.

After doing this the SVD results into the same values as IGOR does.