Hello,

I have a python written code, and it uses svd from numpy. And I am trying to port the same code in Julia.

The problem I am facing is that after some point, even though I read the same files for both programming languages, svd method in two languages decomposes the same matrix differently. Hence, solution changes and I cannot replicate the experiment.

I have multiple svd usages inside the code, the first two gives nearly the exactly decomposition. (Everything is in Float32) But after the third svd function call my results and the original(python) code starts to differ. The only difference from the first 2 svd functions is that the third one removes the ` full_matrices=False`

flag in Python code.

```
u, s, vt = np.linalg.svd(xsim[:idx], full_matrices=False) # get the same result
u, s, vt = np.linalg.svd(zsim[:idx], full_matrices=False) # get the same result
u, s, vt = np.linalg.svd(z[trg_indices].T.dot(x[src_indices])) # third call to svd; where things get messy.
```

On the other hand, in Julia, even thoug I changed the algorithm to QRIteration , and full to true. I am not getting the same results.

I did some research and both numpy and Julia uses LAPACK package underneath. Then what should be the correct way to get the same results ?

This is what I get when I write `print(numpy.show_config())`

```
blas_mkl_info:
NOT AVAILABLE
blis_info:
NOT AVAILABLE
openblas_info:
libraries = ['openblas', 'openblas']
library_dirs = ['/usr/local/lib']
language = c
define_macros = [('HAVE_CBLAS', None)]
blas_opt_info:
libraries = ['openblas', 'openblas']
library_dirs = ['/usr/local/lib']
language = c
define_macros = [('HAVE_CBLAS', None)]
lapack_mkl_info:
NOT AVAILABLE
openblas_lapack_info:
libraries = ['openblas', 'openblas']
library_dirs = ['/usr/local/lib']
language = c
define_macros = [('HAVE_CBLAS', None)]
lapack_opt_info:
libraries = ['openblas', 'openblas']
library_dirs = ['/usr/local/lib']
language = c
define_macros = [('HAVE_CBLAS', None)]
```

For julia could someone please help me if all set is the same as Python ?

B.R.