Kirchhoff formulation fails at loading Intel MKL libraries

[BartelsJ]

When switching from angles to the kirchhoff formulation at network.lopf (several scenarios and data set versions), I receive a 'Kernel died, restarting' remark from spyder and following information at my terminal:
Intel MKL FATAL ERROR: Cannot load libmkl_avx512.so or libmkl_def.so.
(mkl is installed at version 2017.0.3)

As suggested here, the example command line
'python -c 'import sklearn.linear_model.tests.test_randomized_l1'
runs withour errors. Hence, anaconda seems to run correctly.

Any experiences on that error? Are you using a different mkl version or does it run without problems?

[wolfbunke]

Does this issue needs a Milestone?

[BartelsJ]

No. I guess this is still an open issue but the functionality of eTraGo is not endangered.