user1591487
user1591487

Reputation: 425

python matplotlib MKL fatal error on ubuntu 12.04

Here's my python code

1 import matplotlib.pyplot as plt                                                 
2 plt.plot([1, 2, 3, 4])                                                          
3 plt.ylabel('some numbers')                                                      
4 plt.show()                                                                      

When I run it, I got the following error message

$ python test.py 

*** libmkl_avx.so *** failed with error : /usr/local/enthought/lib/libmkl_avx.so: undefined symbol: i_free
*** libmkl_def.so *** failed with error : /usr/local/enthought/lib/libmkl_def.so: undefined symbol: i_free
MKL FATAL ERROR: Cannot load neither libmkl_avx.so nor libmkl_def.so

Here's my python version.

$ python --version
Python 2.7.2 -- EPD 7.2-2 (64-bit)

Upvotes: 2

Views: 2333

Answers (1)

brettb
brettb

Reputation: 809

I realize this is a pretty old question, but I was having the same issue and found a solution without needing to reinstall, so for any future readers here is what I did:

@root commented above with a link to the intel forum. One of the answers there suggested preloading some of the MKL libraries. (dated Tue, 10/20/2009 - 10:32, I can't figure out how to link to specific comment on there).

This solved the issue for me. Though, you'll need to find where tthe libmkl_core.so and libmkl_sequential.so libraries are stored for you. For me, it was:

export LD_PRELOAD=/usr/local/Canopy/v152/CanopyCore/appdata/canopy-1.5.2.2785.rh5-x86_64/lib/libmkl_sequential.so:/usr/local/Canopy/v152/CanopyCore/appdata/canopy-1.5.2.2785.rh5-x86_64/lib/libmkl_core.so

There were also copies in /usr/local/Canopy/v152/Canopy_64bit/User/lib/, but these versions did NOT work. Only the one above.

Hope this helps someone out in the future!

Upvotes: 1

Related Questions