Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ImportError: libmklml_intel.so: cannot open shared object file: No such file or directory #3213

Closed
helinwang opened this issue Aug 3, 2017 · 6 comments · Fixed by #3461
Closed
Assignees

Comments

@helinwang
Copy link
Contributor

helinwang commented Aug 3, 2017

Just tried our latest production docker image, and got this error. I think we need to include the generated libmklml_intel.so and libiomp5.so into the Python package (or link into paddle statically).

Traceback (most recent call last):
  File "go/pserver/client/c/test/test_train.py", line 77, in <module>
    main()
  File "go/pserver/client/c/test/test_train.py", line 15, in main
    paddle.init(use_gpu=False, trainer_count=1)
  File "/usr/local/lib/python2.7/dist-packages/paddle/v2/__init__.py", line 66, in init
    import py_paddle.swig_paddle as api
  File "/usr/local/lib/python2.7/dist-packages/py_paddle/__init__.py", line 15, in <module>
    from util import DataProviderWrapperConverter
  File "/usr/local/lib/python2.7/dist-packages/py_paddle/util.py", line 18, in <module>
    import swig_paddle
  File "/usr/local/lib/python2.7/dist-packages/py_paddle/swig_paddle.py", line 28, in <module>
    _swig_paddle = swig_import_helper()
  File "/usr/local/lib/python2.7/dist-packages/py_paddle/swig_paddle.py", line 24, in swig_import_helper
    _mod = imp.load_module('_swig_paddle', fp, pathname, description)
ImportError: libmklml_intel.so: cannot open shared object file: No such file or directory

...

ImportError: libiomp5.so: cannot open shared object file: No such file or directory
@gangliao
Copy link
Contributor

gangliao commented Aug 4, 2017

It looks Intel's PR default enabled MKL_DNN feature. But they didn't package the corresponding libraries libmklml_intel and libiomp5 into setup.py.in

@gangliao
Copy link
Contributor

gangliao commented Aug 4, 2017

@helinwang @luotao1

@tensor-tang
Copy link
Contributor

tensor-tang commented Aug 4, 2017

@gangliao Yes, It should be the problem about the dynamic libraries.

Actually we found the setup.py.in before, but did not find an appropriate way to add libmklml_intel and libiomp5.

Like cudnn and mkl, which are supposed installed on local machine, MKLML are kind of like the same level as them. So we should found a better way to include MKLML and MKLDNN as third party.

I think you can comment in the issue #3162, since they are the same issue.
Any method is very appreciated.

@Yancey1989
Copy link
Contributor

I have the same problem with the latest Docker image, maybe #3162 does not fix this error?

@luotao1
Copy link
Contributor

luotao1 commented Aug 10, 2017

Since -WITH_MKLML and -WITH_MKLDNN are opened later, I will set off again. @Yancey1989

@Joseph996
Copy link

So, any solutions to this problem?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

6 participants