[Numpy-discussion] Instaling numpy without root access

Ignat Harczuk ignathe at gmail.com
Fri Oct 17 04:22:35 EDT 2014


Have you considered virtual environments?

http://docs.python-guide.org/en/latest/dev/virtualenvs/


Inside of each environment you can build a local python version and
packages with different versions through pip.
Maybe not exactly what you need help with but it is a good tool to have so
that you have less dependency issues.

On Mon, Oct 13, 2014 at 2:52 AM, Lahiru Samarakoon <lahiruts at gmail.com>
wrote:

> Guys, any advice is highly appreciated. I am a little new to building in
> Linux.
> Thanks,
> Lahiru
>
> On Sat, Oct 11, 2014 at 9:43 AM, Lahiru Samarakoon <lahiruts at gmail.com>
> wrote:
>
>> I switched to numpy-1.8.2. . Now getting following error. I am using
>> LAPACK that comes with atlast installation. Can this be a problem?
>>
>> Traceback (most recent call last):
>>   File "<stdin>", line 1, in <module>
>>   File
>> "/home/svu/a0095654/.local/lib/python2.7/site-packages/numpy/__init__.py",
>> line 170, in <module>
>>     from . import add_newdocs
>>   File
>> "/home/svu/a0095654/.local/lib/python2.7/site-packages/numpy/add_newdocs.py",
>> line 13, in <module>
>>     from numpy.lib import add_newdoc
>>   File
>> "/home/svu/a0095654/.local/lib/python2.7/site-packages/numpy/lib/__init__.py",
>> line 18, in <module>
>>     from .polynomial import *
>>   File
>> "/home/svu/a0095654/.local/lib/python2.7/site-packages/numpy/lib/polynomial.py",
>> line 19, in <module>
>>     from numpy.linalg import eigvals, lstsq, inv
>>   File
>> "/home/svu/a0095654/.local/lib/python2.7/site-packages/numpy/linalg/__init__.py",
>> line 51, in <module>
>>     from .linalg import *
>>   File
>> "/home/svu/a0095654/.local/lib/python2.7/site-packages/numpy/linalg/linalg.py",
>> line 29, in <module>
>>     from numpy.linalg import lapack_lite, _umath_linalg
>> ImportError:
>> /home/svu/a0095654/.local/lib/python2.7/site-packages/numpy/linalg/lapack_lite.so:
>> undefined symbol: zgesdd_
>>
>> On Sat, Oct 11, 2014 at 1:30 AM, Julian Taylor <
>> jtaylor.debian at googlemail.com> wrote:
>>
>>> On 10.10.2014 19:26, Lahiru Samarakoon wrote:
>>> > Red Hat Enterprise Linux release 5.8
>>> > gcc (GCC) 4.1.2
>>> >
>>> > I am also trying to install numpy 1.9.
>>>
>>> that is the broken platform, please try the master branch or the
>>> maintenance/1.9.x branch, those should work now.
>>>
>>> Are there volunteers to report this to redhat?
>>>
>>> >
>>> > On Sat, Oct 11, 2014 at 12:59 AM, Julian Taylor
>>> > <jtaylor.debian at googlemail.com <mailto:jtaylor.debian at googlemail.com>>
>>> > wrote:
>>> >
>>> >     On 10.10.2014 18:51, Lahiru Samarakoon wrote:
>>> >     > Dear all,
>>> >     >
>>> >     > I am trying to install numpy without root access. So I am
>>> building from
>>> >     > the source.  I have installed atlas which also has lapack with
>>> it.  I
>>> >     > changed the site.cfg file as given below
>>> >     >
>>> >     > [DEFAULT]
>>> >     > library_dirs = /home/svu/a0095654/ATLAS/build/lib
>>> >     > include_dirs = /home/svu/a0095654/ATLAS/build/include
>>> >     >
>>> >     >
>>> >     > However, I am getting a segmentation fault when importing numpy.
>>> >     >
>>> >     > Please advise. I also put the build log file at the end of the
>>> email if
>>> >     > necessary.
>>> >
>>> >
>>> >     Which platform are you working on? Which compiler version?
>>> >     We just solved a segfault on import on red hat 5 gcc 4.1.2. Very
>>> likely
>>> >     caused by a compiler bug. See
>>> https://github.com/numpy/numpy/issues/5163
>>> >
>>> >     The build log is complaining about your atlas being to small,
>>> possibly
>>> >     the installation is broken?
>>> >
>>> >
>>>
>>>
>>
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> http://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20141017/b5274452/attachment.html>


More information about the NumPy-Discussion mailing list