<div dir="ltr"><div><br></div>Hello all,<div><br></div><div>I am struggling with a python crash of an ipython notebook (kernel) that I do not know how to debug.</div><div><br></div><div>If I run this: (valgt is a pandas dataframe with 354 lines and 14 numerical columns)</div><div><br></div><div><div>sns.pairplot(valgt)</div><div>plt.savefig('parvis.png', dpi=600)</div></div><div><br></div><div>in the same cell, the kernel consistently crashes. If I run in two separate cells, or exporting to a Python script and running there (changing the inline render to qt), everything is OK.</div><div><br></div><div><br></div><div>This is win64 on Win7, both the free build from conda, the accelerate build from conda, and also a version I have built myself. (VS2010 and Intel Fortran XE 2015)</div><div><br></div><div>The following is the package list from the 'vanilla' conda:</div><div><br></div><div><div># packages in environment [...]:</div><div>#</div><div>dateutil 2.1 py34_2</div><div>ipython 2.3.1 py34_0</div><div>jinja2 2.7.3 py34_1</div><div>markupsafe 0.23 py34_0</div><div>matplotlib 1.4.2 np19py34_0</div><div>numexpr 2.3.1 np19py34_0</div><div>numpy 1.9.1 py34_0</div><div>pandas 0.15.2 np19py34_0</div><div>pip 6.0.6 py34_0</div><div>pyparsing 2.0.1 py34_0</div><div>pyqt 4.10.4 py34_0</div><div>pyreadline 2.0 py34_0</div><div>python 3.4.2 1</div><div>pytz 2014.9 py34_0</div><div>pywin32 219 py34_0</div><div>pyzmq 14.5.0 py34_0</div><div>rpy2 2.5.5 <pip></div><div>scipy 0.15.1 np19py34_0</div><div>seaborn 0.5.1 np19py34_0</div><div>setuptools 12.0.5 py34_0</div><div>six 1.9.0 py34_0</div><div>tornado 4.0.2 py34_0</div></div><div><br></div><div>rpy2 is from the Gohlke python collection. <br></div><div><br></div><div><br></div><div>The home-built version is a python 3.4.3rc1, with</div><div><br></div><div><div>certifi (14.5.14)</div><div>Cython (0.21.2)</div><div>ipython (2.4.1)</div><div>Jinja2 (2.7.3)</div><div>jsonschema (2.4.0)</div><div>MarkupSafe (0.23)</div><div>matplotlib (1.4.2)</div><div>mistune (0.5)</div><div>nose (1.3.4)</div><div>numexpr (2.4.1.dev0)</div><div>numpy (1.10.0.dev0+98a8fe3)</div><div>pandas (0.15.2)</div><div>pip (6.0.8)</div><div>pyparsing (2.0.3)</div><div>pyreadline (2.0)</div><div>PySide (1.2.2)</div><div>python-dateutil (2.4.0)</div><div>pytz (2014.10)</div><div>pywin32 (219)</div><div>pyzmq (14.5.0)</div><div>rpy2 (2.5.6)</div><div>scipy (0.15.1)</div><div>seaborn (0.6.dev0)</div><div>setuptools (12.0.5)</div><div>six (1.9.0)</div><div>tornado (4.1)</div><div>wheel (0.24.0)</div></div><div><br></div><div><br></div><div>They crash in the same way, though, the same applies to a test I did with ipython 3</div><div><br></div><div><br></div><div>I am wondering how important it is to have everything built with the same compiler and MKL library. Does all packages that use numpy need to be built against the specific numpy+mkl version? For example rpy2 (which is not used near the cells where it crashes), should I always rebuild this as well if I build differently, or...?</div><div><br></div><div>And, how can I generate debug information from the kernel that ipython notebook starts?</div><div><br></div><div>Any pointers are most appreciated.</div><div><br></div><div><br></div><div><br clear="all"><div><br></div>-- <br><div class="gmail_signature">mvh,<br>Åsmund Hjulstad<br></div>
</div></div>