Unable to install sfepy 2017.1 in python 3.6 environment using anaconda3
Hi, I've been unable to install sfepy 2017.1 using conda-forge in the python 3.6 environment.
The error was:
"UnsatisfiableError: The following specifications were found to be in conflict:
- python 3.6*
- sfepy 2017.1* -> python 2.7* Use "conda info <package>" to see the dependencies for each package."
After using conda info sfepy this is what I've got:
"sfepy 2017.1 np111py35_0
file name : sfepy-2017.1-np111py35_0.tar.bz2 name : sfepy version : 2017.1 build string: np111py35_0 build number: 0 channel : conda-forge size : 7.3 MB arch : x86_64 has_prefix : True license : BSD-3-Clause md5 : b38a92b0ecdf0ead6b390cd984e00de1 noarch : None platform : linux requires : () subdir : linux-64 url : https://conda.anaconda.org/conda-forge/linux-64/sfepy-2017.1-np111py35_0.tar... dependencies: matplotlib mpi4py numpy 1.11* pyparsing pytables python 3.5* scipy sympy"
What am I missing in order to install sfepy 2007.1 in python 3.6? Thanks, woiski
Hi,
we have been skipping python 3.6 in the conda-forge recipe, as there were probably some compatibility problems (I do not recall exactly). SfePy now works with 3.6, so I am trying to update the recipe.
r.
On 03/26/2017 11:50 PM, Emanuel Woiski wrote:
Hi, I've been unable to install sfepy 2017.1 using conda-forge in the python 3.6 environment.
The error was:
"UnsatisfiableError: The following specifications were found to be in conflict:
- python 3.6*
- sfepy 2017.1* -> python 2.7* Use "conda info <package>" to see the dependencies for each package."
After using conda info sfepy this is what I've got:
"sfepy 2017.1 np111py35_0
file name : sfepy-2017.1-np111py35_0.tar.bz2 name : sfepy version : 2017.1 build string: np111py35_0 build number: 0 channel : conda-forge size : 7.3 MB arch : x86_64 has_prefix : True license : BSD-3-Clause md5 : b38a92b0ecdf0ead6b390cd984e00de1 noarch : None platform : linux requires : () subdir : linux-64 url : https://conda.anaconda.org/conda-forge/linux-64/sfepy-2017.1-np111py35_0.tar... dependencies: matplotlib mpi4py numpy 1.11* pyparsing pytables python 3.5* scipy sympy"
What am I missing in order to install sfepy 2007.1 in python 3.6? Thanks, woiski
Awesome, thanks. Eventually I will try it again. woiski
2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi,
we have been skipping python 3.6 in the conda-forge recipe, as there were probably some compatibility problems (I do not recall exactly). SfePy now works with 3.6, so I am trying to update the recipe.
r.
On 03/26/2017 11:50 PM, Emanuel Woiski wrote:
Hi, I've been unable to install sfepy 2017.1 using conda-forge in the python 3.6 environment.
The error was:
"UnsatisfiableError: The following specifications were found to be in conflict:
- python 3.6*
- sfepy 2017.1* -> python 2.7* Use "conda info <package>" to see the dependencies for each package."
After using conda info sfepy this is what I've got:
"sfepy 2017.1 np111py35_0
file name : sfepy-2017.1-np111py35_0.tar.bz2 name : sfepy version : 2017.1 build string: np111py35_0 build number: 0 channel : conda-forge size : 7.3 MB arch : x86_64 has_prefix : True license : BSD-3-Clause md5 : b38a92b0ecdf0ead6b390cd984e00de1 noarch : None platform : linux requires : () subdir : linux-64 url : https://conda.anaconda.org/conda-forge/linux-64/sfepy-2017. 1-np111py35_0.tar.bz2 dependencies: matplotlib mpi4py numpy 1.11* pyparsing pytables python 3.5* scipy sympy"
What am I missing in order to install sfepy 2007.1 in python 3.6? Thanks, woiski
-- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
Hi Emanuel,
the python 3.6 installation should work now out of the box, except the two tests that depend on igakit.
r.
On 03/27/2017 05:47 PM, Emanuel Woiski wrote:
Awesome, thanks. Eventually I will try it again. woiski
2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi,
we have been skipping python 3.6 in the conda-forge recipe, as there were probably some compatibility problems (I do not recall exactly). SfePy now works with 3.6, so I am trying to update the recipe.
r.
On 03/26/2017 11:50 PM, Emanuel Woiski wrote:
Hi, I've been unable to install sfepy 2017.1 using conda-forge in the python 3.6 environment.
The error was:
"UnsatisfiableError: The following specifications were found to be in conflict:
- python 3.6*
- sfepy 2017.1* -> python 2.7* Use "conda info <package>" to see the dependencies for each package."
After using conda info sfepy this is what I've got:
"sfepy 2017.1 np111py35_0
file name : sfepy-2017.1-np111py35_0.tar.bz2 name : sfepy version : 2017.1 build string: np111py35_0 build number: 0 channel : conda-forge size : 7.3 MB arch : x86_64 has_prefix : True license : BSD-3-Clause md5 : b38a92b0ecdf0ead6b390cd984e00de1 noarch : None platform : linux requires : () subdir : linux-64 url : https://conda.anaconda.org/conda-forge/linux-64/sfepy-2017. 1-np111py35_0.tar.bz2 dependencies: matplotlib mpi4py numpy 1.11* pyparsing pytables python 3.5* scipy sympy"
What am I missing in order to install sfepy 2007.1 in python 3.6? Thanks, woiski
-- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
Thanks! There was no problem with the installation using 'conda install sfepy'. However 'python run_tests.py' does not work. So I was trying to test using 'nosetests sfepy' and the following message was produced:
EEEE....EEEEEEEEEEEEEEEEEEEEEEEE..Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so.
According to conda, mkl is already installed.
Any further help or suggestion? woiski
2017-03-28 4:17 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi Emanuel,
the python 3.6 installation should work now out of the box, except the two tests that depend on igakit.
r.
On 03/27/2017 05:47 PM, Emanuel Woiski wrote:
Awesome, thanks. Eventually I will try it again. woiski
2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi,
we have been skipping python 3.6 in the conda-forge recipe, as there were probably some compatibility problems (I do not recall exactly). SfePy now works with 3.6, so I am trying to update the recipe.
r.
[snip]
On 29.3.2017 21:44, Emanuel Woiski wrote:
Thanks! There was no problem with the installation using 'conda install sfepy'. However 'python run_tests.py' does not work. So I was trying to test using
'python run_tests.py' is meant to be used from the directory of sfepy sources (e.g. if you clone the code from github). To test an installed sfepy, try using 'sfepy-run run_tests'. The instructions on the web seem to be wrong.
r.
'nosetests sfepy' and the following message was produced:
EEEE....EEEEEEEEEEEEEEEEEEEEEEEE..Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so.
According to conda, mkl is already installed.
Any further help or suggestion? woiski
2017-03-28 4:17 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi Emanuel,
the python 3.6 installation should work now out of the box, except the two tests that depend on igakit.
r.
On 03/27/2017 05:47 PM, Emanuel Woiski wrote:
Awesome, thanks. Eventually I will try it again. woiski
2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi,
we have been skipping python 3.6 in the conda-forge recipe, as there were probably some compatibility problems (I do not recall exactly). SfePy now works with 3.6, so I am trying to update the recipe.
r.
[snip]
Concerning the MKL errors - do numpy and scipy work?
r.
On 29.3.2017 23:26, Robert Cimrman wrote:
On 29.3.2017 21:44, Emanuel Woiski wrote:
Thanks! There was no problem with the installation using 'conda install sfepy'. However 'python run_tests.py' does not work. So I was trying to test using
'python run_tests.py' is meant to be used from the directory of sfepy sources (e.g. if you clone the code from github). To test an installed sfepy, try using 'sfepy-run run_tests'. The instructions on the web seem to be wrong.
r.
'nosetests sfepy' and the following message was produced:
EEEE....EEEEEEEEEEEEEEEEEEEEEEEE..Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so.
According to conda, mkl is already installed.
Any further help or suggestion? woiski
2017-03-28 4:17 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi Emanuel,
the python 3.6 installation should work now out of the box, except the two tests that depend on igakit.
r.
On 03/27/2017 05:47 PM, Emanuel Woiski wrote:
Awesome, thanks. Eventually I will try it again. woiski
2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi,
we have been skipping python 3.6 in the conda-forge recipe, as there were probably some compatibility problems (I do not recall exactly). SfePy now works with 3.6, so I am trying to update the recipe.
r.
[snip]
yes, they work. It seems that the installed sfepy is unable to find the mkl libraries. woiski
2017-03-30 3:20 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Concerning the MKL errors - do numpy and scipy work?
r.
On 29.3.2017 23:26, Robert Cimrman wrote:
On 29.3.2017 21:44, Emanuel Woiski wrote:
Thanks! There was no problem with the installation using 'conda install sfepy'. However 'python run_tests.py' does not work. So I was trying to test using
'python run_tests.py' is meant to be used from the directory of sfepy sources (e.g. if you clone the code from github). To test an installed sfepy, try using 'sfepy-run run_tests'. The instructions on the web seem to be wrong.
r.
'nosetests sfepy' and the following message was produced:
EEEE....EEEEEEEEEEEEEEEEEEEEEEEE..Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so.
According to conda, mkl is already installed.
Any further help or suggestion? woiski
2017-03-28 4:17 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi Emanuel,
the python 3.6 installation should work now out of the box, except the two tests that depend on igakit.
r.
On 03/27/2017 05:47 PM, Emanuel Woiski wrote:
Awesome, thanks.
Eventually I will try it again. woiski
2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi,
we have been skipping python 3.6 in the conda-forge recipe, as there were probably some compatibility problems (I do not recall exactly). SfePy now works with 3.6, so I am trying to update the recipe.
r.
[snip]
-- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
Well, sfepy does not use MKL by itself anywhere, only through other packages, like numpy.
"sfepy-run run_tests" also fails? If yes, could you send us the output of "sfepy-run run_tests --raise"?
Thanks, r.
On 03/30/2017 12:27 PM, Emanuel Woiski wrote:
yes, they work. It seems that the installed sfepy is unable to find the mkl libraries. woiski
2017-03-30 3:20 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Concerning the MKL errors - do numpy and scipy work?
r.
On 29.3.2017 23:26, Robert Cimrman wrote:
On 29.3.2017 21:44, Emanuel Woiski wrote:
Thanks! There was no problem with the installation using 'conda install sfepy'. However 'python run_tests.py' does not work. So I was trying to test using
'python run_tests.py' is meant to be used from the directory of sfepy sources (e.g. if you clone the code from github). To test an installed sfepy, try using 'sfepy-run run_tests'. The instructions on the web seem to be wrong.
r.
'nosetests sfepy' and the following message was produced:
EEEE....EEEEEEEEEEEEEEEEEEEEEEEE..Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so.
According to conda, mkl is already installed.
Any further help or suggestion? woiski
2017-03-28 4:17 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi Emanuel,
the python 3.6 installation should work now out of the box, except the two tests that depend on igakit.
r.
On 03/27/2017 05:47 PM, Emanuel Woiski wrote:
Awesome, thanks.
Eventually I will try it again. woiski
2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi,
> > we have been skipping python 3.6 in the conda-forge recipe, as there > were > probably some compatibility problems (I do not recall exactly). SfePy > now > works with 3.6, so I am trying to update the recipe. > > r. > > > > [snip] >
-- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
Sure. Please find attached. Thanks for looking into it. woiski
2017-03-30 8:33 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Well, sfepy does not use MKL by itself anywhere, only through other packages, like numpy.
"sfepy-run run_tests" also fails? If yes, could you send us the output of "sfepy-run run_tests --raise"?
Thanks, r.
On 03/30/2017 12:27 PM, Emanuel Woiski wrote:
yes, they work. It seems that the installed sfepy is unable to find the mkl libraries. woiski
2017-03-30 3:20 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Concerning the MKL errors - do numpy and scipy work?
r.
On 29.3.2017 23:26, Robert Cimrman wrote:
On 29.3.2017 21:44, Emanuel Woiski wrote:
Thanks!
There was no problem with the installation using 'conda install sfepy'. However 'python run_tests.py' does not work. So I was trying to test using
'python run_tests.py' is meant to be used from the directory of sfepy sources (e.g. if you clone the code from github). To test an installed sfepy, try using 'sfepy-run run_tests'. The instructions on the web seem to be wrong.
r.
'nosetests sfepy' and the following message was produced:
EEEE....EEEEEEEEEEEEEEEEEEEEEEEE..Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so. Intel MKL FATAL ERROR: Cannot load libmkl_avx2.so or libmkl_def.so.
According to conda, mkl is already installed.
Any further help or suggestion? woiski
2017-03-28 4:17 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Hi Emanuel,
the python 3.6 installation should work now out of the box, except the two tests that depend on igakit.
r.
On 03/27/2017 05:47 PM, Emanuel Woiski wrote:
Awesome, thanks.
> Eventually I will try it again. > woiski > > 2017-03-27 9:38 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz: > > Hi, > > >> we have been skipping python 3.6 in the conda-forge recipe, as there >> were >> probably some compatibility problems (I do not recall exactly). >> SfePy >> now >> works with 3.6, so I am trying to update the recipe. >> >> r. >> >> >> >> [snip] >> >> >
--
You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
-- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
On 30.3.2017 21:47, Emanuel Woiski wrote:
Sure. Please find attached. Thanks for looking into it. woiski
So it fails when the code is about to solve a linear system. Do you have scikit-umfpack installed? Does https://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.sparse.lin... work?
r.
2017-03-30 8:33 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Well, sfepy does not use MKL by itself anywhere, only through other packages, like numpy.
"sfepy-run run_tests" also fails? If yes, could you send us the output of "sfepy-run run_tests --raise"?
Good guess! After installing scikit-umfpack using conda(-forge?), I executed 'sfepy-run run_tests'. The final message was: 98 test file(s) executed in 695.81 s, 2 failure(s) of 165 test(s).
The two failures (expected) are related to navier-stokes2d _iga.py and test_projection_iga_fem . I think it is ok now! I must thank you again for your support. Best regards woiski
2017-03-30 16:56 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
On 30.3.2017 21:47, Emanuel Woiski wrote:
Sure. Please find attached. Thanks for looking into it. woiski
So it fails when the code is about to solve a linear system. Do you have scikit-umfpack installed? Does https://docs.scipy.org/doc/sci py-0.14.0/reference/generated/scipy.sparse.linalg.spsolve.html work?
r.
2017-03-30 8:33 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Well, sfepy does not use MKL by itself anywhere, only through other
packages, like numpy.
"sfepy-run run_tests" also fails? If yes, could you send us the output of "sfepy-run run_tests --raise"?
-- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
On 30.3.2017 22:38, Emanuel Woiski wrote:
Good guess! After installing scikit-umfpack using conda(-forge?), I executed 'sfepy-run run_tests'. The final message was: 98 test file(s) executed in 695.81 s, 2 failure(s) of 165 test(s).
The two failures (expected) are related to navier-stokes2d _iga.py and test_projection_iga_fem . I think it is ok now! I must thank you again for your support. Best regards woiski
OK, hth!
r.
2017-03-30 16:56 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
On 30.3.2017 21:47, Emanuel Woiski wrote:
Sure. Please find attached. Thanks for looking into it. woiski
So it fails when the code is about to solve a linear system. Do you have scikit-umfpack installed? Does https://docs.scipy.org/doc/sci py-0.14.0/reference/generated/scipy.sparse.linalg.spsolve.html work?
r.
2017-03-30 8:33 GMT-03:00 Robert Cimrman cimr...@ntc.zcu.cz:
Well, sfepy does not use MKL by itself anywhere, only through other
packages, like numpy.
"sfepy-run run_tests" also fails? If yes, could you send us the output of "sfepy-run run_tests --raise"?
-- You received this message because you are subscribed to the Google Groups "sfepy-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sfepy-devel...@googlegroups.com. To post to this group, send email to sfepy...@googlegroups.com. Visit this group at https://groups.google.com/group/sfepy-devel.
participants (2)
-
Emanuel Woiski
-
Robert Cimrman