[Numpy-discussion] Installing numpy-mkl binary on top of Python(x, y)
Sun Jan 27 15:34:16 CST 2013
On Sun, Jan 27, 2013 at 2:54 PM, Christoph Gohlke <email@example.com> wrote:
> On 1/27/2013 11:40 AM, firstname.lastname@example.org wrote:
>> if I want to have a painless Python installation build against Intel MKL on Windows, one obvious choice is to just buy the EPD package. However,
>> as I already do have a C++ licence of the MKL library I was wondering if I could just install the Python(x,y) -distribution and then take one of the NumPy-MKL binaries provided
>> by Christoph Gohlke. Is it simple as that? Any downsides, will SciPy work as well? On the plus side, I would get Spyder2 without hassle and it looks nice to a former Matlab user.
>> I apologize for such a simple question, I would have tried it myself but this is for my work where only IT support has the admin rights and I have mac at home. I want it to be as
>> clearcut for them as possible so I get things up and running. I did try to search the internet and the list but did not find a conclusive answer.
>> Many thanks in advance for any help.
>> All the best,
> Try WinPython <http://code.google.com/p/winpython/>. It repackages
> numpy-MKL and other packages from
> <http://www.lfd.uci.edu/~gohlke/pythonlibs/>, contains Spyder and all
> dependencies, is available as 64 bit, and does not require admin rights
> to install.
You can replace python xy installed packages but it's necessary to
watch out for dependencies.
If you replace numpy with the mkl version, then you also have to
replace scipy with the mkl version, as far as I understand.
I initially installed python xy on a new computer and updated many
packages since, using standard python not the python xy updates.
The only problem I have is that I have some incompatibilities between
QT, pyQT, pyside, spyder and the ipython qt console, the later doesn't
work in my current setup.
> NumPy-Discussion mailing list
More information about the NumPy-Discussion