[SciPy-User] Building NumPy 1.6.0 with MKL and MSVS9 on XP 32bit

Ralf Gommers ralf.gommers at googlemail.com
Tue May 17 03:07:46 EDT 2011


On Tue, May 17, 2011 at 9:04 AM, Christoph Gohlke <cgohlke at uci.edu> wrote:

>
>
> On 5/16/2011 10:11 PM, Klonuo Umom wrote:
> >> Try the Intel MKL Link Line Advisor
> >> <http://software.intel.com/en-us/articles/intel-mkl-link-line-advisor/>
> >>
> >> I use the following multi-thread static libraries with MKL 11.1:
> >>
> mkl_lapack95,mkl_blas95,mkl_intel_c,mkl_intel_thread,mkl_core,libiomp5md,libifportmd
> >>
> >
> > Thanks for the tip, Christoph
> > After that install was 'successful' but not the results.
> >
> > I installed this on P4 with 768 MB RAM (ifort 11.1, MSVC9). This PC is
> just one
> > of many nodes I plan to use and others are similarly single CPU P4 with
> 1GB RAM
> >
> > Running:
> > import numpy as np
> > A=np.ones((1000,1000))
> > B=np.ones((1000,1000))
> > %timeit np.dot(A, B)
> >
> > In IPython, gives this results:
> >
> > Cygwin/ATLAS   : 1 loops, best of 3: 843 ms per loop
> > Windows/ATLAS* : 1 loops, best of 3: 1.29 s per loop
> > Windows/MKL    : 1 loops, best of 3: 2.37 s per loop
> >
> > *from offical binary installer
> >
> >
> > I then run np.test() (with MKL build numpy) and there were indeed errors
> :(
> >
> > Now what? :)
> >
> >
> > Running unit tests for numpy
> > NumPy version 1.6.0
> > NumPy is installed in C:\Python26\lib\site-packages\numpy
> > Python version 2.6.6 (r266:84297, Aug 24 2010, 18:46:32) [MSC v.1500 32
> bit (Intel)]
> > nose version 1.0.0
>
> <snip>
>
> > ----------------------------------------------------------------------
> > Ran 3345 tests in 89.437s
> >
> > OK (KNOWNFAIL=8, SKIP=4)
> > Out[7]:<nose.result.TextTestResult run=3345 errors=0 failures=0>
> >
>
> Which errors are you referring to? np.test() passes without errors and
> failures.
>
> I'm guessing the warnings spit out by the test suite. Do you see that too?
The output from gcc builds is much cleaner.

Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20110517/412f9698/attachment.html>


More information about the SciPy-User mailing list