[Numpy-tickets] [NumPy] #589: problem installing with latest version of Intel MKL
NumPy
numpy-tickets@scipy....
Wed Nov 28 23:07:03 CST 2007
#589: problem installing with latest version of Intel MKL
-----------------------------+----------------------------------------------
Reporter: DavidPowell | Owner: cookedm
Type: defect | Status: new
Priority: normal | Milestone: 1.0.5
Component: numpy.distutils | Version: none
Severity: normal | Resolution:
Keywords: |
-----------------------------+----------------------------------------------
Comment (by cdavid):
No need to create symlink: just use the site.cfg. This is not well
documented, but you can use lapack_libs option in the [mkl] section of the
site.cfg.
I added an example in the site.cfg.example (it works for mewith mkl
9.0.21):
{{{
#!python
numpy.show_config()
lapack_opt_info:
libraries = ['mkl_lapack', 'mkl', 'vml', 'guide', 'pthread']
library_dirs = ['/home/david/opt/intel/mkl/9.1.023/lib/32/']
define_macros = [('SCIPY_MKL_H', None)]
include_dirs = ['/usr/local/include', '/usr/include']
blas_opt_info:
libraries = ['mkl', 'vml', 'guide', 'pthread']
library_dirs = ['/home/david/opt/intel/mkl/9.1.023/lib/32/']
define_macros = [('SCIPY_MKL_H', None)]
include_dirs = ['/usr/local/include', '/usr/include']
lapack_mkl_info:
libraries = ['mkl_lapack', 'mkl', 'vml', 'guide', 'pthread']
library_dirs = ['/home/david/opt/intel/mkl/9.1.023/lib/32/']
define_macros = [('SCIPY_MKL_H', None)]
include_dirs = ['/usr/local/include', '/usr/include']
blas_mkl_info:
libraries = ['mkl', 'vml', 'guide', 'pthread']
library_dirs = ['/home/david/opt/intel/mkl/9.1.023/lib/32/']
define_macros = [('SCIPY_MKL_H', None)]
include_dirs = ['/usr/local/include', '/usr/include']
mkl_info:
libraries = ['mkl', 'vml', 'guide', 'pthread']
library_dirs = ['/home/david/opt/intel/mkl/9.1.023/lib/32/']
define_macros = [('SCIPY_MKL_H', None)]
include_dirs = ['/usr/local/include', '/usr/include']
}}}
I think the ticket can be closed.
--
Ticket URL: <http://scipy.org/scipy/numpy/ticket/589#comment:2>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.
More information about the Numpy-tickets
mailing list