[Numpy-discussion] numpy with icc

rex rex@nosyntax....
Sun May 18 05:13:38 CDT 2008


> I am trying to build numpy with intel icc and mkl.  I don't understand
> a lot of what I am doing.  

Me, too. I have built it with icc & MKL several times in the past,
but cannot build the numpy svn with MKL now. I can build it with
icc and no MKL, and it passes all the tests with no errors.

I've interleaved my setup with yours to make it easy to compare them. 


> ----------------------------------------------------------------------------------
> site.cfg:

> [DEFAULT]
> library_dirs = /opt/intel/mkl/10.0.3.020/lib/em64t,/usr/lib64,/usr/local/lib64,/usr/local/python2.5.2-intel/lib,/usr/lib,/usr/local/lib
> include_dirs = /opt/intel/mkl/10.0.3.020/include,/usr/include,/usr/local/include,/usr/local/python2.5.2-intel/include

> [mkl]
> include_dirs = /opt/intel/mkl/10.0.3.020/include
> library_dirs = /opt/intel/mkl/10.0.3.020/lib/em64t
> lapack_libs = mkl_lapack

> [lapack_src]
> libraries=mkl_lapack,mkl,guide

> [lapack_info]
> libraries=mkl_lapack,mkl,guide
> -------------------------------------------------------------------------------

[DEFAULT]
library_dirs = /opt/intel/mkl/10.0.3.020/lib/32
include_dirs = /opt/intel/mkl/10.0.3.020/10.0.3.020/include

[mkl]
library_dirs = /opt/intel/mkl/10.0.3.020/lib/32
lapack_libs = mkl, mkl_lapack, guide

#my reading of the Intel MKL docs suggests that vml does not need
to be explicitly added, but that mkl_lapack does.


> intelccompiler.py:

> from distutils.unixccompiler import UnixCCompiler
> from numpy.distutils.exec_command import find_executable

> class IntelCCompiler(UnixCCompiler):

>    """ A modified Intel compiler compatible with an gcc built Python.
>    """

>    compiler_type = 'intel'
>    cc_exe = 'icc -g -O3 -w -fPIC -parallel -ipo -xT -axT'

cc_exe = 'icc -msse3 -fast'    #adjust to suit your cpu

I think some of those flags can be omitted: drop -g, and -fast
implies several other flags, e.g., -xT, IIRC. (I'm building on a
Core 2 Duo)


>    def __init__ (self, verbose=0, dry_run=0, force=0):
>        UnixCCompiler.__init__ (self, verbose,dry_run, force)
>        compiler = self.cc_exe
>        self.set_executables(compiler=compiler,
>                             compiler_so=compiler,
>                             compiler_cxx=compiler,
>                             linker_exe=compiler,
>                             linker_so=compiler + ' -shared')

> class IntelItaniumCCompiler(IntelCCompiler):
>    compiler_type = 'intele'

>    # On Itanium, the Intel Compiler used to be called ecc, let's search for
>    # it (now it's also icc, so ecc is last in the search).
>    for cc_exe in map(find_executable,['icc','ecc']):
>        if cc_exe:
>            break
> ----------------------------------------------------------------

> system_info.py:

> ....
> class lapack_mkl_info(mkl_info):

>    def calc_info(self):
>        mkl = get_info('mkl')
>        if not mkl:
>            return
>        if sys.platform == 'win32':
>            lapack_libs = self.get_libs('lapack_libs',['mkl_lapack'])
>        else:
>            lapack_libs = self.get_libs('lapack_libs',['mkl_lapack'])

>        info = {'libraries': lapack_libs}
>        dict_append(info,**mkl)
>        self.set_info(**info)
> ...
> ------------------------------------------------------------------------

I made the same change in the above. It was originally:

if sys.platform == 'win32':
            lapack_libs = self.get_libs('lapack_libs',['mkl_lapack'])
        else:
            lapack_libs = self.get_libs('lapack_libs',['mkl_lapack32','mkl_lapack64]')

Which does not work because all 3 libs are now named mkl_lapack.

I made another change in system.info.py:

class mkl_info(system_info):
    section = 'mkl'
    dir_env_var = 'MKL'
    _lib_mkl = ['mkl','mkl_lapack','guide']

It originally had 'vml' instead of 'mkl_lapack'


> i then use this command to compile:

> /usr/local/python2.5.2-intel/bin/python setup.py config
> --compiler=intel config_fc --fcompiler=intel \
> --opt='-fPIC -O3 -w -axT -xT' install > build.out

I used:

numpy# python setup.py config --compiler=intel build_clib \
 --compiler=intel build_ext --compiler=intel install \
 --prefix=/usr/local> build51        

> build.out has this in it:

> F2PY Version 2_4422
> blas_opt_info:
> blas_mkl_info:
>  libraries mkl,vml,guide not found in /opt/intel/mkl/10.0.3.020/lib/em64t
>  NOT AVAILABLE

> atlas_blas_threads_info:
> Setting PTATLAS=ATLAS
>  NOT AVAILABLE

> atlas_blas_info:
>  NOT AVAILABLE

> blas_info:
>  NOT AVAILABLE

> blas_src_info:
>  NOT AVAILABLE

>  NOT AVAILABLE

> lapack_opt_info:
> lapack_mkl_info:
> mkl_info:
>  libraries mkl,vml,guide not found in /opt/intel/mkl/10.0.3.020/lib/em64t
>  NOT AVAILABLE

>  NOT AVAILABLE

> atlas_threads_info:
> Setting PTATLAS=ATLAS
> numpy.distutils.system_info.atlas_threads_info
>  NOT AVAILABLE

> atlas_info:
> numpy.distutils.system_info.atlas_info
>  NOT AVAILABLE

> lapack_info:
>  NOT AVAILABLE

> lapack_src_info:
>  NOT AVAILABLE

>  NOT AVAILABLE


build53 (I used svn 5183) has:

F2PY Version 2_5183
blas_opt_info:
blas_mkl_info:
  FOUND:
    libraries = ['mkl', 'mkl_lapack', 'guide', 'pthread']
    library_dirs = ['/opt/intel/mkl/10.0.3.020/lib/32']
    define_macros = [('SCIPY_MKL_H', None)]
    include_dirs = []

  FOUND:
    libraries = ['mkl', 'mkl_lapack', 'guide', 'pthread']
    library_dirs = ['/opt/intel/mkl/10.0.3.020/lib/32']
    define_macros = [('SCIPY_MKL_H', None)]
    include_dirs = []

lapack_opt_info:
lapack_mkl_info:
mkl_info:
  FOUND:
    libraries = ['mkl', 'mkl_lapack', 'guide', 'pthread']
    library_dirs = ['/opt/intel/mkl/10.0.3.020/lib/32']
    define_macros = [('SCIPY_MKL_H', None)]
    include_dirs = []

  FOUND:
    libraries = ['mkl', 'mkl_lapack', 'guide', 'mkl', 'mkl_lapack', 'guide', 'pthread']
    library_dirs = ['/opt/intel/mkl/10.0.3.020/lib/32']
    define_macros = [('SCIPY_MKL_H', None)]
    include_dirs = []

  FOUND:
    libraries = ['mkl', 'mkl_lapack', 'guide', 'mkl', 'mkl_lapack', 'guide', 'pthread']
    library_dirs = ['/opt/intel/mkl/10.0.3.020/lib/32']

running config
running build_clib
[...]


> running config
> running config_fc
> unifing config_fc, config, build_clib, build_ext, build commands
> --fcompiler options
> running install
> running build
> running config_cc
> unifing config_cc, config, build_clib, build_ext, build commands
> --compiler options
> running build_src
> building py_modules sources
> creating build
> creating build/src.linux-x86_64-2.5
> creating build/src.linux-x86_64-2.5/numpy
> creating build/src.linux-x86_64-2.5/numpy/distutils
> building extension "numpy.core.multiarray" sources
> creating build/src.linux-x86_64-2.5/numpy/core
> Generating build/src.linux-x86_64-2.5/numpy/core/config.h
> Found executable /opt/intel/cce/10.1.015/bin/icc
> Could not locate executable ecc
> customize IntelFCompiler
> Found executable /opt/intel/fce/10.1.015/bin/ifort
> C compiler: icc -g -O3 -w -fPIC -parallel -ipo -xT -axT

> ......

> it seems to compile and install fine....  then i start python and try
> to run numpy.test().  that is where i am stuck.  this is what happens:

> # pwd
> /usr/local/python2.5.2-intel/bin
> # ./python
> Python 2.5.2 (r252:60911, May 13 2008, 11:22:16)
> [GCC Intel(R) C++ gcc 4.1 mode] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> >>> import numpy
> >>> numpy.test()
> Numpy is installed in
> /usr/local/python2.5.2-intel/lib/python2.5/site-packages/numpy
> Numpy version 1.0.4
> Python version 2.5.2 (r252:60911, May 13 2008, 11:22:16) [GCC Intel(R)
> C++ gcc 4.1 mode]
>  Found 10/10 tests for numpy.core.defmatrix
>  Found 36/36 tests for numpy.core.ma
>  Found 223/223 tests for numpy.core.multiarray
>  Found 65/65 tests for numpy.core.numeric
>  Found 31/31 tests for numpy.core.numerictypes
>  Found 12/12 tests for numpy.core.records
>  Found 6/6 tests for numpy.core.scalarmath
>  Found 14/14 tests for numpy.core.umath
>  Found 4/4 tests for numpy.ctypeslib
>  Found 5/5 tests for numpy.distutils.misc_util
>  Found 1/1 tests for numpy.fft.fftpack
>  Found 3/3 tests for numpy.fft.helper
>  Found 9/9 tests for numpy.lib.arraysetops
>  Found 46/46 tests for numpy.lib.function_base
>  Found 5/5 tests for numpy.lib.getlimits
>  Found 4/4 tests for numpy.lib.index_tricks
>  Found 3/3 tests for numpy.lib.polynomial
>  Found 49/49 tests for numpy.lib.shape_base
>  Found 15/15 tests for numpy.lib.twodim_base
>  Found 43/43 tests for numpy.lib.type_check
>  Found 1/1 tests for numpy.lib.ufunclike
>  Found 40/40 tests for numpy.linalg
>  Found 2/2 tests for numpy.random
>  Found 0/0 tests for __main__
> MKL FATAL ERROR: /opt/intel/mkl/10.0.3.020/lib/em64t/: cannot read
> file data: Is a directory
> #

I get the same error:

python
Python 2.5 (release25-maint, Dec  9 2006, 14:35:53)
[GCC 4.1.2 20061115 (prerelease) (Debian 4.1.1-20)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>> numpy.test()
Numpy is installed in /usr/local/lib/python2.5/site-packages/numpy
Numpy version 1.1.0.dev5183
Python version 2.5 (release25-maint, Dec  9 2006, 14:35:53) [GCC 4.1.2 20061115 (prerelease) (Debian 4.1.1-20)]
  Found 18/18 tests for numpy.core.defmatrix
  Found 3/3 tests for numpy.core.memmap
  Found 283/283 tests for numpy.core.multiarray
  Found 70/70 tests for numpy.core.numeric
  Found 36/36 tests for numpy.core.numerictypes
  Found 12/12 tests for numpy.core.records
  Found 7/7 tests for numpy.core.scalarmath
  Found 16/16 tests for numpy.core.umath
  Found 5/5 tests for numpy.ctypeslib
  Found 5/5 tests for numpy.distutils.misc_util
  Found 2/2 tests for numpy.fft.fftpack
  Found 3/3 tests for numpy.fft.helper
  Found 24/24 tests for numpy.lib._datasource
  Found 10/10 tests for numpy.lib.arraysetops
  Found 1/1 tests for numpy.lib.financial
  Found 0/0 tests for numpy.lib.format
  Found 53/53 tests for numpy.lib.function_base
  Found 5/5 tests for numpy.lib.getlimits
  Found 6/6 tests for numpy.lib.index_tricks
  Found 15/15 tests for numpy.lib.io
  Found 1/1 tests for numpy.lib.machar
  Found 4/4 tests for numpy.lib.polynomial
  Found 49/49 tests for numpy.lib.shape_base
  Found 15/15 tests for numpy.lib.twodim_base
  Found 43/43 tests for numpy.lib.type_check
  Found 1/1 tests for numpy.lib.ufunclike
  Found 89/89 tests for numpy.linalg
  Found 94/94 tests for numpy.ma.core
  Found 15/15 tests for numpy.ma.extras
  Found 7/7 tests for numpy.random
  Found 16/16 tests for numpy.testing.utils
  Found 0/0 tests for __main__
MKL FATAL ERROR: /opt/intel/mkl/10.0.3.020/lib/32/: cannot read file data: Is a directory

> not sure how to get past this...  although, I am sure, if numpy is not
> working right then i will not be able to go on and compile scipy.....

I don't understand how MKL got linked into your system. Whenever I
get no messages of the type:

  FOUND:
    libraries = ['mkl', 'mkl_lapack', 'guide', 'pthread']
    library_dirs = ['/opt/intel/mkl/10.0.3.020/lib/32']
    define_macros = [('SCIPY_MKL_H', None)]
    include_dirs = []

in the build log, MKL doesn't get linked. If it IS linked it will
show up here:

# ldd /usr/local/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so
        linux-gate.so.1 =>  (0xffffe000)
        /opt/intel/mkl/10.0.3.020/lib/32/libmkl_intel.so (0xb7e62000)
        /opt/intel/mkl/10.0.3.020/lib/32/libmkl_intel_thread.so (0xb7c65000)
        /opt/intel/mkl/10.0.3.020/lib/32/libmkl_core.so (0xb7c01000)
        libmkl_lapack.so => /opt/intel/mkl/10.0.3.020/lib/32/libmkl_lapack.so (0xb7736000)
        libguide.so => /opt/intel/cc/10.1.015/lib/libguide.so (0xb76d4000)
        libpthread.so.0 => /lib/tls/i686/cmov/libpthread.so.0 (0xb76a8000)
        libimf.so => /opt/intel/cc/10.1.015/lib/libimf.so (0xb7478000)
        libsvml.so => /opt/intel/cc/10.1.015/lib/libsvml.so (0xb73a7000)
        libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0xb7382000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xb7377000)
        libintlc.so.5 => /opt/intel/cc/10.1.015/lib/libintlc.so.5 (0xb7334000)
        libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0xb7202000)
        libdl.so.2 => /lib/tls/i686/cmov/libdl.so.2 (0xb71fe000)
        /lib/ld-linux.so.2 (0x80000000)



More information about the Numpy-discussion mailing list