[SciPy-user] Problem with scipy on FC5 (PentiumM/PentiumIV)

Gerard Vermeulen gerard.vermeulen at grenoble.cnrs.fr
Mon Oct 16 06:55:31 CDT 2006


I have the same unit test failure on Mac OSX where I failed to build ATLAS from
source (it looks like that SciPy falls back on a different BLAS expecting a
different memory layout).

You can try to use something like the following site.cfg when rebuilding NumPy
(put it in the NumPy source tree before building):

cat >site.cfg <<EOF
[atlas]
library_dirs = /usr/lib
atlas_libs = lapack, f77blas, cblas, atlas
EOF

I needed to do this for my home made ATLAS rpm on Mandrake-2005: you have
to adapt it for your system.

Once I had built NumPy with the correct ATLAS settings, SciPy picked it up
and didn't complain about missing symbols anymore.

Gerard

PS: I consider this to be a NumPy bug.  Pre-NumPy versions of SciPy could
pick up my lapack, f77blas, cblas, atlas libraries without my intervention


On Mon, 16 Oct 2006 11:57:32 +0200
Grzegorz Koczyk <gkoczyk at echostar.pl> wrote:

> I am not sure whether this an FC5-specific problem, so I am submitting
> this here and not as a ticket. 
> 
> For Fedora Core 5 running on Pentium4-based servers following scipy
> tests fail and I am at a complete loss, at what can be done (and more
> importantly to what extent other modules depends on the tested code):  
> 
> [ Note: I did try compiling ATLAS (3.6) from source, however then both
> numpy and scipy complain about gfortran missing symbols in
> lapack_lite.so - so I am unable to verify whether problem stems from
> some peculiarity of vanilla FC5 RPMs. ]
> 
> ======================================================================
> FAIL: check_dot (scipy.linalg.tests.test_blas.test_fblas1_simple)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File
> "/usr/lib/python2.4/site-packages/scipy/linalg/tests/test_blas.py", line
> 75, in check_dot
>     assert_almost_equal(f([3j,-4,3-4j],[2,3,1]),-9+2j)
>   File "/usr/lib/python2.4/site-packages/numpy/testing/utils.py", line
> 156, in assert_almost_equal
>     assert round(abs(desired - actual),decimal) == 0, msg
> AssertionError:
> Items are not equal:
>  ACTUAL: (3.3613915289673437e-35-1.0143136978149414j)
>  DESIRED: (-9+2j)
> ----------------------------------------------------------------------
> Scipy configuration:
> 
> >>> numpy.__version__
> '1.0.dev3341'
> >>> scipy.__version__
> '0.5.2.dev2288'
> >>> scipy.show_config()
> umfpack_info:
>   NOT AVAILABLE
> 
> atlas_threads_info:
>   NOT AVAILABLE
> 
> blas_opt_info:
>     libraries = ['f77blas', 'cblas', 'atlas']
>     library_dirs = ['/usr/lib/sse2']
>     define_macros = [('ATLAS_INFO', '"\\"3.6.0\\""')]
>     language = c
>     include_dirs = ['/usr/include/atlas']
> 
> atlas_blas_threads_info:
>   NOT AVAILABLE
> 
> djbfft_info:
>   NOT AVAILABLE
> 
> lapack_opt_info:
>     libraries = ['lapack', 'f77blas', 'cblas', 'atlas']
>     library_dirs = ['/usr/lib/sse2']
>     define_macros = [('ATLAS_INFO', '"\\"3.6.0\\""')]
>     language = c
>     include_dirs = ['/usr/include/atlas']
> 
> fftw3_info:
>     libraries = ['fftw3']
>     library_dirs = ['/usr/lib']
>     define_macros = [('SCIPY_FFTW3_H', None)]
>     include_dirs = ['/usr/include']
> 
> atlas_info:
>     libraries = ['lapack', 'f77blas', 'cblas', 'atlas']
>     library_dirs = ['/usr/lib/sse2']
>     language = c
>     include_dirs = ['/usr/include/atlas']
> 
> lapack_mkl_info:
>   NOT AVAILABLE
> 
> blas_mkl_info:
>   NOT AVAILABLE
> 
> atlas_blas_info:
>     libraries = ['f77blas', 'cblas', 'atlas']
>     library_dirs = ['/usr/lib/sse2']
>     language = c
>     include_dirs = ['/usr/include/atlas']
> 
> mkl_info:
>   NOT AVAILABLE
> ----------------------------------------------------------------------
> Relevant RPMs installed:
> 
> atlas-sse2-3.6.0-10.fc5
> atlas-sse2-devel-3.6.0-10.fc5
> blas-3.0-37.fc5
> blas-devel-3.0-37.fc5
> fftw-3.1.1-1.fc5
> fftw-devel-3.1.1-1.fc5
> lapack-3.0-37.fc5
> lapack-devel-3.0-37.fc5
> python-2.4.3-8.FC5
> ----------------------------------------------------------------------
> /proc/cpuinfo:
> 
> processor       : 0
> vendor_id       : GenuineIntel
> cpu family      : 15
> model           : 4
> model name      : Intel(R) Pentium(R) 4 CPU 3.00GHz
> stepping        : 3
> cpu MHz         : 2992.931
> cache size      : 2048 KB
> physical id     : 0
> siblings        : 2
> core id         : 0
> cpu cores       : 1
> fdiv_bug        : no
> hlt_bug         : no
> f00f_bug        : no
> coma_bug        : no
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 5
> wp              : yes
> flags           : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca
> cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe lm
> constant_tsc pni monitor ds_cpl est cid cx16 xtpr
> bogomips        : 5997.83
> 
> processor       : 1
> vendor_id       : GenuineIntel
> cpu family      : 15
> model           : 4
> model name      : Intel(R) Pentium(R) 4 CPU 3.00GHz
> stepping        : 3
> cpu MHz         : 2992.931
> cache size      : 2048 KB
> physical id     : 0
> siblings        : 2
> core id         : 0
> cpu cores       : 1
> fdiv_bug        : no
> hlt_bug         : no
> f00f_bug        : no
> coma_bug        : no
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 5
> wp              : yes
> flags           : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca
> cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe lm
> constant_tsc pni monitor ds_cpl est cid cx16 xtpr
> bogomips        : 5985.40
> ----------------------------------------------------------------------
> 
> Any help or information concerning the matter would be highly
> appreciated.
> 
> Best regards,
> Grzegorz Koczyk
> 
> 
> 
> 
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
> 


More information about the SciPy-user mailing list