which installation of pysparse would you suggest ?
I have currently 1.1.1
I tried to install pysparse-1.2-dev203 or pysparse-1.2-dev213, but I have
errors at the installation;
first in the spmatrixmodule.c : there is a weird beginning of function
DL_EXPORT(void)
initspmatrix(void)
{
........
return ;
}
I replaced it by
DL_EXPORT(void){}
initspmatrix(void)
{
........
return 0;
}
bu then I get another error :
pysparse/sparse/src/spmatrixmodule.c:364: error: ‘PyTypeObject’ has no
member named ‘ob_type’
pysparse/sparse/src/spmatrixmodule.c:365: error: ‘PyTypeObject’ has no
member named ‘ob_type’
pysparse/sparse/src/spmatrixmodule.c:366: error: ‘PyTypeObject’ has no
member named ‘ob_type’
On Thu, Feb 24, 2011 at 11:48 AM, Julien Derr <[email protected]> wrote:
> indeed!
>
> I have a segmentation fault when I run the test!
> I am going to reinstall pyspare or maybe another version ??
>
>
>
> julien@derr-Precision-T1500:~/Téléchargements/pysparse-1.1.1/Test$ python
> testSuperlu.py
> Test RelErr Tol nnz(A) nnz(L+U) Fact Solve
> -------------------------------------------------------------------
> poi1d-dflt 1.73e-18 2.25e-07 99999 299936 0.06 0.00
> poi1d-size 0.00e+00 2.25e-07 99999 299936 0.03 0.00
> . poi1d-relx 8.64e-19 2.25e-07 99999 301354 0.06 0.00
> . poi1d-trsh 1.73e-18 2.25e-07 99999 299936 0.05 0.00
> poi1d-prm0 1.44e-12 2.25e-07 99999 199998 0.02 0.00
> . poi1d-prm1 1.78e-12 2.25e-07 99999 199998 0.02 0.00
> . poi1d-prm2 1.73e-18 2.25e-07 99999 299936 0.05 0.00
> poi1d-prm3 1.44e-12 2.25e-07 99999 200000 0.03 0.00
> . poi2d-dftl 3.49e-16 3.60e-12 119600 1916364 0.15 0.01
> . poi2d-size 3.62e-16 3.60e-12 119600 1916364 0.13 0.01
> poi2d-relx 3.50e-16 3.60e-12 119600 2029678 0.17 0.01
> . poi2d-trsh 3.49e-16 3.60e-12 119600 1916364 0.16 0.00
> . poi2d-prm0 2.69e-15 3.60e-12 119600 16000398 4.14 0.04
> poi2d-prm1 8.13e-16 3.60e-12 119600 3506204 0.26 0.01
> . poi2d-prm2 3.49e-16 3.60e-12 119600 1916364 0.16 0.01
> . poi2d-prm3 2.01e-15 3.60e-12 119600 3472176 026 0.01
>
> spdgs-trsh 4.44e-16 2.22e-14 29998 59946 0.01 0.00
> . spdgs-prm0 4.44e-16 2.22e-14 29998 40000 0.00 0.00
> . spdgs-prm1 4.44e-16 2.22e-14 29998 40002 0.01 0.00
> spdgs-prm3 4.44e-16 2.22e-14 29998 40002 0.01 0.00
> .
> ----------------------------------------------------------------------
> Ran 20 tests in 6.355s
>
> OK
> Erreur de segmentation
>
>
>
>
> "Erreur de segmentation" means segmentation fault!
>
>
>
> On Wed, Feb 23, 2011 at 8:46 PM, Jonathan Guyer <[email protected]> wrote:
>
>>
>>
>> On Feb 23, 2011, at 12:35 PM, Julien Derr wrote:
>>
>> > Unfortunately it seems there is an additional problem as it creates now
>> a segmentation fault :
>> :
>> :
>> > [derr-Precision-T1500:09248] [ 1]
>> /usr/lib/libsuperlu.so.3(dpivotL+0x16d) [0x7fa5dac6fa9d]
>> > [derr-Precision-T1500:09248] [ 2] /usr/lib/libsuperlu.so3(dgstrf+0x609)
>> [0x7fa5dac69339]
>> > [derr-Precision-T1500:09248] [ 3]
>> /usr/lib/python2.6/dist-packages/pysparse/superlu.so(+0x213f)
>> [0x7fa5db8e113f]
>>
>>
>> How frustrating.
>>
>> This looks to be failing in the PySparse LU solver.
>>
>>
>> Can you run the PySparse tests by themselves?
>>
>> http://pysparse.sourceforge.net/introduction.html#testing-pysparse
>>
>>
>> How large is the mesh you are solving?
>>
>>
>> Does the simpletrenchSystem.py script run acceptably?
>>
>>
>>
>>
>>
>>
>
