[SciPy-Dev] Gmres with Sparse Complex Matrices gives Segmentation fault (or NULL result)

[Martin Campos Pinto]

Hi,

I have just tried with scipy version 0.11.0, installed with mac ports 
and gmres runs without crashing.
Since scipy 0.10.1 was the package provided with EPD I first thought it 
should run fine...

Thank you,
Martin

-- ps. Also, just to check I have run your test `scipy.test("full", 
verbose=2)' with scipy 0.11.0 and there still is a problem.


Python 2.7.3 (default, Oct 22 2012, 06:12:28)
[GCC 4.2.1 (Apple Inc. build 5666) (dot 3)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
 >>> import scipy
 >>> scipy.test("full", verbose=2)
Running unit tests for scipy
NumPy version 1.6.2
NumPy is installed in 
/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/numpy
SciPy version 0.11.0
SciPy is installed in 
/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/scipy
Python version 2.7.3 (default, Oct 22 2012, 06:12:28) [GCC 4.2.1 (Apple 
Inc. build 5666) (dot 3)]
nose version 1.2.1
Tests cophenet(Z) on tdist data set. ... ok
Tests cophenet(Z, Y) on tdist data set. ... ok
Tests correspond(Z, y) on linkage and CDMs over observation sets of 
different sizes. ... ok
Tests correspond(Z, y) on linkage and CDMs over observation sets of 
different sizes. Correspondance should be false. ... ok
Tests correspond(Z, y) on linkage and CDMs over observation sets of 
different sizes. Correspondance should be false. ... ok
Tests correspond(Z, y) with empty linkage and condensed distance matrix. 
... ok
Tests num_obs_linkage with observation matrices of multiple sizes. ... ok
Tests dendrogram calculation on single linkage of the tdist data set. ... ok
Tests fcluster(Z, criterion='maxclust', t=2) on a random 3-cluster data 
set. ... ok

(...)

test_exclusive_end (test_slice_handler.TestBuildSliceAtom) ... ok
match slice from a[1:] ... ok
match slice from a[1::] ... ok
match slice from a[1:2] ... ok
match slice from a[1:2:] ... ok
match slice from a[1:2:3] ... ok
match slice from a[1::3] ... ok
match slice from a[:] ... ok
match slice from a[::] ... ok
match slice from a[:2] ... ok
match slice from a[:2:] ... ok
match slice from a[:2:3] ... ok
match slice from a[:1+i+2:] ... ok
match slice from a[0] ... ok
match slice from a[::3] ... ok
transform a[:] to slice notation ... ok
transform a[:,:] = b[:,1:1+2:3] *(c[1-2+i:,:] - c[:,:]) ... ok
test_type_match_array (test_standard_array_spec.TestArrayConverter) ... ok
test_type_match_int (test_standard_array_spec.TestArrayConverter) ... ok
test_type_match_string (test_standard_array_spec.TestArrayConverter) ... ok

======================================================================
FAIL: test_iv_cephes_vs_amos_mass_test (test_basic.TestBessel)
----------------------------------------------------------------------
Traceback (most recent call last):
   File 
"/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/scipy/special/tests/test_basic.py", 
line 1659, in test_iv_cephes_vs_amos_mass_test
     assert_(dc[k] < 2e-7, (v[k], x[k], special.iv(v[k], x[k]), 
special.iv(v[k], x[k]+0j)))
   File 
"/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/numpy/testing/utils.py", 
line 34, in assert_
     raise AssertionError(msg)
AssertionError: (189.2947429454936, 3.0238805556481037, 
4.089165443940765e-317, 0j)

----------------------------------------------------------------------
Ran 6217 tests in 1046.384s

FAILED (KNOWNFAIL=15, SKIP=32, failures=1)
<nose.result.TextTestResult run=6217 errors=0 failures=1>





Le 11/24/12 10:40 AM, Pauli Virtanen a écrit :
> 24.11.2012 07:50, Martin Campos Pinto kirjoitti:
> [clip]
>> I have looked on the web for solutions but haven't found any. Some very old posts
>> indicate similar errors but they don't come with an answer,
>> and I imagine that if those were due to bugs, they should have been fixed by now...
>>
>> Am I doing something stupid here, or is that a real problem ? Is somebody aware of a solution ?
>> (I am using scipy version 0.10.1)
> I don't get a crash with this code example. Valgrind also does not show
> anything, so I suppose this is something occurring only on OSX.
>
> There are some known issues on OSX with Accelerate, but I'm not sure if
> they are in play here.
>
> Can you run `scipy.test("full", verbose=2)` without seeing problems?
>

-- 
Martin Campos Pinto
http://www-irma.u-strasbg.fr/~campos