On Tue, Jun 4, 2013 at 1:20 AM, Tim Burgess <tim.burgess@noaa.gov> wrote:

On Sat, 2013-06-01 at 20:09 -0400, Warren Weckesser wrote:


> I'm using Ubuntu 12.04, so I suspect I won't be the only one who sees
> these.
>

gcc on 12.04 (precise) should be 4.6.3

See
http://packages.ubuntu.com/search?keywords=gcc&searchon=names&suite=precise&section=all

Yes, that's what it is.   The python from the Anaconda package includes "[GCC 4.1.2 20080704 (Red Hat 4.1.2-52)] on linux2" in its banner, so I guess Anaconda is built with an even older gcc.

If I build python myself (with ubuntu's gcc 4.6.3), I get the same failure and two errors that I originally reported.  In this case, I used python 3.3.1:

$ python3 -c "import numpy as np; np.test('full')"

Running unit tests for numpy

NumPy version 1.8.0.dev-e9e490a
NumPy is installed in /home/warren/local_numpy/lib/python3.3/site-packages/numpy
Python version 3.3.1 (default, Apr 13 2013, 13:42:07) [GCC 4.6.3]
nose version 1.2.1

<snip>

======================================================================
ERROR: test_numeric.TestIsclose.test_ip_isclose_allclose([1e-08, 1, 1000020.0000000099], [0, nan, 1000000.0])
----------------------------------------------------------------------
Traceback (most recent call last):

  File "/home/warren/local_py331/lib/python3.3/site-packages/nose-1.2.1-py3.3.egg/nose/case.py", line 198, in runTest
    self.test(*self.arg)
  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/core/tests/test_numeric.py", line 1253, in tst_isclose_allclose

    assert_array_equal(isclose(x, y).all(), allclose(x, y), msg % (x, y))

  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/core/numeric.py", line 2008, in allclose

    return all(less_equal(abs(x-y), atol + rtol * abs(y)))
RuntimeWarning: invalid value encountered in absolute

======================================================================
ERROR: test_numeric.TestIsclose.test_ip_isclose_allclose(nan, [nan, nan, nan])
----------------------------------------------------------------------
Traceback (most recent call last):

  File "/home/warren/local_py331/lib/python3.3/site-packages/nose-1.2.1-py3.3.egg/nose/case.py", line 198, in runTest
    self.test(*self.arg)
  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/core/tests/test_numeric.py", line 1253, in tst_isclose_allclose

    assert_array_equal(isclose(x, y).all(), allclose(x, y), msg % (x, y))

  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/core/numeric.py", line 2008, in allclose

    return all(less_equal(abs(x-y), atol + rtol * abs(y)))
RuntimeWarning: invalid value encountered in absolute

======================================================================
FAIL: Test numpy dot with different order C, F
----------------------------------------------------------------------
Traceback (most recent call last):

  File "/home/warren/local_py331/lib/python3.3/site-packages/nose-1.2.1-py3.3.egg/nose/case.py", line 198, in runTest
    self.test(*self.arg)
  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/core/tests/test_blasdot.py", line 114, in test_dot_array_order
    assert_almost_equal(a.dot(a), a.T.dot(a.T).T, decimal=30)
  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/testing/utils.py", line 458, in assert_almost_equal

    return assert_array_almost_equal(actual, desired, decimal, err_msg)

  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/testing/utils.py", line 819, in assert_array_almost_equal

    header=('Arrays are not almost equal to %d decimals' % decimal))

  File "/home/warren/local_numpy/lib/python3.3/site-packages/numpy/testing/utils.py", line 652, in assert_array_compare

    raise AssertionError(msg)
AssertionError:
Arrays are not almost equal to 30 decimals

(mismatch 10.0%)
 x: array([[ 0.60970883,  1.6909554 , -1.0885194 , -1.82058004, -3.95746616,
         1.52435604, -0.59853062, -3.72278631,  3.82375941,  5.51367039],
       [-3.58154905, -2.0623123 , -0.06567267,  1.47373436,  2.60687462,...
 y: array([[ 0.60970883,  1.6909554 , -1.0885194 , -1.82058004, -3.95746616,
         1.52435604, -0.59853062, -3.72278631,  3.82375941,  5.51367039],
       [-3.58154905, -2.0623123 , -0.06567267,  1.47373436,  2.60687462,...

----------------------------------------------------------------------
Ran 5150 tests in 33.772s

FAILED (KNOWNFAIL=6, SKIP=18, errors=2, failures=1)

Warren

 

_______________________________________________
NumPy-Discussion mailing list
NumPy-Discussion@scipy.org
http://mail.scipy.org/mailman/listinfo/numpy-discussion