FAIL: pygpu.tests.test_elemwise.test_divmod('int8', 'float32', (50,))

Question

FAIL: pygpu.tests.test_elemwise.test_divmod('int8', 'float32', (50,))

Opened this issue 7 years ago · 4 comments

After installing libgpuarray and runnig the tests I got this error:

$ DEVICE=cuda python2 -c "import pygpu.tests.test_elemwise; pygpu.tests.test_elemwise.divmod_array('int8', 'float32', (50,))"
*** Testing for GeForce GTX TITAN X
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/home/elacruzca/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/support.py", line 39, in f
    func(*args, **kwargs)
  File "/home/elacruzca/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/test_elemwise.py", line 236, in divmod_array
    assert numpy.allclose(out_c[1], numpy.asarray(out_g[1]))
AssertionError

No other test fails.

pygpu version 0.6.5

Answer 1 · 2017-05-16T12:34:56.000Z

Can you print the values? Le ven. 12 mai 2017 19:28, helq <notifications@github.com> a écrit :

…

After installing libgpuarray and runnig the tests I got this error: $ DEVICE=cuda python2 -c "import pygpu.tests.test_elemwise; pygpu.tests.test_elemwise.divmod_array('int8', 'float32', (50,))" *** Testing for GeForce GTX TITAN X Traceback (most recent call last): File "<string>", line 1, in <module> File "/home/elacruzca/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/support.py", line 39, in f func(*args, **kwargs) File "/home/elacruzca/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/test_elemwise.py", line 236, in divmod_array assert numpy.allclose(out_c[1], numpy.asarray(out_g[1])) AssertionError No other test fails. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#433>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AALC-34dwGK0g3ITPEea27Hc_6n3PIiLks5r5OsngaJpZM4NZ4S4> .

Answer 2 · 2017-05-17T18:41:02.000Z

These are the values:

print out_c[1]
[ 2.          2.96456456  1.80960321  1.          2.          5.          1.
  0.37324905  4.          2.          2.          4.          5.          6.
  0.          1.09210575  2.92230988  4.          2.          1.31337512
  1.          2.          0.31895924  2.          4.          1.
  2.18828678  0.          0.35755444  0.          0.76005101  5.
  0.48180008  5.          2.69929767  6.          5.          0.          8.
  2.          5.          3.          0.          7.          9.
  1.02005863  2.          1.          1.64588308  0.27146184]

print numpy.asarray(out_g[1])
[ 2.          2.96456456  1.80960321  1.          2.          5.          1.
  0.37324905  4.          2.          2.          4.          5.          6.
         nan  1.09210575  2.92230988  4.          2.          1.31337512
  1.          2.          0.31895924  2.          4.          1.
  2.18828678         nan  0.35755444         nan  0.76005101  5.
  0.48180008  5.          2.69929767  6.          5.                 nan
  8.          2.          5.          3.                 nan  7.          9.
  1.02005863  2.          1.          1.64588308  0.27146184]

print out_g[1]
[ 2.          2.96456456  1.80960321  1.          2.          5.          1.
  0.37324905  4.          2.          2.          4.          5.          6.
         nan  1.09210575  2.92230988  4.          2.          1.31337512
  1.          2.          0.31895924  2.          4.          1.
  2.18828678         nan  0.35755444         nan  0.76005101  5.
  0.48180008  5.          2.69929767  6.          5.                 nan
  8.          2.          5.          3.                 nan  7.          9.
  1.02005863  2.          1.          1.64588308  0.27146184]

Answer 3 · 2017-06-26T15:58:01.000Z

It's probably just because nan != nan. Maybe we should adjust the input values to avoid nans.

Answer 4 · 2017-06-26T15:58:47.000Z

Or rather, we get nans instead of 0. That a bit worse.