This was gh-13134. @r-devulap could you have a look at this?

yup, taking a look.

Started a PR #13415 to fix this. Will fix the issue with log soon.

PR #13415 is merged, perhaps this issue can be closed?

@r-devulap just before it gets lost, was this comment a possible problem with the current fix: #13415 (comment)

yes, unfortunately I don't have access to a mac and I cannot reproduce that problem on my SKX with ubuntu and gcc/clang. But since that portion of test test_exp_values only cares about verifying the output of exp, I can may be ignore floating point errors by updating the test with np.errstate(all='ignore'). Would that be acceptable?

@tylerjereddy Is the occasional RuntimeWarning: invalid value encountered in exp on macOS still a valid concern or have they dissapeared?

@mattip Yes, I still get the following test failure output from the latest master branch commit (8a421d9) when running full test suite on my Mac laptop:

________________________ TestSpecialFloats.test_exp_values _________________________
[gw1] darwin -- Python 3.6.5 /Users/treddy/miniconda3/envs/numpy_dev_py36/bin/python

self = <numpy.core.tests.test_umath.TestSpecialFloats object at 0x11bccdac8>

    def test_exp_values(self):
        x = [np.nan,  np.nan, np.inf, 0.]
        y = [np.nan, -np.nan, np.inf, -np.inf]
        for dt in ['f', 'd', 'g']:
            xf = np.array(x, dtype=dt)
            yf = np.array(y, dtype=dt)
>           assert_equal(np.exp(yf), xf)
E           RuntimeWarning: invalid value encountered in exp

dt         = 'f'
self       = <numpy.core.tests.test_umath.TestSpecialFloats object at 0x11bccdac8>
x          = [nan, nan, inf, 0.0]
xf         = array([nan, nan, inf,  0.], dtype=float32)
y          = [nan, nan, inf, -inf]
yf         = array([ nan,  nan,  inf, -inf], dtype=float32)

numpy/core/tests/test_umath.py:659: RuntimeWarning
1 failed, 8641 passed, 63 skipped, 15 xfailed, 3 xpassed in 120.79 seconds

I don't think it is occasional either--I haven't see any fluke passes lately. I could use pytest-repeat if you really need likelihood of failure, but seems to happen all the time for me on mac.

Hmm, this is unfortunate with 1.17 coming on soon.

If I am not wrong, this particular issue was fixed. The test failure seems to be because of a bug in clang7.0 and a new issue was opened to address it (see #13586). Although, I am not fully sure how to address it..

@juliantaylor Just want to check if this can be closed.

Closing as it seems to be fixed. I expect it will come up again if it persists.