flaky test: test_operator.test_depthwise_convolution

[mseth10]

http://jenkins.mxnet-ci.amazon-ml.com/blue/organizations/jenkins/mxnet-validation%2Funix-cpu/detail/PR-14016/4/pipeline/

ERROR: test_operator.test_depthwise_convolution

Traceback (most recent call last):
File "/usr/local/lib/python2.7/dist-packages/nose/case.py", line 197, in runTest
self.test(*self.arg)
File "/work/mxnet/tests/python/unittest/common.py", line 173, in test_new
orig_test(*args, **kwargs)
File "/work/mxnet/tests/python/unittest/test_operator.py", line 1676, in test_depthwise_convolution
np.testing.assert_allclose(arr1.asnumpy(), arr2.asnumpy(), rtol=1e-3, atol=1e-3)
File "/work/mxnet/python/mxnet/ndarray/ndarray.py", line 1995, in asnumpy
ctypes.c_size_t(data.size)))
File "/work/mxnet/python/mxnet/base.py", line 252, in check_call
raise MXNetError(py_str(_LIB.MXGetLastError()))
MXNetError: [01:40:38] src/operator/nn/mkldnn/mkldnn_base.cc:576: Check failed: similar

NOTE: The test fails on CPU when MKLDNN is enabled and passes when MKL-DNN is disabled. So, the issue is mostly due to MKL-DNN.

[mxnet-label-bot]

Hey, this is the MXNet Label Bot.
Thank you for submitting the issue! I will try and suggest some labels so that the appropriate MXNet community members can help resolve it.
Here are my recommended labels: Test, Flaky

[vdantu]

@mxnet-label-bot add [mkldnn, test, flaky]

[ZhennanQin]

@mseth10, Thanks for reporting this. Confirmed it's caused by mkldnn bug. The fix is already in mkldnn master, but not available in its latest release. We will upgrade mkldnn when its next release comes out.

[mseth10]

@ZhennanQin Thanks for confirming it's an MKLDNN bug. Good to know that it's fixed. Can you also give an ETA on when the next release containing the fix be rolled out?

[pengzhao-intel]

Resolved so closing.