Hi Ole,
The TensorFlow tests are known to be a PITA (to put it mildly).
Alexander (who I included in CC since he's not on the mailing list) has
spent quite a bit of time weeding things out there, so maybe he has
specific suggestions.
An easy way out if you're willing to ignore failing tests is to use "eb
--skip-test-step" to do the installation.
You can also filter out specific tests via the test_targets custom
easyconfig parameter for TensorFlow, which is already used in
TensorFlow-2.4.1-foss-2020b.eb .
From the output you shared it's not clear how badly that specific test
is failing though, you should be able to dig up more information from
the EasyBuild log file on that...
regards,
Kenneth
On 26/05/2021 16:07, Ole Holm Nielsen wrote:
I'm trying to build TensorFlow with EB 4.3.4 but get an error:
$ eb TensorFlow-2.4.1-fosscuda-2020b.eb
--cuda-compute-capabilities=8.0,8.6 --tmpdir=/scratch/modules
(lines deleted)
== installing extension TensorFlow 2.4.1 (28/28)...
== configuring...
== building...
== testing...
== FAILED: Installation ended unsuccessfully (build directory:
/run/user/983/eb_build/TensorFlow/2.4.1/fosscuda-2020b): build failed
(first 300 chars): At least 1 cpu tests failed:
//tensorflow/core/common_runtime:graph_constructor_test (took 43 min 58
sec)
== Results of the build can be found in the log file(s)
/scratch/modules/eb-KPZu0P/easybuild-TensorFlow-2.4.1-20210526.144651.PuIWy.log
ERROR: Build of
/home/modules/software/EasyBuild/4.3.4/easybuild/easyconfigs/t/TensorFlow/TensorFlow-2.4.1-fosscuda-2020b.eb
failed (err: 'build failed (first 300 chars): At least 1 cpu tests
failed:\n//tensorflow/core/common_runtime:graph_constructor_test')
The EB log file reports an error:
//tensorflow/core/common_runtime:graph_constructor_test FAILED
TO BUILD
and the log file ends with:
Executed 137 out of 814 tests: 137 tests pass, 1 fails to build and 676
were skipped.
FAILED: Build did NOT complete successfully
== 2021-05-26 15:30:49,719 build_log.py:169 ERROR EasyBuild crashed with
an error (at easybuild/base/exceptions.py:124 in __init__): At least 1
cpu tests failed:
//tensorflow/core/common_runtime:graph_constructor_test (at
easybuild/easyblocks/t/tensorflow.py:973 in test_step)
== 2021-05-26 15:30:49,719 filetools.py:1810 INFO Removing lock
/home/modules/software/.locks/_home_modules_software_TensorFlow_2.4.1-fosscuda-2020b.lock...
== 2021-05-26 15:30:49,721 filetools.py:347 INFO Path
/home/modules/software/.locks/_home_modules_software_TensorFlow_2.4.1-fosscuda-2020b.lock
successfully removed.
== 2021-05-26 15:30:49,721 filetools.py:1814 INFO Lock removed:
/home/modules/software/.locks/_home_modules_software_TensorFlow_2.4.1-fosscuda-2020b.lock
== 2021-05-26 15:30:49,721 easyblock.py:3414 WARNING build failed (first
300 chars): At least 1 cpu tests failed:
//tensorflow/core/common_runtime:graph_constructor_test
== 2021-05-26 15:30:49,721 easyblock.py:298 INFO Closing log for
application name TensorFlow version 2.4.1
Can anyone suggest a fix for this issue?
Thanks,
Ole
