sdc icon indicating copy to clipboard operation
sdc copied to clipboard

python -m unittest fails with large number of processes

Open TocarIP opened this issue 7 years ago • 2 comments

Running mpiexec -n 4 python -m unittest on master fails with:

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 17975 RUNNING AT antsatel03
=   EXIT CODE: 139
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================

We need to investigate this

TocarIP avatar Sep 10 '18 22:09 TocarIP

Guilty test is hpat.tests.test_hiframes.TestHiFrames.test_crosstab_parallel1

TocarIP avatar Sep 10 '18 22:09 TocarIP

In many cases, parallel crashes are caused by some parallel routine called with small chunks of data on some processor, which triggers some corner cases.

ehsantn avatar Sep 10 '18 23:09 ehsantn