[BEAM-9085] Performance regression in np.random.RandomState() skews performance test results across Python 2/3 on Dataflow

• Type: Bug
• Status: Resolved
• Priority: P2
• Resolution: Fixed
• Fix Version/s: Not applicable
• Component/s: testing
• Assignee: Kamil Wasilewski
• Reporter: Kamil Wasilewski
• Created: 10/Jan/20 12:15
• Updated: 24/Jul/20 19:52
• Resolved: 15/Apr/20 21:16

Description

Tests show that the performance of core Beam operations in Python 3.x on Dataflow can be a few time slower than in Python 2.7. We should investigate what's the cause of the problem.

Currently, we have one ParDo test that is run both in Py3 and Py2 [1]. A dashboard with runtime results can be found here [2].

[1] sdks/python/apache_beam/testing/load_tests/pardo_test.py

[2] https://apache-beam-testing.appspot.com/explore?dashboard=5678187241537536

Issue Links

• is a child of: BEAM-8671 Migrate Python version to 3.7
• links to: GitHub Pull Request #10885
• links to: GitHub Pull Request #10924
• links to: GitHub Pull Request #11092