[ https://issues.apache.org/jira/browse/BEAM-9085?focusedWorklogId=419086&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-419086 ]
ASF GitHub Bot logged work on BEAM-9085: ---------------------------------------- Author: ASF GitHub Bot Created on: 09/Apr/20 02:33 Start Date: 09/Apr/20 02:33 Worklog Time Spent: 10m Work Description: tvalentyn commented on pull request #11092: [BEAM-9085] Fix performance regression in SyntheticSource URL: https://github.com/apache/beam/pull/11092#discussion_r405921080 ########## File path: sdks/python/apache_beam/testing/synthetic_pipeline.py ########## @@ -61,6 +65,35 @@ np = None +class _Random(Random): + """A subclass of `random.Random` from the Python Standard Library that + provides a method returning random bytes of arbitrary length. + """ + + # `numpy.random.RandomState` does not provide `random()` method, we keep this + # for compatibility reasons. + random_sample = Random.random + + def bytes(self, length): + """Returns random bytes. + + Args: + length (int): Number of random bytes. + """ + n = length // 8 + 1 + # pylint: disable=map-builtin-not-iterating + return struct.pack( + '{}Q'.format(n), + *map(self.getrandbits, itertools.repeat(64, n)))[:length] Review comment: Since we don't need py2 compatibility anymore, consider using `to_bytes`. Here's an equivalent for chunk_size=8. It seems to be somewhat slower than current method (perhaps since I'm not using `map+repeat()`), but with larger `chuck_size`, seems to be more efficient. Large chunk size may be less efficient for short bytesequences. ``` chunk_size_bytes = 8 // TBD - larger chunks seem to improve performance. chunk_size_bits = chunk_size_bytes * 8 num_chunks = length // chunk_size_bytes + 1 return b''.join([self.getrandbits(chunk_size_bits).to_bytes(chunk_size_bytes, sys.byteorder) for _ in range(num_chunks)])[:length] ``` If you decide to keep current implementation - please add a comment explaining the mechanics for readers not familiar with this code (we generate 8-byte stings, and then fit them into a representation of C++'s long-long, which also takes up 8 bytes). ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking ------------------- Worklog Id: (was: 419086) Time Spent: 8h (was: 7h 50m) > Performance regression in np.random.RandomState() skews performance test > results across Python 2/3 on Dataflow > -------------------------------------------------------------------------------------------------------------- > > Key: BEAM-9085 > URL: https://issues.apache.org/jira/browse/BEAM-9085 > Project: Beam > Issue Type: Bug > Components: testing > Reporter: Kamil Wasilewski > Assignee: Kamil Wasilewski > Priority: Major > Time Spent: 8h > Remaining Estimate: 0h > > 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 -- This message was sent by Atlassian Jira (v8.3.4#803005)