[
https://issues.apache.org/jira/browse/BEAM-3645?focusedWorklogId=267864&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-267864
]
ASF GitHub Bot logged work on BEAM-3645:
----------------------------------------
Author: ASF GitHub Bot
Created on: 26/Jun/19 19:02
Start Date: 26/Jun/19 19:02
Worklog Time Spent: 10m
Work Description: Hannah-Jiang commented on pull request #8872:
[BEAM-3645] add ParallelBundleManager
URL: https://github.com/apache/beam/pull/8872#discussion_r297809318
##########
File path: sdks/python/apache_beam/runners/portability/fn_api_runner.py
##########
@@ -1506,6 +1534,39 @@ def process_bundle(self, inputs, expected_outputs):
return result, split_results
+class ParallelBundleManager(BundleManager):
+
+ def process_bundle(self, inputs, expected_outputs, num_workers=None):
Review comment:
I added it in case we want to change num_workers for a certain bundle only,
instead of changing at BundleManager level. Do we have a case that process
multi bundles parallel with different num_workers? I now changed back to
support only self._num_workers, I would like to grab your advice on if we even
need to consider such cases.
----------------------------------------------------------------
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:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 267864)
Time Spent: 13h (was: 12h 50m)
> Support multi-process execution on the FnApiRunner
> --------------------------------------------------
>
> Key: BEAM-3645
> URL: https://issues.apache.org/jira/browse/BEAM-3645
> Project: Beam
> Issue Type: Improvement
> Components: sdk-py-core
> Affects Versions: 2.2.0, 2.3.0
> Reporter: Charles Chen
> Assignee: Hannah Jiang
> Priority: Major
> Time Spent: 13h
> Remaining Estimate: 0h
>
> https://issues.apache.org/jira/browse/BEAM-3644 gave us a 15x performance
> gain over the previous DirectRunner. We can do even better in multi-core
> environments by supporting multi-process execution in the FnApiRunner, to
> scale past Python GIL limitations.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
