[
https://issues.apache.org/jira/browse/BEAM-10409?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yifan Mai reassigned BEAM-10409:
--------------------------------
Assignee: Yifan Mai
> Add combiner packing to graph optimizer phases
> ----------------------------------------------
>
> Key: BEAM-10409
> URL: https://issues.apache.org/jira/browse/BEAM-10409
> Project: Beam
> Issue Type: Improvement
> Components: runner-core
> Reporter: Yifan Mai
> Assignee: Yifan Mai
> Priority: P2
> Time Spent: 7.5h
> Remaining Estimate: 0h
>
> Some use cases of Beam (e.g. [TensorFlow
> Transform|https://github.com/tensorflow/transform]) create thousands of
> Combine stages with a common parent. The large number of stages can cause
> performance issues on some runners. To alleviate, a graph optimization phase
> could be added to the translations module that packs compatible Combine
> stages into a single stage.
> The graph optimization for CombinePerKey would work as follows: If
> CombinePerKey stages have a common input, one input each, and one output
> each, pack the stages into a single stage that runs all CombinePerKeys and
> outputs resulting tuples to a new PCollection. A subsequent stage unpacks
> tuples from this PCollection and sends them to the original output
> PCollections.
> There is an additional issue with supporting this for CombineGlobally:
> because of the intermediate KeyWithVoid stage between the CombinePerKey
> stages and the input stage, the CombinePerKey stages do not have a common
> input stage, and cannot be packed. To support CombineGlobally, a common
> sibling elimination graph optimization phase can be used to combine the
> KeyWithVoid stages. After this, the CombinePerKey stages would have a common
> input and can be packed.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
