[
https://issues.apache.org/jira/browse/AVRO-3527?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17958772#comment-17958772
]
ASF subversion and git services commented on AVRO-3527:
-------------------------------------------------------
Commit c880f4729e0dcb99bab82dd5b5efb15f0da55890 in avro's branch
refs/heads/main from Steven Aerts
[ https://gitbox.apache.org/repos/asf?p=avro.git;h=c880f4729e ]
AVRO-3527: codegen equals and hashCode for Records (#1708)
Update the compiler to generate the implementation of the `.equals()`
and `.hashCode() function, instead of relying on the
implementation of GenericData. This improves the performance of
those functions significantly.
The generated implementations are factor 10 to 20 faster for
`.equals()` and a factor 5 to 10 for `.hashCode()`.
Result of Perf test before the change:
```
Benchmark Mode Cnt Score Error Units
SpecficTest.equals thrpt 3 12598610.194 +/- 11160265.279 ops/s
SpecficTest.hashCode thrpt 3 24729446.862 +/- 29051332.794 ops/s
```
Results using generated functions:
```
Benchmark Mode Cnt Score Error Units
SpecficTest.equals thrpt 3 211314296.950 +/- 104154793.126 ops/s
SpecficTest.hashCode thrpt 3 180349506.632 +/- 143639246.771 ops/s
```
Signed-off-by: Steven Aerts <[email protected]>
> Generated equals() and hashCode() for SpecificRecords
> -----------------------------------------------------
>
> Key: AVRO-3527
> URL: https://issues.apache.org/jira/browse/AVRO-3527
> Project: Apache Avro
> Issue Type: Improvement
> Components: java
> Reporter: Steven Aerts
> Assignee: Christophe Le Saec
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.12.0
>
> Attachments: equals_hashcode_after.txt, equals_hashcode_before.txt,
> flame_graph.jpeg
>
> Time Spent: 3h 50m
> Remaining Estimate: 0h
>
> When profiling our production system, we found that it was spending almost
> 40% of its overall time in the {{SpecificRecordBase.hashCode()}} and
> {{SpecificRecordBase.equals()}} implementations.
> In some sections of its logic we see that almost all time is spend in those
> function, as can be seen in attached flame graph (blue "pyramids")
> !flame_graph.jpeg|width=385,height=99!
> By generating the {{.equals()}} and {{.hashCode()}} all this overhead
> disappeared and this application became 35% faster overall.
> Also on other AVRO heavy applications we saw noticeable performance gains
> where we hadn't expect them due to this improvement.
> A generated implementation of {{.hashCode()}} becomes 5 to 10 times faster
> than its generic counterpart. For {{.equals()}} it is 10 to 20 times faster.
> Which is also visible in the attached JMH benchmarks.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
