[
https://issues.apache.org/jira/browse/FLINK-8172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16272437#comment-16272437
]
ASF GitHub Bot commented on FLINK-8172:
---------------------------------------
Github user NicoK commented on a diff in the pull request:
https://github.com/apache/flink/pull/5104#discussion_r154024973
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/io/network/api/serialization/SpanningRecordSerializer.java
---
@@ -107,6 +110,7 @@ public SerializationResult addRecord(T record) throws
IOException {
@Override
public SerializationResult setNextBuffer(Buffer buffer) throws
IOException {
--- End diff --
Actually, whether we are the sole owner of the `buffer` now depends on the
outside and this adds a pretty heavy implicit assumption. Why not make the
`SpanningRecordSerializer` work with the `BufferProvider` of the
`targetPartition` directly, e.g. either by adding it at construction time or as
a parameter. We'd then probably have to rename the `setNextBuffer` method but
can keep the logic and make sure we are the sole owner.
> Remove unnecessary synchronisation in RecordSerializer
> ------------------------------------------------------
>
> Key: FLINK-8172
> URL: https://issues.apache.org/jira/browse/FLINK-8172
> Project: Flink
> Issue Type: Improvement
> Components: Network
> Affects Versions: 1.4.0, 1.3.2
> Reporter: Piotr Nowojski
> Assignee: Piotr Nowojski
> Fix For: 1.5.0
>
>
> While writing the records, RecordSerializer is the only owner of the `Buffer`
> into which data are written. Yet we are synchronisation twice per record
> while accessing MemorySegment. Removing this synchronisation speeds up the
> Network throughput in point to point benchmark by a factor of two (from
> ~12500records/ms up to 23000 records/ms).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
