[
https://issues.apache.org/jira/browse/IGNITE-22941?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Konstantin Orlov updated IGNITE-22941:
--------------------------------------
Fix Version/s: 3.0.0-beta2
> Improve performance of TupleMarshaller
> --------------------------------------
>
> Key: IGNITE-22941
> URL: https://issues.apache.org/jira/browse/IGNITE-22941
> Project: Ignite
> Issue Type: Improvement
> Reporter: Konstantin Orlov
> Assignee: Konstantin Orlov
> Priority: Major
> Labels: ignite-3
> Fix For: 3.0.0-beta2
>
> Attachments: flamegraph.html
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> According to [^flamegraph.html], statistics gathering for buffer
> preallocation takes ~78.9% (2.91% out of 3.69% of the whole graph) of an
> entire tuple marshaling. Perhaps, it would be better to switch to best effort
> estimation, and re-allocate buffer when we underestimate actual number of
> bytes required to marshal the tuple.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
