[
https://issues.apache.org/jira/browse/IGNITE-25147?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17944721#comment-17944721
]
Ivan Bessonov commented on IGNITE-25147:
Tested using the following scenario (same as in
https://issues.apache.org/jira/browse/IGNITE-16879):
* {{putAll}} transactions with 1000 entries.
* 3 replicas, 1 client with many threads.
* Just enough load to always have checkpoints with _"too many dirty pages"_
reason.
h3. Results
Before merging IGNITE-16879:
* Throughput in main is *198946* (don't know why so low, must be a fluke,
should be around 201-202k).
* Throughput with the fix is {*}241756{*}, which is roughly *20%* higher.
After manually merging IGNITE-16879 to IGNITE-25147:
* Throughput in main is {*}201118{*}.
* Throughput in my branch is {*}249376{*}, which is roughly *23-24%* higher.
After merging IGNITE-16879 to main:
* Throughput in main is {*}219099{*}.
* Throughput in my branch is {*}244857{*}, which is roughly *11-12%* higher.
h3. Graphs
Main:
!Screenshot from 2025-04-15 16-44-20.png|width=683,height=198!
This branch:
!Screenshot from 2025-04-15 16-44-23.png|width=683,height=198!
> Update default throttling bounds
>
>
> Key: IGNITE-25147
> URL: https://issues.apache.org/jira/browse/IGNITE-25147
> Project: Ignite
> Issue Type: Improvement
>Reporter: Ivan Bessonov
>Assignee: Ivan Bessonov
>Priority: Major
> Labels: ignite-3
> Attachments: Screenshot from 2025-04-15 16-44-20.png, Screenshot from
> 2025-04-15 16-44-23.png
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Lower bound of 3% is way too small, I propose raising it up to 50%.
> Upper bound could also technically be relaxed, although it should not affect
> the system too much. Anyway, let's try having a 90% there and run some
> benchmarks.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
