[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15420195#comment-15420195
]
Ben Manes commented on KAFKA-3973:
--
Measuring on both put() and removeEldestEntry() is slow, as you will
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15420061#comment-15420061
]
Bill Bejeck commented on KAFKA-3973:
Attaching JMH benchmark results, removing the previous
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15412838#comment-15412838
]
Ben Manes commented on KAFKA-3973:
--
It would also be impactful to consider the policy's hit rate. If a
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15411821#comment-15411821
]
Bill Bejeck commented on KAFKA-3973:
I used JMH to benchmark the performance of caching bytes vs
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15394761#comment-15394761
]
Bill Bejeck commented on KAFKA-3973:
[~ijuma]
I re-ran the tests with no instrumentation using the
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15393079#comment-15393079
]
Bill Bejeck commented on KAFKA-3973:
[~enothereska] Will do, just need to finish up a few details.
>
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15392234#comment-15392234
]
Eno Thereska commented on KAFKA-3973:
-
[~bbejeck] Thank you for the results. As part of this JIRA I'm
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15391766#comment-15391766
]
Bill Bejeck commented on KAFKA-3973:
Yes I ran the tests using instrumentation (-javaagent:), sorry I
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15391519#comment-15391519
]
Michael Noll commented on KAFKA-3973:
-
+1 to having repeatable tests/benchmarks (e.g. with JMH).
>
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15391502#comment-15391502
]
Ismael Juma commented on KAFKA-3973:
Also, MemoryMeasurer has the following enum:
{code}
public
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15391500#comment-15391500
]
Ismael Juma commented on KAFKA-3973:
"These benchmarks should be re-written with JMH
[
https://issues.apache.org/jira/browse/KAFKA-3973?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15391229#comment-15391229
]
Bill Bejeck commented on KAFKA-3973:
The results for the LRU investigation are below. There were three
12 matches
Mail list logo
