[
https://issues.apache.org/jira/browse/CASSANDRA-15213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17011952#comment-17011952
]
Jordan West commented on CASSANDRA-15213:
-----------------------------------------
Ok I figured out what was going on. My linked test was generating only large
values and the approximation of only the last bucket is very inaccurate as
value grows but also easy to account for. Fixing this the test confirms your
findings (sometimes I see a distance of 4 indexes but still a huge improvement):
Sample from 10k test run:
{code:java}
model = 80, estimate = 83
model = 83, estimate = 86
model = 81, estimate = 84
model = 82, estimate = 85
model = 69, estimate = 72
model = 79, estimate = 82
model = 81, estimate = 84
model = 79, estimate = 82
model = 75, estimate = 79
model = 83, estimate = 86
model = 77, estimate = 81
model = 72, estimate = 75
model = 84, estimate = 88
model = 61, estimate = 64
{code}
> DecayingEstimatedHistogramReservoir Inefficiencies
> --------------------------------------------------
>
> Key: CASSANDRA-15213
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15213
> Project: Cassandra
> Issue Type: Bug
> Components: Observability/Metrics
> Reporter: Benedict Elliott Smith
> Assignee: Jordan West
> Priority: Normal
> Fix For: 4.0-beta
>
>
> * {{LongAdder}} introduced to trunk consumes 9MiB of heap without user
> schemas, and this will grow significantly under contention and user schemas
> with many tables. This is because {{LongAdder}} is a very heavy class
> designed for single contended values.
> ** This can likely be improved significantly, without significant loss of
> performance in the contended case, by simply increasing the size of our
> primitive backing array and providing multiple buckets, with each thread
> picking a bucket to increment, or simply multiple backing arrays. Probably a
> better way still to do this would be to introduce some competition detection
> to the update, much like {{LongAdder}} utilises, that increases the number of
> backing arrays under competition.
> ** To save memory this approach could partition the space into chunks that
> are likely to be updated together, so that we do not need to duplicate the
> entire array under competition.
> * Similarly, binary search is costly and a measurable cost as a share of the
> new networking work (without filtering it was > 10% of the CPU used overall).
> We can compute an approximation floor(log2 n / log2 1.2) extremely cheaply,
> to save the random memory access costs.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
