[
https://issues.apache.org/jira/browse/CASSANDRA-15396?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17098155#comment-17098155
]
David Capwell edited comment on CASSANDRA-15396 at 5/3/20, 1:26 AM:
--------------------------------------------------------------------
I rebased to trunk and started perf testing this against a k/v workflow where
the value ranges between 100-102400 bytes, in this test I see a noticeable
difference between trunk and this patch (p99 of 64ms on trunk vs 25ms with the
patch). I plan to test this more with different value lengths and with other
profiles to make sure there isn't a regression under different access patterns.
was (Author: dcapwell):
I rebased to trunk and started perf testing this against a k/v workflow where
the value ranges between 100-102400 bytes, in this test I see a noticeable
difference between trunk and this patch (p99 of 64ms on trunk, and 25ms with
the patch). I plan to test this more with different value lengths and with
other profiles to make sure there isn't a regression under different access
patterns.
> RandomAccessReader does not override skip or skipBytes so will do a lot of
> disk io when n is large
> --------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-15396
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15396
> Project: Cassandra
> Issue Type: Improvement
> Components: Local/SSTable
> Reporter: David Capwell
> Assignee: David Capwell
> Priority: Normal
> Attachments: Histogram-2.png, Histogram.png
>
>
> RandomAccessReader does not override skip or skipBytes which becomes a
> problem when the size of n (the bytes to skip) is larger than a single
> buffer; in these cases we can rely on seek to avoid the extra disk io.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
