[
https://issues.apache.org/jira/browse/CASSANDRA-15215?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17437451#comment-17437451
]
Branimir Lambov commented on CASSANDRA-15215:
---------------------------------------------
A small suggestion: we also need a benchmark that randomly chooses a length, so
that the branch doesn't always hit the same target and isn't predictable.
Because in our code we call this encoding for many different types of values, I
expect unpredictable length to be common.
Also, because we are switching on the exact value of {{size}}, it makes sense
to put the mask explicitly in the specific cases. To me that also makes it a
bit easier to read.
If you do use the magic numbers above, please add a comment that they are
hand-picked to match {{9 - ((magnitude - 1) / 7)}} for {{magnitude}} between 0
and 63.
> VIntCoding should read and write more efficiently
> -------------------------------------------------
>
> Key: CASSANDRA-15215
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15215
> Project: Cassandra
> Issue Type: Bug
> Components: Local/Compaction, Local/SSTable
> Reporter: Benedict Elliott Smith
> Assignee: Aleksandr Sorokoumov
> Priority: Normal
> Fix For: 3.0.x, 3.11.x, 4.x
>
>
> Most vints occupy significantly fewer than 8 bytes, and most buffers have >=
> 8 bytes spare, in which case we can construct the relevant bytes in a
> register and memcpy them to the correct position. Since we read and write a
> lot of vints, this waste is probably measurable, particularly during
> compaction and flush, and can probably be considered a performance bug.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
