[jira] [Comment Edited] (CASSANDRA-15511) Utilising BTree Improvements

["Benedict Elliott Smith (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-15511?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17559176#comment-17559176 ] 

Benedict Elliott Smith edited comment on CASSANDRA-15511 at 6/27/22 1:44 PM:
-----------------------------------------------------------------------------

Had a quick skim through the patch, and noted a handful of things we might want to address:

# The new {{UpdateFunction.retain}} method doesn't need to be declared in the interface, and can live solely in {{ColumnData.Reconciler}} - this is helpful since we don't use {{retain}} in {{BTree.update}} which might be surprising semantically to users in future. It means we also don't have to rename {{UpdateFunction.insert}} if we don't want to, and can avoid allocating lambdas to invoke it.
# I think in {{ColumnData.Reconciler}} we probably don't need to expunge records from the existing tree if the existing active deletion supersedes the new one.
# I think there has been an accidental removal of an {{updateF}} parameter when copying from the {{insert}} tree, on which we should probably be invoking {{updateF.insert}} still in this branch.

Looks great to me otherwise though, and a more modest patch than I remember. Look forward to it landing.


was (Author: benedict):
Had a quick skim through the patch, and noted a handful of things we might want to address:

# The new {{UpdateFunction.retain}} method doesn't need to be declared in the interface, and can live solely in {{ColumnData.Reconciler}} - this is helpful since we don't use {{retain}} in {{BTree.update}} which might be surprising semantically to users in future. It means we also don't have to rename {{UpdateFunction.insert}} if we don't want to, and can avoid allocating lambdas to invoke it.
# I think in {{ColumnData.Reconciler}} we probably don't need to expunge records from the existing tree if the existing active deletion supersedes the new one.
# I think there has been an accidental removal of an {{updateF}} parameter when copying from the {{insert}} tree, on which we should probably be invoking {{updateF.insert}} still in this branch.

> Utilising BTree Improvements
> ----------------------------
>
>                 Key: CASSANDRA-15511
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-15511
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Local/Other
>            Reporter: Benedict Elliott Smith
>            Assignee: Benedict Elliott Smith
>            Priority: Normal
>             Fix For: 4.0.x, 4.x
>
>         Attachments: atomicbtreepartition.ods, atomicbtreepartition.xlsx.zip, perfsh.tar.gz
>
>
> This patch utilises CASSANDRA-15510 to improve throughput and reduce garbage produced by a number of common operations, by employing {{transformAndFilter}}, {{transform}} and {{FastBuilder}}
> * {{Row}}, {{Cell}} and {{ComplexColumnData}} cloning are implemented with {{BTree.transform}}, so no special builders are necessary; 
> ** {{Rows.copy}} removed
> * {{Rows.merge}} implemented using {{BTree.update}} and a {{ColumnData}} reconciler
> ** Zero-allocations if result of merge is same as {{existing}}
> ** Fewer comparisons
> * {{ColumnData}} reconciler implemented in same manner
> ** {{Cells.reconcileComplex}} is retired
> ** {{ComplexColumnData}} reconciliation now
> *** Garbage-free if the merge has no effect
> *** Always fewer allocations
> *** Fewer comparisons
> * {{FastBuilder}} employed widely:
> ** {{ClusteringIndexNamesFilter}} deserialization
> ** {{Columns}} deserialization
> ** {{PartitionUpdate}} deserialization
> ** {{AbstractBTreePartition}} construction
> ** Misc others
> The upshot of this work when combined with the proposed patch for CASSANDRA-15367 has a dramatic impact on operations over collection types - under contention, as much as 100x improved throughput, and hundreds of megabytes of reduced allocations.  For all operations, allocations under contention and no contention are significantly reduced and throughput improved.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org