[ 
https://issues.apache.org/jira/browse/CALCITE-6763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17911265#comment-17911265
 ] 

Julian Hyde edited comment on CALCITE-6763 at 1/8/25 11:48 PM:
---------------------------------------------------------------

[~mbudiu], There are two possible strategies.
 # Add a micro-benchmark in 
[ubenchmark|https://github.com/apache/calcite/tree/main/ubenchmark] folder, run 
the benchmark manually before and after the fix, and paste the output into the 
Jira case. CALCITE-4994 and CALCITE-3827 are examples of this.
 # If the performance bug is extreme - e.g. the fix improves running time from 
O(2^N) to O(N log N) - you can add a test case that is very expensive before 
the fix but negligible after the fix. 


was (Author: julianhyde):
[~mbudiu], There are two possible strategies.
 # Add a micro-benchmark in 
[ubenchmark|https://github.com/apache/calcite/tree/main/ubenchmark] folder, run 
the benchmark manually before and after the fix, and paste the output into the 
Jira case. CALCITE-4994 and CALCITE-3827 are examples of this.
 # If the performance bug is extreme, you can add a test case that is very 
expensive before the fix but negligible after the fix. 

> Optimize logic to select the tiles with the fewest rows 
> --------------------------------------------------------
>
>                 Key: CALCITE-6763
>                 URL: https://issues.apache.org/jira/browse/CALCITE-6763
>             Project: Calcite
>          Issue Type: Improvement
>          Components: core
>            Reporter: xiaochen.zhou
>            Priority: Minor
>              Labels: pull-request-available
>         Attachments: image-2025-01-08-12-39-52-295.png, 
> image-2025-01-08-12-40-26-332.png
>
>
> Our current logic using a PriorityQueue to select the tile with the fewest 
> rows when multiple tiles are available. However, there are several potential 
> issues:
> 1. The PriorityQueue stores all satisfiable tiles. When the number of tiles 
> is large, maintaining the heap structure during element insertion has a time 
> complexity of O(log n), which also increases memory usage.
> 2. The initial size of the PriorityQueue is difficult to estimate and is 
> currently set to 1, causing frequent resizing of the PriorityQueue.
> We can optimize the code by keeping only a single bestCandidate tile to 
> improve performance.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to