jimczi commented on PR #12551:
URL: https://github.com/apache/lucene/pull/12551#issuecomment-1720078533

   Adding some charts together to compare how effective it is to use a dynamic 
efSearch.
   
   The first chart shows how well different efSearch values work on one 
segment, on multiple segments with a fixed efSearch, and on multiple segments 
with a dynamic efSearch that changes based on segment size.
   
   The second chart compares how many comparisons are needed for each efSearch 
value, with the single-segment case as the starting point. For example, if you 
set efSearch to 10, you'll need more than 16 times as many comparisons when 
using the same value for all segments compared to when you're dealing with just 
one segment.
   
   ## SIFT (10-200)
   
   
![sift_10_200](https://github.com/apache/lucene/assets/15977469/6a9c14b3-3afa-4b5e-b0b4-bdd48e74bcab)
   
   ## SIFT (100-2000)
   
   
![sift_100_2000](https://github.com/apache/lucene/assets/15977469/01d2f678-17ae-4d04-8201-3127bd71df69)
   
   ## Cohere (10-200)
   
   
![cohere_10_200](https://github.com/apache/lucene/assets/15977469/94acb49b-30df-4f35-8024-0614ba63d177)
   
   ## Cohere (100-2000)
   
![cohere_100_2000](https://github.com/apache/lucene/assets/15977469/232bac74-c3f4-40c4-8628-40e2b26b277b)
   
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org

Reply via email to