yugushihuang commented on PR #13572:
URL: https://github.com/apache/lucene/pull/13572#issuecomment-2435780436

   We have measured performance using 
[knnPerfTest.py](https://github.com/mikemccand/luceneutil/blob/main/src/python/knnPerfTest.py)
 in lucene util with this PR 
[commit](https://github.com/goankur/lucene/commit/85d78116f87b679078a80cf606cd4bc7219ee793)
 as candidate branch. 
   ### cmd
   ```
   '/usr/lib/jvm/java-21-amazon-corretto/bin/java', '-cp', [...], 
'--add-modules', 'jdk.incubator.vector', 
'-Djava.library.path=/home/[user_name]/lucene_candidate/lucene/native/build/libs/dotProduct/shared',
 'knn.KnnGraphTester', '-quantize', '-ndoc', '1500000', '-maxConn', '32', 
'-beamWidthIndex', '50', '-fanout', '6', '-quantizeBits', '7', 
'-numMergeWorker', '12', '-numMergeThread', '4', '-encoding', 'float32', 
'-topK', '10', '-dim', '768', '-docs', 'enwiki-20120502-lines-1k-mpnet.vec', 
'-reindex', '-search-and-stats', 'enwiki-20120502-mpnet.vec', '-forceMerge', 
'-quiet'
   ```
   ### Lucene_Baseline
   ```
   Graph level=3 size=46, connectedness=1.00
   Graph level=2 size=1405, connectedness=1.00
   Graph level=1 size=46174, connectedness=1.00
   Graph level=0 size=1500000, connectedness=1.00
   
   Results:
   recall  latency (ms)     nDoc  topK  fanout  maxConn  beamWidth  quantized  
index s  force merge s  num segments  index size (MB)
    0.332         0.333  1500000    10       6       32         50     7 bits   
432.69         271.51             1          5558.90
   ```
   ### Lucene_Candidate
   ```
   Graph level=3 size=46, connectedness=1.00
   Graph level=2 size=1410, connectedness=1.00
   Graph level=1 size=46205, connectedness=1.00
   Graph level=0 size=1500000, connectedness=1.00
   
   Results:
   recall  latency (ms)     nDoc  topK  fanout  maxConn  beamWidth  quantized  
index s  force merge s  num segments  index size (MB)
    0.337         0.260  1500000    10       6       32         50     7 bits   
441.25         293.41             1          5558.91
   ```
   
   The latency has dropped from 0.333ms to 0.26ms.


-- 
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