diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt
index 08f379c23db6..8ecc7015817c 100644
--- a/lucene/CHANGES.txt
+++ b/lucene/CHANGES.txt
@@ -168,6 +168,8 @@ Optimizations
 
 * GITHUB#15597, GITHUB#15777: Reduce memory usage of NeighborArray (Viliam Durina)
 
+* GITHUB#12324: ELiminate redundant work in AbstractKnnVectorQuery.rewrite (Mike Sokolov)
+
 Bug Fixes
 ---------------------
 * GITHUB#14049: Randomize KNN codec params in RandomCodec. Fixes scalar quantization div-by-zero
diff --git a/lucene/core/src/java/org/apache/lucene/search/AbstractKnnVectorQuery.java b/lucene/core/src/java/org/apache/lucene/search/AbstractKnnVectorQuery.java
index 4bdd821ccba0..b877e81a1997 100644
--- a/lucene/core/src/java/org/apache/lucene/search/AbstractKnnVectorQuery.java
+++ b/lucene/core/src/java/org/apache/lucene/search/AbstractKnnVectorQuery.java
@@ -146,7 +146,9 @@ public Query rewrite(IndexSearcher indexSearcher) throws IOException {
       }
       assert leafReaderContexts.size() == tasks.size();
       assert perLeafResults.size() == reader.leaves().size();
-      topK = runSearchTasks(tasks, taskExecutor, perLeafResults, leafReaderContexts);
+      if (tasks.size() > 0) {
+        topK = runSearchTasks(tasks, taskExecutor, perLeafResults, leafReaderContexts);
+      }
     }
     if (topK.scoreDocs.length == 0) {
       return MatchNoDocsQuery.INSTANCE;