Improved code

CHANGELOG.md

@@ -1,3 +1,7 @@
+## 0.6.2 (unreleased)
+
+- Improved performance of parallel HNSW index builds
+
 ## 0.6.1 (2024-03-04)
 
 - Fixed error with `ANALYZE` and vectors with different dimensions

README.md

@@ -581,6 +581,10 @@ or choose to store vectors inline:
 ALTER TABLE items ALTER COLUMN embedding SET STORAGE PLAIN;
 ```
 
+#### Why are there less results for a query after adding an HNSW index?
+
+Results are limited by the size of the dynamic candidate list (`hnsw.ef_search`). There may be even less results due to dead tuples or filtering conditions in the query. We recommend setting `hnsw.ef_search` to at least twice the `LIMIT` of the query. If you need more than 500 results, use an IVFFlat index instead.
+
 #### Why are there less results for a query after adding an IVFFlat index?
 
 The index was likely created with too little data for the number of lists. Drop the index until the table has more data.
@@ -589,6 +593,8 @@ The index was likely created with too little data for the number of lists. Drop
 DROP INDEX index_name;
 ```
 
+Results can also be limited by the number of probes (`ivfflat.probes`).
+
 ## Reference
 
 ### Vector Type

src/hnswbuild.c

@@ -400,7 +400,7 @@ UpdateNeighborsInMemory(char *base, FmgrInfo *procinfo, Oid collation, HnswEleme
  * Update graph in memory
  */
 static void
-UpdateGraphInMemory(FmgrInfo *procinfo, Oid collation, HnswElement element, int m, int efConstruction, HnswElement entryPoint, HnswBuildState * buildstate)
+UpdateGraphInMemory(FmgrInfo *procinfo, Oid collation, HnswElement element, int m, int efConstruction, bool updateEntryPoint, HnswBuildState * buildstate)
 {
 	HnswGraph  *graph = buildstate->graph;
 	char	   *base = buildstate->hnswarea;
@@ -416,7 +416,7 @@ UpdateGraphInMemory(FmgrInfo *procinfo, Oid collation, HnswElement element, int
 	UpdateNeighborsInMemory(base, procinfo, collation, element, m);
 
 	/* Update entry point if needed (already have lock) */
-	if (entryPoint == NULL || element->level > entryPoint->level)
+	if (updateEntryPoint)
 		HnswPtrStore(base, graph->entryPoint, element);
 }
 
@@ -434,32 +434,29 @@ InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
 	int			efConstruction = buildstate->efConstruction;
 	int			m = buildstate->m;
 	char	   *base = buildstate->hnswarea;
+	bool		updateEntryPoint;
 
 	/* Get entry point */
-	LWLockAcquire(entryLock, LW_SHARED);
+	LWLockAcquire(entryLock, LW_EXCLUSIVE);
 	entryPoint = HnswPtrAccess(base, graph->entryPoint);
 
-	/* Prevent concurrent inserts when likely updating entry point */
+	/* Pause new inserts when updating entry point */
-	if (entryPoint == NULL || element->level > entryPoint->level)
+	/* May still be in-flight inserts */
-	{
+	updateEntryPoint = entryPoint == NULL || element->level > entryPoint->level;
-		/* Release shared lock */
+
+	/* Release entry lock */
+	if (!updateEntryPoint)
 		LWLockRelease(entryLock);
 
-		/* Get exclusive lock */
-		LWLockAcquire(entryLock, LW_EXCLUSIVE);
-
-		/* Get latest entry point after lock is acquired */
-		entryPoint = HnswPtrAccess(base, graph->entryPoint);
-	}
-
 	/* Find neighbors for element */
 	HnswFindElementNeighbors(base, element, entryPoint, NULL, procinfo, collation, m, efConstruction, false);
 
 	/* Update graph in memory */
-	UpdateGraphInMemory(procinfo, collation, element, m, efConstruction, entryPoint, buildstate);
+	UpdateGraphInMemory(procinfo, collation, element, m, efConstruction, updateEntryPoint, buildstate);
 
 	/* Release entry lock */
-	LWLockRelease(entryLock);
+	if (updateEntryPoint)
+		LWLockRelease(entryLock);
 }
 
 /*