Version bump to 0.6.0 [skip ci]

These places called PageIndexTupleOverwrite(), with the new tuple
pointing directly to the original page. The PageIndexTupleOverwrite()
call is unnecessary in these cases, as we have already modified the
tuple on the page directly. Moreover, PageIndexTupleOverWrite() will
call memcpy with same src and dst arguments, which is undefined
behavior.

It's OK to modify the pages on disk directly, and we don't need
critical sections, because we either use the generic xlog functions
which create a temporary copy of the page, or we are building a new
index so if we crash the whole index is invisible and will be dropped
anyway.

CHANGELOG.md

@@ -1,11 +1,11 @@
-## 0.6.0 (unreleased)
+## 0.6.0 (2024-01-29)
 
-If upgrading with Postgres 12 or Docker, see [these notes](https://github.com/pgvector/pgvector#060-unreleased).
+If upgrading with Postgres 12 or Docker, see [these notes](https://github.com/pgvector/pgvector#060).
 
 - Changed storage for vector from `extended` to `external`
-- Improved performance of HNSW
 - Added support for parallel index builds for HNSW
 - Added validation for GUC parameters
+- Improved performance of HNSW
 - Reduced memory usage for HNSW index builds
 - Reduced WAL generation for HNSW index builds
 - Fixed error with logical replication

Dockerfile

@@ -1,4 +1,4 @@
-ARG PG_MAJOR=15
+ARG PG_MAJOR=16
 FROM postgres:$PG_MAJOR
 ARG PG_MAJOR
 

META.json

@@ -2,7 +2,7 @@
 	"name": "vector",
 	"abstract": "Open-source vector similarity search for Postgres",
 	"description": "Supports L2 distance, inner product, and cosine distance",
-	"version": "0.5.1",
+	"version": "0.6.0",
 	"maintainer": [
 		"Andrew Kane <andrew@ankane.org>"
 	],
@@ -20,7 +20,7 @@
 		"vector": {
 			"file": "sql/vector.sql",
 			"docfile": "README.md",
-			"version": "0.5.1",
+			"version": "0.6.0",
 			"abstract": "Open-source vector similarity search for Postgres"
 		}
 	},

Makefile

@@ -1,5 +1,5 @@
 EXTENSION = vector
-EXTVERSION = 0.5.1
+EXTVERSION = 0.6.0
 
 MODULE_big = vector
 DATA = $(wildcard sql/*--*.sql)

Makefile.win

@@ -1,5 +1,5 @@
 EXTENSION = vector
-EXTVERSION = 0.5.1
+EXTVERSION = 0.6.0
 
 OBJS = src\hnsw.obj src\hnswbuild.obj src\hnswinsert.obj src\hnswscan.obj src\hnswutils.obj src\hnswvacuum.obj src\ivfbuild.obj src\ivfflat.obj src\ivfinsert.obj src\ivfkmeans.obj src\ivfscan.obj src\ivfutils.obj src\ivfvacuum.obj src\vector.obj
 HEADERS = src\vector.h

README.md

@@ -20,7 +20,7 @@ Compile and install the extension (supports Postgres 11+)
 
 ```sh
 cd /tmp
-git clone --branch v0.5.1 https://github.com/pgvector/pgvector.git
+git clone --branch v0.6.0 https://github.com/pgvector/pgvector.git
 cd pgvector
 make
 make install # may need sudo
@@ -44,7 +44,7 @@ Then use `nmake` to build:
 
 ```cmd
 set "PGROOT=C:\Program Files\PostgreSQL\16"
-git clone --branch v0.5.1 https://github.com/pgvector/pgvector.git
+git clone --branch v0.6.0 https://github.com/pgvector/pgvector.git
 cd pgvector
 nmake /F Makefile.win
 nmake /F Makefile.win install
@@ -264,7 +264,7 @@ HINT:  Increase maintenance_work_mem to speed up builds.
 
 Note: Do not set `maintenance_work_mem` so high that it exhausts the memory on the server
 
-Starting with 0.6.0 [unreleased], you can also speed up index creation by increasing the number of parallel workers (2 by default)
+Starting with 0.6.0, you can also speed up index creation by increasing the number of parallel workers (2 by default)
 
 ```sql
 SET max_parallel_maintenance_workers = 7; -- plus leader
@@ -651,10 +651,10 @@ If compilation fails and the output includes `warning: no such sysroot directory
 
 ### Docker
 
-Get the [Docker image](https://hub.docker.com/r/ankane/pgvector) with:
+Get the [Docker image](https://hub.docker.com/r/pgvector/pgvector) with:
 
 ```sh
-docker pull ankane/pgvector
+docker pull pgvector/pgvector:pg16
 ```
 
 This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (run it the same way).
@@ -662,9 +662,9 @@ This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (r
 You can also build the image manually:
 
 ```sh
-git clone --branch v0.5.1 https://github.com/pgvector/pgvector.git
+git clone --branch v0.6.0 https://github.com/pgvector/pgvector.git
 cd pgvector
-docker build --build-arg PG_MAJOR=15 -t myuser/pgvector .
+docker build --build-arg PG_MAJOR=16 -t myuser/pgvector .
 ```
 
 ### Homebrew
@@ -741,7 +741,7 @@ SELECT extversion FROM pg_extension WHERE extname = 'vector';
 
 ## Upgrade Notes
 
-### 0.6.0 [unreleased]
+### 0.6.0
 
 #### Postgres 12
 

src/hnsw.h

@@ -60,6 +60,7 @@
 #define PROGRESS_HNSW_PHASE_LOAD		2
 
 #define HNSW_MAX_SIZE (BLCKSZ - MAXALIGN(SizeOfPageHeaderData) - MAXALIGN(sizeof(HnswPageOpaqueData)) - sizeof(ItemIdData))
+#define HNSW_TUPLE_ALLOC_SIZE BLCKSZ
 
 #define HNSW_ELEMENT_TUPLE_SIZE(size)	MAXALIGN(offsetof(HnswElementTupleData, data) + (size))
 #define HNSW_NEIGHBOR_TUPLE_SIZE(level, m)	MAXALIGN(offsetof(HnswNeighborTupleData, indextids) + ((level) + 2) * (m) * sizeof(ItemPointerData))
@@ -354,6 +355,7 @@ typedef struct HnswVacuumState
 	/* Variables */
 	struct tidhash_hash *deleted;
 	BufferAccessStrategy bas;
+	HnswNeighborTuple ntup;
 	HnswElementData highestPoint;
 
 	/* Memory */

src/hnswbuild.c

@@ -148,7 +148,6 @@ CreateGraphPages(HnswBuildState * buildstate)
 {
 	Relation	index = buildstate->index;
 	ForkNumber	forkNum = buildstate->forkNum;
-	Size		etupAllocSize;
 	Size		maxSize;
 	HnswElementTuple etup;
 	HnswNeighborTuple ntup;
@@ -160,12 +159,11 @@ CreateGraphPages(HnswBuildState * buildstate)
 	char	   *base = buildstate->hnswarea;
 
 	/* Calculate sizes */
-	etupAllocSize = BLCKSZ;
 	maxSize = HNSW_MAX_SIZE;
 
 	/* Allocate once */
-	etup = palloc0(etupAllocSize);
-	ntup = palloc0(BLCKSZ);
+	etup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
+	ntup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
 
 	/* Prepare first page */
 	buf = HnswNewBuffer(index, forkNum);
@@ -184,7 +182,7 @@ CreateGraphPages(HnswBuildState * buildstate)
 		iter = element->next;
 
 		/* Zero memory for each element */
-		MemSet(etup, 0, etupAllocSize);
+		MemSet(etup, 0, HNSW_TUPLE_ALLOC_SIZE);
 
 		/* Calculate sizes */
 		etupSize = HNSW_ELEMENT_TUPLE_SIZE(VARSIZE_ANY(valuePtr));
@@ -192,7 +190,7 @@ CreateGraphPages(HnswBuildState * buildstate)
 		combinedSize = etupSize + ntupSize + sizeof(ItemIdData);
 
 		/* Initial size check */
-		if (etupSize > etupAllocSize)
+		if (etupSize > HNSW_TUPLE_ALLOC_SIZE)
 			elog(ERROR, "index tuple too large");
 
 		HnswSetElementTuple(base, etup, element);
@@ -254,38 +252,43 @@ WriteNeighborTuples(HnswBuildState * buildstate)
 	int			m = buildstate->m;
 	HnswElementPtr iter = buildstate->graph->head;
 	char	   *base = buildstate->hnswarea;
+	HnswNeighborTuple ntup;
+
+	/* Allocate once */
+	ntup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
 
 	while (!HnswPtrIsNull(base, iter))
 	{
-		HnswElement e = HnswPtrAccess(base, iter);
+		HnswElement element = HnswPtrAccess(base, iter);
 		Buffer		buf;
 		Page		page;
-		ItemId		itemid;
-		HnswNeighborTuple ntup;
+		Size		ntupSize = HNSW_NEIGHBOR_TUPLE_SIZE(element->level, m);
 
 		/* Update iterator */
-		iter = e->next;
+		iter = element->next;
+
+		/* Zero memory for each element */
+		MemSet(ntup, 0, HNSW_TUPLE_ALLOC_SIZE);
 
 		/* Can take a while, so ensure we can interrupt */
 		/* Needs to be called when no buffer locks are held */
 		CHECK_FOR_INTERRUPTS();
 
-		buf = ReadBufferExtended(index, forkNum, e->neighborPage, RBM_NORMAL, NULL);
+		buf = ReadBufferExtended(index, forkNum, element->neighborPage, RBM_NORMAL, NULL);
 		LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
 		page = BufferGetPage(buf);
-		itemid = PageGetItemId(page, e->neighborOffno);
-		ntup = (HnswNeighborTuple) PageGetItem(page, itemid);
 
-		/* Check expected size */
-		Assert(ItemIdGetLength(itemid) == HNSW_NEIGHBOR_TUPLE_SIZE(e->level, m));
+		HnswSetNeighborTuple(base, ntup, element, m);
 
-		/* Update page in-place */
-		HnswSetNeighborTuple(base, ntup, e, m);
+		if (!PageIndexTupleOverwrite(page, element->neighborOffno, (Item) ntup, ntupSize))
+			elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
 
 		/* Commit */
 		MarkBufferDirty(buf);
 		UnlockReleaseBuffer(buf);
 	}
+
+	pfree(ntup);
 }
 
 /*

src/hnswinsert.c

@@ -355,9 +355,7 @@ HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, Hns
 			Buffer		buf;
 			Page		page;
 			GenericXLogState *state;
-			ItemId		itemid;
 			HnswNeighborTuple ntup;
-			Size		ntupSize;
 			int			idx = -1;
 			int			startIdx;
 			HnswElement neighborElement = HnswPtrAccess(base, hc->element);
@@ -396,9 +394,7 @@ HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, Hns
 			}
 
 			/* Get tuple */
-			itemid = PageGetItemId(page, offno);
-			ntup = (HnswNeighborTuple) PageGetItem(page, itemid);
-			ntupSize = ItemIdGetLength(itemid);
+			ntup = (HnswNeighborTuple) PageGetItem(page, PageGetItemId(page, offno));
 
 			/* Calculate index for update */
 			startIdx = (neighborElement->level - lc) * m;
@@ -427,13 +423,9 @@ HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, Hns
 			{
 				ItemPointer indextid = &ntup->indextids[idx];
 
-				/* Update neighbor */
+				/* Update neighbor on the buffer */
 				ItemPointerSet(indextid, e->blkno, e->offno);
 
-				/* Overwrite tuple */
-				if (!PageIndexTupleOverwrite(page, offno, (Item) ntup, ntupSize))
-					elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
-
 				/* Commit */
 				if (building)
 					MarkBufferDirty(buf);
@@ -457,9 +449,7 @@ AddDuplicateOnDisk(Relation index, HnswElement element, HnswElement dup, bool bu
 	Buffer		buf;
 	Page		page;
 	GenericXLogState *state;
-	ItemId		itemid;
 	HnswElementTuple etup;
-	Size		etupSize;
 	int			i;
 
 	/* Read page */
@@ -477,9 +467,7 @@ AddDuplicateOnDisk(Relation index, HnswElement element, HnswElement dup, bool bu
 	}
 
 	/* Find space */
-	itemid = PageGetItemId(page, dup->offno);
-	etup = (HnswElementTuple) PageGetItem(page, itemid);
-	etupSize = ItemIdGetLength(itemid);
+	etup = (HnswElementTuple) PageGetItem(page, PageGetItemId(page, dup->offno));
 	for (i = 0; i < HNSW_HEAPTIDS; i++)
 	{
 		if (!ItemPointerIsValid(&etup->heaptids[i]))
@@ -495,13 +483,9 @@ AddDuplicateOnDisk(Relation index, HnswElement element, HnswElement dup, bool bu
 		return false;
 	}
 
-	/* Add heap TID */
+	/* Add heap TID, modifying the tuple on the page directly */
 	etup->heaptids[i] = element->heaptids[0];
 
-	/* Overwrite tuple */
-	if (!PageIndexTupleOverwrite(page, dup->offno, (Item) etup, etupSize))
-		elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
-
 	/* Commit */
 	if (building)
 		MarkBufferDirty(buf);

src/hnswvacuum.c

@@ -60,8 +60,7 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 		/* Iterate over nodes */
 		for (offno = FirstOffsetNumber; offno <= maxoffno; offno = OffsetNumberNext(offno))
 		{
-			ItemId		itemid = PageGetItemId(page, offno);
-			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, itemid);
+			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, PageGetItemId(page, offno));
 			int			idx = 0;
 			bool		itemUpdated = false;
 
@@ -92,15 +91,10 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 
 				if (itemUpdated)
 				{
-					Size		etupSize = ItemIdGetLength(itemid);
-
 					/* Mark rest as invalid */
 					for (int i = idx; i < HNSW_HEAPTIDS; i++)
 						ItemPointerSetInvalid(&etup->heaptids[i]);
 
-					if (!PageIndexTupleOverwrite(page, offno, (Item) etup, etupSize))
-						elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
-
 					updated = true;
 				}
 			}
@@ -198,8 +192,8 @@ RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswEleme
 	FmgrInfo   *procinfo = vacuumstate->procinfo;
 	Oid			collation = vacuumstate->collation;
 	BufferAccessStrategy bas = vacuumstate->bas;
-	ItemId		itemid;
-	HnswNeighborTuple ntup;
+	HnswNeighborTuple ntup = vacuumstate->ntup;
+	Size		ntupSize = HNSW_NEIGHBOR_TUPLE_SIZE(element->level, m);
 	char	   *base = NULL;
 
 	/* Skip if element is entry point */
@@ -213,23 +207,23 @@ RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswEleme
 	/* Find neighbors for element, skipping itself */
 	HnswFindElementNeighbors(base, element, entryPoint, index, procinfo, collation, m, efConstruction, true);
 
+	/* Zero memory for each element */
+	MemSet(ntup, 0, HNSW_TUPLE_ALLOC_SIZE);
+
+	/* Update neighbor tuple */
+	/* Do this before getting page to minimize locking */
+	HnswSetNeighborTuple(base, ntup, element, m);
+
 	/* Get neighbor page */
 	buf = ReadBufferExtended(index, MAIN_FORKNUM, element->neighborPage, RBM_NORMAL, bas);
 	LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
 	state = GenericXLogStart(index);
 	page = GenericXLogRegisterBuffer(state, buf, 0);
 
-	/* Get neighbor tuple */
-	itemid = PageGetItemId(page, element->neighborOffno);
-	ntup = (HnswNeighborTuple) PageGetItem(page, itemid);
-
-	/* Check expected size */
-	if (ItemIdGetLength(itemid) != HNSW_NEIGHBOR_TUPLE_SIZE(element->level, m))
+	/* Overwrite tuple */
+	if (!PageIndexTupleOverwrite(page, element->neighborOffno, (Item) ntup, ntupSize))
 		elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
 
-	/* Update page in-place */
-	HnswSetNeighborTuple(base, ntup, element, m);
-
 	/* Commit */
 	GenericXLogFinish(state);
 	UnlockReleaseBuffer(buf);
@@ -482,11 +476,8 @@ MarkDeleted(HnswVacuumState * vacuumstate)
 		/* Update element and neighbors together */
 		for (offno = FirstOffsetNumber; offno <= maxoffno; offno = OffsetNumberNext(offno))
 		{
-			ItemId		itemid = PageGetItemId(page, offno);
-			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, itemid);
+			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, PageGetItemId(page, offno));
 			HnswNeighborTuple ntup;
-			Size		etupSize;
-			Size		ntupSize;
 			Buffer		nbuf;
 			Page		npage;
 			BlockNumber neighborPage;
@@ -510,10 +501,6 @@ MarkDeleted(HnswVacuumState * vacuumstate)
 			if (ItemPointerIsValid(&etup->heaptids[0]))
 				continue;
 
-			/* Calculate sizes */
-			etupSize = ItemIdGetLength(itemid);
-			ntupSize = HNSW_NEIGHBOR_TUPLE_SIZE(etup->level, vacuumstate->m);
-
 			/* Get neighbor page */
 			neighborPage = ItemPointerGetBlockNumber(&etup->neighbortid);
 			neighborOffno = ItemPointerGetOffsetNumber(&etup->neighbortid);
@@ -540,13 +527,10 @@ MarkDeleted(HnswVacuumState * vacuumstate)
 			for (int i = 0; i < ntup->count; i++)
 				ItemPointerSetInvalid(&ntup->indextids[i]);
 
-			/* Overwrite element tuple */
-			if (!PageIndexTupleOverwrite(page, offno, (Item) etup, etupSize))
-				elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
-
-			/* Overwrite neighbor tuple */
-			if (!PageIndexTupleOverwrite(npage, neighborOffno, (Item) ntup, ntupSize))
-				elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
+			/*
+			 * We modified the tuples in place, no need to call
+			 * PageIndexTupleOverwrite
+			 */
 
 			/* Commit */
 			GenericXLogFinish(state);
@@ -591,6 +575,7 @@ InitVacuumState(HnswVacuumState * vacuumstate, IndexVacuumInfo *info, IndexBulkD
 	vacuumstate->bas = GetAccessStrategy(BAS_BULKREAD);
 	vacuumstate->procinfo = index_getprocinfo(index, 1, HNSW_DISTANCE_PROC);
 	vacuumstate->collation = index->rd_indcollation[0];
+	vacuumstate->ntup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
 	vacuumstate->tmpCtx = AllocSetContextCreate(CurrentMemoryContext,
 												"Hnsw vacuum temporary context",
 												ALLOCSET_DEFAULT_SIZES);
@@ -610,6 +595,7 @@ FreeVacuumState(HnswVacuumState * vacuumstate)
 {
 	tidhash_destroy(vacuumstate->deleted);
 	FreeAccessStrategy(vacuumstate->bas);
+	pfree(vacuumstate->ntup);
 	MemoryContextDelete(vacuumstate->tmpCtx);
 }
 

vector.control

@@ -1,4 +1,4 @@
 comment = 'vector data type and ivfflat and hnsw access methods'
-default_version = '0.5.1'
+default_version = '0.6.0'
 module_pathname = '$libdir/vector'
 relocatable = true