Updated SparsevecInnerProduct [skip ci]

.github/workflows/build.yml

@@ -97,7 +97,7 @@ jobs:
         with:
           postgres-version: ${{ matrix.postgres }}
       - run: |
-          call "C:\Program Files\Microsoft Visual Studio\${{ matrix.os == 'windows-2025' && 18 || 2022 }}\Enterprise\VC\Auxiliary\Build\vcvars64.bat" && ^
+          call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvars64.bat" && ^
           nmake /NOLOGO /F Makefile.win && ^
           nmake /NOLOGO /F Makefile.win install && ^
           nmake /NOLOGO /F Makefile.win installcheck ${{ matrix.postgres != 17 && 'PG_REGRESS=$(PGROOT)\bin\pg_regress' || '' }} && ^

CHANGELOG.md

@@ -1,8 +1,3 @@
-## 0.8.3 (unreleased)
-
-- Fixed possible index corruption with HNSW vacuuming
-- Fixed performance regression with Hamming distance and Jaccard distance with Postgres 18
-
 ## 0.8.2 (2026-02-25)
 
 - Fixed buffer overflow with parallel HNSW index build - [more info](https://github.com/pgvector/pgvector/issues/959)

README.md

@@ -11,8 +11,6 @@ Store your vectors with the rest of your data. Supports:
 
 Plus [ACID](https://en.wikipedia.org/wiki/ACID) compliance, point-in-time recovery, JOINs, and all of the other [great features](https://www.postgresql.org/about/) of Postgres
 
-Have a lot of vectors? Use [quantization](#scaling) to scale
-
 [![Build Status](https://github.com/pgvector/pgvector/actions/workflows/build.yml/badge.svg)](https://github.com/pgvector/pgvector/actions)
 
 ## Installation
@@ -316,8 +314,6 @@ For a large number of workers, you may need to increase `max_parallel_workers` (
 
 The [index options](#index-options) also have a significant impact on build time (use the defaults unless seeing low recall)
 
-Use [binary quantization](#binary-quantization) for faster build times at scale
-
 ### Indexing Progress
 
 Check [indexing progress](https://www.postgresql.org/docs/current/progress-reporting.html#CREATE-INDEX-PROGRESS-REPORTING)
@@ -447,7 +443,13 @@ Exact indexes work well for conditions that match a low percentage of rows. Othe
 CREATE INDEX ON items USING hnsw (embedding vector_l2_ops);
 ```
 
-With approximate indexes, filtering is applied *after* the index is scanned. If a condition matches 10% of rows, with HNSW and the default `hnsw.ef_search` of 40, only 4 rows will match on average. For more rows, enable [iterative index scans](#iterative-index-scans), which will automatically scan more of the index when needed.
+With approximate indexes, filtering is applied *after* the index is scanned. If a condition matches 10% of rows, with HNSW and the default `hnsw.ef_search` of 40, only 4 rows will match on average. For more rows, increase `hnsw.ef_search`.
+
+```sql
+SET hnsw.ef_search = 200;
+```
+
+Starting with 0.8.0, you can enable [iterative index scans](#iterative-index-scans), which will automatically scan more of the index when needed.
 
 ```sql
 SET hnsw.iterative_scan = strict_order;
@@ -671,10 +673,6 @@ SHOW shared_buffers;
 
 Be sure to restart Postgres for changes to take effect.
 
-### Storing
-
-Use the `halfvec` type instead of `vector` for a smaller working set.
-
 ### Loading
 
 Use `COPY` for bulk loading data ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/loading/example.py)).
@@ -689,8 +687,6 @@ Add any indexes *after* loading the initial data for best performance.
 
 See index build time for [HNSW](#index-build-time) and [IVFFlat](#index-build-time-1).
 
-Use [binary quantization](#binary-quantization) for smaller indexes and faster build times at scale.
-
 In production environments, create indexes concurrently to avoid blocking writes.
 
 ```sql
@@ -721,8 +717,6 @@ SELECT * FROM items ORDER BY embedding <#> '[3,1,2]' LIMIT 5;
 
 #### Approximate Search
 
-Use [binary quantization](#binary-quantization) with re-ranking to keep indexes in-memory at scale.
-
 To speed up queries with an IVFFlat index, increase the number of inverted lists (at the expense of recall).
 
 ```sql
@@ -738,20 +732,21 @@ REINDEX INDEX CONCURRENTLY index_name;
 VACUUM table_name;
 ```
 
-## Scaling
-
-For a smaller working set:
-
-1. Use the `halfvec` type instead of `vector` for tables
-2. Use [binary quantization](#binary-quantization) for indexes (with re-ranking for search)
-
-Scale vertically by increasing memory, CPU, and storage on a single instance. Use existing tools to [tune parameters](#tuning) and [monitor performance](#monitoring).
-
-Scale horizontally with [replicas](https://www.postgresql.org/docs/current/hot-standby.html), or use [Citus](https://github.com/citusdata/citus), [PgDog](https://github.com/pgdogdev/pgdog), or another approach for sharding ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/citus/example.py)).
-
 ## Monitoring
 
-Use existing tools like [pg_stat_statements](https://www.postgresql.org/docs/current/pgstatstatements.html) or [PgHero](https://github.com/ankane/pghero) to monitor performance.
+Monitor performance with [pg_stat_statements](https://www.postgresql.org/docs/current/pgstatstatements.html) (be sure to add it to `shared_preload_libraries`).
+
+```sql
+CREATE EXTENSION pg_stat_statements;
+```
+
+Get the most time-consuming queries with:
+
+```sql
+SELECT query, calls, ROUND((total_plan_time + total_exec_time) / calls) AS avg_time_ms,
+    ROUND((total_plan_time + total_exec_time) / 60000) AS total_time_min
+    FROM pg_stat_statements ORDER BY total_plan_time + total_exec_time DESC LIMIT 20;
+```
 
 Monitor recall by comparing results from approximate search with exact search.
 
@@ -762,6 +757,14 @@ SELECT ...
 COMMIT;
 ```
 
+## Scaling
+
+Scale pgvector the same way you scale Postgres.
+
+Scale vertically by increasing memory, CPU, and storage on a single instance. Use existing tools to [tune parameters](#tuning) and [monitor performance](#monitoring).
+
+Scale horizontally with [replicas](https://www.postgresql.org/docs/current/hot-standby.html), or use [Citus](https://github.com/citusdata/citus) or another approach for sharding ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/citus/example.py)).
+
 ## Languages
 
 Use pgvector from any language with a Postgres client. You can even generate and store vectors in one language and query them in another.
@@ -875,8 +878,6 @@ No, but like other index types, you’ll likely see better performance if they d
 SELECT pg_size_pretty(pg_relation_size('index_name'));
 ```
 
-Use [half-precision indexing](#half-precision-indexing) or [binary quantization](#binary-quantization) for smaller indexes.
-
 ## Troubleshooting
 
 #### Why isn’t a query using an index?
@@ -1329,7 +1330,7 @@ make clean && PG_CFLAGS="-DUSE_ASSERT_CHECKING" make && make install
 To enable benchmarking:
 
 ```sh
-make clean && PG_CFLAGS="-DHNSW_BENCH -DIVFFLAT_BENCH" make && make install
+make clean && PG_CFLAGS="-DIVFFLAT_BENCH" make && make install
 ```
 
 To show memory usage:

src/bitutils.c

@@ -31,12 +31,10 @@
 #define BIT_TARGET_CLONES
 #endif
 
-/* Use built-ins when possible for Postgres < 19 for inlining */
+/* Use built-ins when possible for inlining */
-#if PG_VERSION_NUM >= 190000
+#if defined(HAVE__BUILTIN_POPCOUNT) && defined(HAVE_LONG_INT_64)
-#define popcount64(x) pg_popcount64(x)
-#elif defined(HAVE__BUILTIN_POPCOUNT) && (defined(HAVE_LONG_INT_64) || SIZEOF_LONG == 8)
 #define popcount64(x) __builtin_popcountl(x)
-#elif defined(HAVE__BUILTIN_POPCOUNT) && (defined(HAVE_LONG_LONG_INT_64) || SIZEOF_LONG_LONG == 8)
+#elif defined(HAVE__BUILTIN_POPCOUNT) && defined(HAVE_LONG_LONG_INT_64)
 #define popcount64(x) __builtin_popcountll(x)
 #elif !defined(_MSC_VER)
 /* Fails to resolve with MSVC */
@@ -171,7 +169,7 @@ BitJaccardDistanceAvx512Popcount(uint32 bytes, unsigned char *ax, unsigned char
 #endif
 
 TARGET_XSAVE static bool
-SupportsAvx512Popcount(void)
+SupportsAvx512Popcount()
 {
 	unsigned int exx[4] = {0, 0, 0, 0};
 

src/hnsw.c

@@ -13,7 +13,6 @@
 #include "hnsw.h"
 #include "miscadmin.h"
 #include "nodes/pg_list.h"
-#include "storage/lwlock.h"
 #include "utils/float.h"
 #include "utils/guc.h"
 #include "utils/relcache.h"
@@ -290,11 +289,7 @@ hnswhandler(PG_FUNCTION_ARGS)
 		.amcanparallel = false,
 		.amcanbuildparallel = true,
 		.amcaninclude = false,
-#if PG_VERSION_NUM >= 170000
-		.amusemaintenanceworkmem = true,
-#else
 		.amusemaintenanceworkmem = false,
-#endif
 		.amsummarizing = false,
 		.amparallelvacuumoptions = VACUUM_OPTION_PARALLEL_BULKDEL,
 		.amkeytype = InvalidOid,
@@ -355,11 +350,7 @@ hnswhandler(PG_FUNCTION_ARGS)
 	amroutine->amcanbuildparallel = true;
 #endif
 	amroutine->amcaninclude = false;
-#if PG_VERSION_NUM >= 170000
+	amroutine->amusemaintenanceworkmem = false; /* not used during VACUUM */
-	amroutine->amusemaintenanceworkmem = true;
-#else
-	amroutine->amusemaintenanceworkmem = false;
-#endif
 #if PG_VERSION_NUM >= 160000
 	amroutine->amsummarizing = false;
 #endif

src/hnsw.h

@@ -10,28 +10,14 @@
 #include "lib/pairingheap.h"
 #include "nodes/execnodes.h"
 #include "port.h"				/* for random() */
-#include "storage/bufpage.h"
-#include "storage/condition_variable.h"
-#include "storage/lwlock.h"
-#include "storage/s_lock.h"
 #include "utils/relptr.h"
 #include "utils/sampling.h"
 #include "vector.h"
 
-#ifdef HNSW_BENCH
-#include "portability/instr_time.h"
-#endif
-
 #if PG_VERSION_NUM >= 190000
 typedef Pointer Item;
 #endif
 
-#if PG_VERSION_NUM >= 170000
-#define HnswTidStore TidStore
-#else
-#define HnswTidStore tidhash_hash
-#endif
-
 #define HNSW_MAX_DIM 2000
 #define HNSW_MAX_NNZ 1000
 
@@ -88,21 +74,6 @@ typedef Pointer Item;
 #define HnswPageGetOpaque(page)	((HnswPageOpaque) PageGetSpecialPointer(page))
 #define HnswPageGetMeta(page)	((HnswMetaPageData *) PageGetContents(page))
 
-#ifdef HNSW_BENCH
-#define HnswBench(name, code) \
-	do { \
-		instr_time	start; \
-		instr_time	duration; \
-		INSTR_TIME_SET_CURRENT(start); \
-		(code); \
-		INSTR_TIME_SET_CURRENT(duration); \
-		INSTR_TIME_SUBTRACT(duration, start); \
-		elog(INFO, "%s: %.3f ms", name, INSTR_TIME_GET_MILLISEC(duration)); \
-	} while (0)
-#else
-#define HnswBench(name, code) (code)
-#endif
-
 #if PG_VERSION_NUM >= 150000
 #define RandomDouble() pg_prng_double(&pg_global_prng_state)
 #define SeedRandom(seed) pg_prng_seed(&pg_global_prng_state, seed)
@@ -433,11 +404,10 @@ typedef struct HnswVacuumState
 	HnswSupport support;
 
 	/* Variables */
-	struct HnswTidStore *deleted;
+	struct tidhash_hash *deleted;
 	BufferAccessStrategy bas;
 	HnswNeighborTuple ntup;
 	HnswElementData highestPoint;
-	HnswElementData fallbackPoint;
 
 	/* Memory */
 	MemoryContext tmpCtx;

src/hnswbuild.c

@@ -54,7 +54,6 @@
 #include "nodes/execnodes.h"
 #include "optimizer/optimizer.h"
 #include "storage/bufmgr.h"
-#include "storage/condition_variable.h"
 #include "tcop/tcopprot.h"
 #include "utils/datum.h"
 #include "utils/memutils.h"
@@ -803,11 +802,7 @@ HnswParallelScanAndInsert(Relation heapRel, Relation indexRel, HnswShared * hnsw
 	buildstate.hnswarea = hnswarea;
 	InitAllocator(&buildstate.allocator, &HnswSharedMemoryAlloc, &buildstate);
 	scan = table_beginscan_parallel(heapRel,
-									ParallelTableScanFromHnswShared(hnswshared)
+									ParallelTableScanFromHnswShared(hnswshared));
-#if PG_VERSION_NUM >= 190000
-									,SO_NONE
-#endif
-		);
 	reltuples = table_index_build_scan(heapRel, indexRel, indexInfo,
 									   true, progress, BuildCallback,
 									   (void *) &buildstate, scan);

src/hnswinsert.c

@@ -6,7 +6,6 @@
 #include "nodes/execnodes.h"
 #include "storage/bufmgr.h"
 #include "storage/lmgr.h"
-#include "storage/lwlock.h"
 #include "utils/datum.h"
 #include "utils/memutils.h"
 #include "utils/rel.h"

src/hnswutils.c

@@ -546,9 +546,6 @@ HnswLoadElementImpl(BlockNumber blkno, OffsetNumber offno, double *distance, Hns
 
 	Assert(HnswIsElementTuple(etup));
 
-	if (unlikely(etup->deleted))
-		elog(ERROR, "cannot load deleted element");
-
 	/* Calculate distance */
 	if (distance != NULL)
 	{

src/hnswvacuum.c

@@ -10,12 +10,6 @@
 #include "utils/memutils.h"
 #include "utils/rel.h"
 
-#if PG_VERSION_NUM >= 170000
-#include "access/tidstore.h"
-#include "miscadmin.h"
-#include "postmaster/autovacuum.h"
-#endif
-
 #if PG_VERSION_NUM >= 160000
 #include "varatt.h"
 #endif
@@ -28,13 +22,9 @@
  * Check if deleted list contains an index TID
  */
 static bool
-DeletedContains(HnswTidStore * deleted, ItemPointer indextid)
+DeletedContains(tidhash_hash * deleted, ItemPointer indextid)
 {
-#if PG_VERSION_NUM >= 170000
-	return TidStoreIsMember(deleted, indextid);
-#else
 	return tidhash_lookup(deleted, *indextid) != NULL;
-#endif
 }
 
 /*
@@ -47,20 +37,17 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 {
 	BlockNumber blkno = HNSW_HEAD_BLKNO;
 	HnswElement highestPoint = &vacuumstate->highestPoint;
-	HnswElement fallbackPoint = &vacuumstate->fallbackPoint;
 	Relation	index = vacuumstate->index;
 	BufferAccessStrategy bas = vacuumstate->bas;
+	HnswElement entryPoint = HnswGetEntryPoint(vacuumstate->index);
 	IndexBulkDeleteResult *stats = vacuumstate->stats;
 
-	/* Store separately since HnswElement level is uint8 */
+	/* Store separately since highestPoint.level is uint8 */
 	int			highestLevel = -1;
-	int			fallbackLevel = -1;
 
-	/* Initialize highest point and fallback point */
+	/* Initialize highest point */
 	highestPoint->blkno = InvalidBlockNumber;
 	highestPoint->offno = InvalidOffsetNumber;
-	fallbackPoint->blkno = InvalidBlockNumber;
-	fallbackPoint->offno = InvalidOffsetNumber;
 
 	while (BlockNumberIsValid(blkno))
 	{
@@ -70,10 +57,6 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 		OffsetNumber offno;
 		OffsetNumber maxoffno;
 		bool		updated = false;
-#if PG_VERSION_NUM >= 170000
-		OffsetNumber deletedoffs[MaxOffsetNumber];
-		int			ndeletedoffs = 0;
-#endif
 
 		vacuum_delay_point();
 
@@ -127,9 +110,6 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 
 			if (!ItemPointerIsValid(&etup->heaptids[0]))
 			{
-#if PG_VERSION_NUM >= 170000
-				deletedoffs[ndeletedoffs++] = offno;
-#else
 				ItemPointerData ip;
 				bool		found;
 
@@ -138,39 +118,17 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 
 				tidhash_insert(vacuumstate->deleted, ip, &found);
 				Assert(!found);
-#endif
 			}
-			else if (etup->level > highestLevel)
+			else if (etup->level > highestLevel && !(entryPoint != NULL && blkno == entryPoint->blkno && offno == entryPoint->offno))
 			{
-				if (BlockNumberIsValid(highestPoint->blkno))
+				/* Keep track of highest non-entry point */
-				{
-					/* Current highest point becomes fallback */
-					fallbackPoint->blkno = highestPoint->blkno;
-					fallbackPoint->offno = highestPoint->offno;
-					fallbackPoint->level = highestPoint->level;
-					fallbackLevel = highestLevel;
-				}
-
-				/* Keep track of highest point */
 				highestPoint->blkno = blkno;
 				highestPoint->offno = offno;
 				highestPoint->level = etup->level;
 				highestLevel = etup->level;
 			}
-			else if (etup->level > fallbackLevel)
-			{
-				/* Keep track of second highest point */
-				fallbackPoint->blkno = blkno;
-				fallbackPoint->offno = offno;
-				fallbackPoint->level = etup->level;
-				fallbackLevel = etup->level;
-			}
 		}
 
-#if PG_VERSION_NUM >= 170000
-		TidStoreSetBlockOffsets(vacuumstate->deleted, blkno, deletedoffs, ndeletedoffs);
-#endif
-
 		blkno = HnswPageGetOpaque(page)->nextblkno;
 
 		if (updated)
@@ -180,10 +138,6 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 
 		UnlockReleaseBuffer(buf);
 	}
-
-#ifdef HNSW_MEMORY
-	elog(INFO, "memory: %zu KB", MemoryContextMemAllocated(CurrentMemoryContext, true) / 1024);
-#endif
 }
 
 /*
@@ -315,27 +269,12 @@ RepairGraphEntryPoint(HnswVacuumState * vacuumstate)
 		/* Get a shared lock */
 		LockPage(index, HNSW_UPDATE_LOCK, ShareLock);
 
-		/* Get latest entry point */
+		/* Load element */
-		entryPoint = HnswGetEntryPoint(index);
+		HnswLoadElement(highestPoint, NULL, NULL, index, support, true, NULL);
 
-		/* Use fallback point if highest point is entry point */
+		/* Repair if needed */
-		if (entryPoint != NULL && entryPoint->blkno == highestPoint->blkno && entryPoint->offno == highestPoint->offno)
+		if (NeedsUpdated(vacuumstate, highestPoint))
-		{
+			RepairGraphElement(vacuumstate, highestPoint, HnswGetEntryPoint(index));
-			highestPoint = &vacuumstate->fallbackPoint;
-
-			if (!BlockNumberIsValid(highestPoint->blkno))
-				highestPoint = NULL;
-		}
-
-		if (highestPoint != NULL)
-		{
-			/* Load element */
-			HnswLoadElement(highestPoint, NULL, NULL, index, support, true, NULL);
-
-			/* Repair if needed */
-			if (NeedsUpdated(vacuumstate, highestPoint))
-				RepairGraphElement(vacuumstate, highestPoint, entryPoint);
-		}
 
 		/* Release lock */
 		UnlockPage(index, HNSW_UPDATE_LOCK, ShareLock);
@@ -502,99 +441,6 @@ RepairGraph(HnswVacuumState * vacuumstate)
 		/* Reset memory context */
 		MemoryContextSwitchTo(oldCtx);
 		MemoryContextReset(vacuumstate->tmpCtx);
-
-#ifdef HNSW_VACUUM_PROGRESS
-		if (!BlockNumberIsValid(blkno) || ((blkno - HNSW_HEAD_BLKNO) % 1000 == 0 && blkno != HNSW_HEAD_BLKNO))
-		{
-			BlockNumber totalBlocks = RelationGetNumberOfBlocks(index);
-			BlockNumber currentBlocks = BlockNumberIsValid(blkno) ? blkno : totalBlocks;
-
-			elog(INFO, "hnsw vacuum progress: %.1f%%", 100.0 * currentBlocks / totalBlocks);
-		}
-#endif
-	}
-}
-
-/*
- * Confirm graph was repaired
- */
-static void
-ConfirmRepaired(HnswVacuumState * vacuumstate)
-{
-	BlockNumber blkno = HNSW_HEAD_BLKNO;
-	Relation	index = vacuumstate->index;
-	BufferAccessStrategy bas = vacuumstate->bas;
-
-	while (BlockNumberIsValid(blkno))
-	{
-		Buffer		buf;
-		Page		page;
-		OffsetNumber offno;
-		OffsetNumber maxoffno;
-
-		vacuum_delay_point();
-
-		buf = ReadBufferExtended(index, MAIN_FORKNUM, blkno, RBM_NORMAL, bas);
-		LockBuffer(buf, BUFFER_LOCK_SHARE);
-		page = BufferGetPage(buf);
-		maxoffno = PageGetMaxOffsetNumber(page);
-
-		/* Iterate over nodes */
-		for (offno = FirstOffsetNumber; offno <= maxoffno; offno = OffsetNumberNext(offno))
-		{
-			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, PageGetItemId(page, offno));
-			HnswNeighborTuple ntup;
-			Buffer		nbuf;
-			Page		npage;
-			BlockNumber neighborPage;
-			OffsetNumber neighborOffno;
-
-			/* Skip neighbor tuples */
-			if (!HnswIsElementTuple(etup))
-				continue;
-
-			/* Skip if being deleted */
-			if (!ItemPointerIsValid(&etup->heaptids[0]))
-				continue;
-
-			/* Get neighbor page */
-			neighborPage = ItemPointerGetBlockNumber(&etup->neighbortid);
-			neighborOffno = ItemPointerGetOffsetNumber(&etup->neighbortid);
-
-			if (neighborPage == blkno)
-			{
-				nbuf = buf;
-				npage = page;
-			}
-			else
-			{
-				nbuf = ReadBufferExtended(index, MAIN_FORKNUM, neighborPage, RBM_NORMAL, bas);
-				LockBuffer(nbuf, BUFFER_LOCK_SHARE);
-				npage = BufferGetPage(nbuf);
-			}
-
-			ntup = (HnswNeighborTuple) PageGetItem(npage, PageGetItemId(npage, neighborOffno));
-
-			/* Check neighbors */
-			for (int i = 0; i < ntup->count; i++)
-			{
-				ItemPointer indextid = &ntup->indextids[i];
-
-				if (!ItemPointerIsValid(indextid))
-					continue;
-
-				/* Check if in deleted list */
-				if (DeletedContains(vacuumstate->deleted, indextid))
-					elog(ERROR, "hnsw graph not repaired");
-			}
-
-			if (nbuf != buf)
-				UnlockReleaseBuffer(nbuf);
-		}
-
-		blkno = HnswPageGetOpaque(page)->nextblkno;
-
-		UnlockReleaseBuffer(buf);
 	}
 }
 
@@ -761,11 +607,7 @@ InitVacuumState(HnswVacuumState * vacuumstate, IndexVacuumInfo *info, IndexBulkD
 	HnswGetMetaPageInfo(index, &vacuumstate->m, NULL);
 
 	/* Create hash table */
-#if PG_VERSION_NUM >= 170000
-	vacuumstate->deleted = TidStoreCreateLocal((AmAutoVacuumWorkerProcess() && autovacuum_work_mem != -1) ? autovacuum_work_mem : maintenance_work_mem, true);
-#else
 	vacuumstate->deleted = tidhash_create(CurrentMemoryContext, 256, NULL);
-#endif
 }
 
 /*
@@ -774,11 +616,7 @@ InitVacuumState(HnswVacuumState * vacuumstate, IndexVacuumInfo *info, IndexBulkD
 static void
 FreeVacuumState(HnswVacuumState * vacuumstate)
 {
-#if PG_VERSION_NUM >= 170000
-	TidStoreDestroy(vacuumstate->deleted);
-#else
 	tidhash_destroy(vacuumstate->deleted);
-#endif
 	FreeAccessStrategy(vacuumstate->bas);
 	pfree(vacuumstate->ntup);
 	MemoryContextDelete(vacuumstate->tmpCtx);
@@ -796,16 +634,13 @@ hnswbulkdelete(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
 	InitVacuumState(&vacuumstate, info, stats, callback, callback_state);
 
 	/* Pass 1: Remove heap TIDs */
-	HnswBench("RemoveHeapTids", RemoveHeapTids(&vacuumstate));
+	RemoveHeapTids(&vacuumstate);
 
 	/* Pass 2: Repair graph */
-	HnswBench("RepairGraph", RepairGraph(&vacuumstate));
+	RepairGraph(&vacuumstate);
 
-	/* Pass 3: Confirm repaired */
+	/* Pass 3: Mark as deleted */
-	HnswBench("ConfirmRepaired", ConfirmRepaired(&vacuumstate));
+	MarkDeleted(&vacuumstate);
-
-	/* Pass 4: Mark as deleted */
-	HnswBench("MarkDeleted", MarkDeleted(&vacuumstate));
 
 	FreeVacuumState(&vacuumstate);
 

src/ivfbuild.c

@@ -22,7 +22,6 @@
 #include "nodes/execnodes.h"
 #include "optimizer/optimizer.h"
 #include "storage/bufmgr.h"
-#include "storage/condition_variable.h"
 #include "tcop/tcopprot.h"
 #include "utils/memutils.h"
 #include "utils/rel.h"
@@ -63,13 +62,15 @@ AddSample(Datum *values, IvfflatBuildState * buildstate)
 	Datum		value = PointerGetDatum(PG_DETOAST_DATUM(values[0]));
 
 	/*
-	 * Check with KMEANS_NORM_PROC that the value can be normalized since
+	 * Normalize with KMEANS_NORM_PROC since spherical distance function
-	 * spherical distance function expects unit vectors
+	 * expects unit vectors
 	 */
 	if (buildstate->kmeansnormprocinfo != NULL)
 	{
 		if (!IvfflatCheckNorm(buildstate->kmeansnormprocinfo, buildstate->collation, value))
 			return;
+
+		value = IvfflatNormValue(buildstate->typeInfo, buildstate->collation, value);
 	}
 
 	if (samples->length < targsamples)
@@ -80,7 +81,7 @@ AddSample(Datum *values, IvfflatBuildState * buildstate)
 	else
 	{
 		if (buildstate->rowstoskip < 0)
-			buildstate->rowstoskip = reservoir_get_next_S(&buildstate->rstate, buildstate->samplerows, targsamples);
+			buildstate->rowstoskip = reservoir_get_next_S(&buildstate->rstate, samples->length, targsamples);
 
 		if (buildstate->rowstoskip <= 0)
 		{
@@ -96,9 +97,6 @@ AddSample(Datum *values, IvfflatBuildState * buildstate)
 
 		buildstate->rowstoskip -= 1;
 	}
-
-	/* Increment after reservoir_get_next_S */
-	buildstate->samplerows += 1;
 }
 
 /*
@@ -135,7 +133,6 @@ SampleRows(IvfflatBuildState * buildstate)
 	int			targsamples = buildstate->samples->maxlen;
 	BlockNumber totalblocks = RelationGetNumberOfBlocks(buildstate->heap);
 
-	buildstate->samplerows = 0;
 	buildstate->rowstoskip = -1;
 
 	BlockSampler_Init(&buildstate->bs, totalblocks, targsamples, RandomInt());
@@ -145,24 +142,8 @@ SampleRows(IvfflatBuildState * buildstate)
 	{
 		BlockNumber targblock = BlockSampler_Next(&buildstate->bs);
 
-		/* Set anyvisible to false like table_index_build_scan */
 		table_index_build_range_scan(buildstate->heap, buildstate->index, buildstate->indexInfo,
-									 false, false, false, targblock, 1, SampleCallback, (void *) buildstate, NULL);
+									 false, true, false, targblock, 1, SampleCallback, (void *) buildstate, NULL);
-	}
-
-	/* Normalize if needed */
-	if (buildstate->kmeansnormprocinfo != NULL)
-	{
-		VectorArray samples = buildstate->samples;
-
-		for (int i = 0; i < samples->length; i++)
-		{
-			Datum		value = PointerGetDatum(VectorArrayGet(samples, i));
-			Datum		normValue = IvfflatNormValue(buildstate->typeInfo, buildstate->collation, value);
-
-			VectorArraySet(samples, i, DatumGetPointer(normValue));
-			pfree(DatumGetPointer(normValue));
-		}
 	}
 }
 
@@ -393,13 +374,9 @@ InitBuildState(IvfflatBuildState * buildstate, Relation heap, Relation index, In
 	TupleDescInitEntry(buildstate->sortdesc, (AttrNumber) 1, "list", INT4OID, -1, 0);
 	TupleDescInitEntry(buildstate->sortdesc, (AttrNumber) 2, "tid", TIDOID, -1, 0);
 	TupleDescInitEntry(buildstate->sortdesc, (AttrNumber) 3, "vector", TupleDescAttr(buildstate->tupdesc, 0)->atttypid, -1, 0);
-#if PG_VERSION_NUM >= 190000
-	TupleDescFinalize(buildstate->sortdesc);
-#endif
 
 	buildstate->slot = MakeSingleTupleTableSlot(buildstate->sortdesc, &TTSOpsVirtual);
 
-	/* TODO Ensure within maintenance_work_mem */
 	buildstate->centers = VectorArrayInit(buildstate->lists, buildstate->dimensions, buildstate->typeInfo->itemSize(buildstate->dimensions));
 	buildstate->listInfo = palloc(sizeof(ListInfo) * buildstate->lists);
 
@@ -458,7 +435,7 @@ ComputeCenters(IvfflatBuildState * buildstate)
 	buildstate->samples = VectorArrayInit(numSamples, buildstate->dimensions, buildstate->centers->itemsize);
 	if (buildstate->heap != NULL)
 	{
-		IvfflatBench("sample rows", SampleRows(buildstate));
+		SampleRows(buildstate);
 
 		if (buildstate->samples->length < buildstate->lists)
 		{
@@ -673,11 +650,7 @@ IvfflatParallelScanAndSort(IvfflatSpool * ivfspool, IvfflatShared * ivfshared, S
 	ivfspool->sortstate = InitBuildSortState(buildstate.sortdesc, sortmem, coordinate);
 	buildstate.sortstate = ivfspool->sortstate;
 	scan = table_beginscan_parallel(ivfspool->heap,
-									ParallelTableScanFromIvfflatShared(ivfshared)
+									ParallelTableScanFromIvfflatShared(ivfshared));
-#if PG_VERSION_NUM >= 190000
-									,SO_NONE
-#endif
-		);
 	reltuples = table_index_build_scan(ivfspool->heap, ivfspool->index, indexInfo,
 									   true, progress, BuildCallback,
 									   (void *) &buildstate, scan);

src/ivfflat.h

@@ -9,7 +9,6 @@
 #include "lib/pairingheap.h"
 #include "nodes/execnodes.h"
 #include "port.h"				/* for random() */
-#include "storage/condition_variable.h"
 #include "utils/sampling.h"
 #include "utils/tuplesort.h"
 #include "vector.h"
@@ -214,8 +213,7 @@ typedef struct IvfflatBuildState
 	/* Sampling */
 	BlockSamplerData bs;
 	ReservoirStateData rstate;
-	double		samplerows;
+	int			rowstoskip;
-	double		rowstoskip;
 
 	/* Sorting */
 	Tuplesortstate *sortstate;

src/ivfscan.c

@@ -298,9 +298,6 @@ ivfflatbeginscan(Relation index, int nkeys, int norderbys)
 	so->tupdesc = CreateTemplateTupleDesc(2);
 	TupleDescInitEntry(so->tupdesc, (AttrNumber) 1, "distance", FLOAT8OID, -1, 0);
 	TupleDescInitEntry(so->tupdesc, (AttrNumber) 2, "heaptid", TIDOID, -1, 0);
-#if PG_VERSION_NUM >= 190000
-	TupleDescFinalize(so->tupdesc);
-#endif
 
 	/* Prep sort */
 	so->sortstate = InitScanSortState(so->tupdesc);

src/sparsevec.c

@@ -182,10 +182,10 @@ sparsevec_isspace(char ch)
 static int
 CompareIndices(const void *a, const void *b)
 {
-	if (((const SparseInputElement *) a)->index < ((const SparseInputElement *) b)->index)
+	if (((SparseInputElement *) a)->index < ((SparseInputElement *) b)->index)
 		return -1;
 
-	if (((const SparseInputElement *) a)->index > ((const SparseInputElement *) b)->index)
+	if (((SparseInputElement *) a)->index > ((SparseInputElement *) b)->index)
 		return 1;
 
 	return 0;
@@ -895,28 +895,24 @@ SparsevecInnerProduct(SparseVector * a, SparseVector * b)
 	float	   *ax = SPARSEVEC_VALUES(a);
 	float	   *bx = SPARSEVEC_VALUES(b);
 	float		distance = 0.0;
-	int			bpos = 0;
+	int			i = 0;
+	int			j = 0;
 
-	for (int i = 0; i < a->nnz; i++)
+	while (i < a->nnz && j < b->nnz)
 	{
 		int			ai = a->indices[i];
+		int			bi = b->indices[j];
 
-		for (int j = bpos; j < b->nnz; j++)
+		if (ai == bi)
 		{
-			int			bi = b->indices[j];
+			distance += ax[i] * bx[j];
-
+			i++;
-			/* Only update when the same index */
+			j++;
-			if (ai == bi)
-				distance += ax[i] * bx[j];
-
-			/* Update start for next iteration */
-			if (ai >= bi)
-				bpos = j + 1;
-
-			/* Found or passed it */
-			if (bi >= ai)
-				break;
 		}
+		else if (ai < bi)
+			i++;
+		else
+			j++;
 	}
 
 	return distance;

src/vector.c

@@ -40,7 +40,7 @@
 #endif
 
 #if PG_VERSION_NUM >= 180000
-PG_MODULE_MAGIC_EXT(.name = "vector", .version = "0.8.2");
+PG_MODULE_MAGIC_EXT(.name = "vector",.version = "0.8.2");
 #else
 PG_MODULE_MAGIC;
 #endif