Compare commits

..

1 Commits

Author SHA1 Message Date
Andrew Kane
4b2c593873 Updated number of parallel workers for IVFFlat index builds 2024-02-28 15:03:50 -08:00
13 changed files with 53 additions and 72 deletions

View File

@@ -60,9 +60,7 @@ jobs:
wget -q https://github.com/postgres/postgres/archive/refs/tags/REL_14_10.tar.gz
tar xf REL_14_10.tar.gz
- run: make prove_installcheck PROVE_FLAGS="-I ./postgres-REL_14_10/src/test/perl" PERL5LIB="/Users/runner/perl5/lib/perl5"
- run: make clean && /usr/local/opt/llvm@15/bin/scan-build --status-bugs make
env:
PG_CFLAGS: -DUSE_ASSERT_CHECKING
- run: make clean && /usr/local/opt/llvm@15/bin/scan-build --status-bugs make PG_CFLAGS="-DUSE_ASSERT_CHECKING"
windows:
runs-on: windows-latest
if: ${{ !startsWith(github.ref_name, 'mac') }}

View File

@@ -1,8 +1,7 @@
## 0.6.1 (2024-03-04)
## 0.6.1 (unreleased)
- Fixed error with `ANALYZE` and vectors with different dimensions
- Fixed error with `shared_preload_libraries`
- Fixed vector subtraction being marked as commutative
## 0.6.0 (2024-01-29)

View File

@@ -1,4 +1,4 @@
Portions Copyright (c) 1996-2024, PostgreSQL Global Development Group
Portions Copyright (c) 1996-2023, PostgreSQL Global Development Group
Portions Copyright (c) 1994, The Regents of the University of California

View File

@@ -2,7 +2,7 @@
"name": "vector",
"abstract": "Open-source vector similarity search for Postgres",
"description": "Supports L2 distance, inner product, and cosine distance",
"version": "0.6.1",
"version": "0.6.0",
"maintainer": [
"Andrew Kane <andrew@ankane.org>"
],
@@ -20,7 +20,7 @@
"vector": {
"file": "sql/vector.sql",
"docfile": "README.md",
"version": "0.6.1",
"version": "0.6.0",
"abstract": "Open-source vector similarity search for Postgres"
}
},

View File

@@ -1,5 +1,5 @@
EXTENSION = vector
EXTVERSION = 0.6.1
EXTVERSION = 0.6.0
MODULE_big = vector
DATA = $(wildcard sql/*--*.sql)

View File

@@ -1,5 +1,5 @@
EXTENSION = vector
EXTVERSION = 0.6.1
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

View File

@@ -20,7 +20,7 @@ Compile and install the extension (supports Postgres 12+)
```sh
cd /tmp
git clone --branch v0.6.1 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 v0.6.1 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
@@ -102,12 +102,6 @@ Insert vectors
INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]');
```
Or load vectors in bulk using `COPY` ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/bulk_loading.py))
```sql
COPY items (embedding) FROM STDIN WITH (FORMAT BINARY);
```
Upsert vectors
```sql
@@ -270,8 +264,6 @@ 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
Like other index types, its faster to create an index after loading your initial data
Starting with 0.6.0, you can also speed up index creation by increasing the number of parallel workers (2 by default)
```sql
@@ -664,14 +656,6 @@ Note: Replace `16` with your Postgres server version
If compilation fails and the output includes `warning: no such sysroot directory` on Mac, reinstall Xcode Command Line Tools.
### Portability
To compile for portability, use:
```sh
make OPTFLAGS=""
```
## Additional Installation Methods
### Docker
@@ -687,7 +671,7 @@ This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (r
You can also build the image manually:
```sh
git clone --branch v0.6.1 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=16 -t myuser/pgvector .
```

View File

@@ -1,16 +0,0 @@
-- complain if script is sourced in psql, rather than via CREATE EXTENSION
\echo Use "ALTER EXTENSION vector UPDATE TO '0.6.1'" to load this file. \quit
DROP OPERATOR - (vector, vector);
CREATE OPERATOR - (
LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_sub
);
ALTER OPERATOR <= (vector, vector) SET (
RESTRICT = scalarlesel, JOIN = scalarlejoinsel
);
ALTER OPERATOR >= (vector, vector) SET (
RESTRICT = scalargesel, JOIN = scalargejoinsel
);

View File

@@ -180,7 +180,8 @@ CREATE OPERATOR + (
);
CREATE OPERATOR - (
LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_sub
LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_sub,
COMMUTATOR = -
);
CREATE OPERATOR * (
@@ -194,10 +195,11 @@ CREATE OPERATOR < (
RESTRICT = scalarltsel, JOIN = scalarltjoinsel
);
-- should use scalarlesel and scalarlejoinsel, but not supported in Postgres < 11
CREATE OPERATOR <= (
LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_le,
COMMUTATOR = >= , NEGATOR = > ,
RESTRICT = scalarlesel, JOIN = scalarlejoinsel
RESTRICT = scalarltsel, JOIN = scalarltjoinsel
);
CREATE OPERATOR = (
@@ -212,10 +214,11 @@ CREATE OPERATOR <> (
RESTRICT = eqsel, JOIN = eqjoinsel
);
-- should use scalargesel and scalargejoinsel, but not supported in Postgres < 11
CREATE OPERATOR >= (
LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_ge,
COMMUTATOR = <= , NEGATOR = < ,
RESTRICT = scalargesel, JOIN = scalargejoinsel
RESTRICT = scalargtsel, JOIN = scalargtjoinsel
);
CREATE OPERATOR > (

View File

@@ -80,7 +80,7 @@
#if PG_VERSION_NUM < 130000
#define list_delete_last(list) list_truncate(list, list_length(list) - 1)
#define list_sort(list, cmp) ((list) = list_qsort(list, cmp))
#define list_sort(list, cmp) list_qsort(list, cmp)
#endif
#define HnswIsElementTuple(tup) ((tup)->type == HNSW_ELEMENT_TUPLE_TYPE)

View File

@@ -860,15 +860,12 @@ HnswSearchLayer(char *base, Datum q, List *ep, int ef, int lc, Relation index, F
static int
#if PG_VERSION_NUM >= 130000
CompareCandidateDistances(const ListCell *a, const ListCell *b)
{
HnswCandidate *hca = lfirst(a);
HnswCandidate *hcb = lfirst(b);
#else
CompareCandidateDistances(const void *a, const void *b)
{
HnswCandidate *hca = lfirst(*(ListCell **) a);
HnswCandidate *hcb = lfirst(*(ListCell **) b);
#endif
{
HnswCandidate *hca = lfirst((ListCell *) a);
HnswCandidate *hcb = lfirst((ListCell *) b);
if (hca->distance < hcb->distance)
return 1;
@@ -891,15 +888,12 @@ CompareCandidateDistances(const void *a, const void *b)
static int
#if PG_VERSION_NUM >= 130000
CompareCandidateDistancesOffset(const ListCell *a, const ListCell *b)
{
HnswCandidate *hca = lfirst(a);
HnswCandidate *hcb = lfirst(b);
#else
CompareCandidateDistancesOffset(const void *a, const void *b)
{
HnswCandidate *hca = lfirst(*(ListCell **) a);
HnswCandidate *hcb = lfirst(*(ListCell **) b);
#endif
{
HnswCandidate *hca = lfirst((ListCell *) a);
HnswCandidate *hcb = lfirst((ListCell *) b);
if (hca->distance < hcb->distance)
return 1;
@@ -958,9 +952,7 @@ SelectNeighbors(char *base, List *c, int lm, int lc, FmgrInfo *procinfo, Oid col
{
List *r = NIL;
List *w = list_copy(c);
HnswCandidate **wd;
int wdlen = 0;
int wdoff = 0;
pairingheap *wd;
HnswNeighborArray *neighbors = HnswGetNeighbors(base, e2, lc);
bool mustCalculate = !neighbors->closerSet;
List *added = NIL;
@@ -969,7 +961,7 @@ SelectNeighbors(char *base, List *c, int lm, int lc, FmgrInfo *procinfo, Oid col
if (list_length(w) <= lm)
return w;
wd = palloc(sizeof(HnswCandidate *) * list_length(w));
wd = pairingheap_allocate(CompareNearestCandidates, NULL);
/* Ensure order of candidates is deterministic for closer caching */
if (sortCandidates)
@@ -1035,21 +1027,21 @@ SelectNeighbors(char *base, List *c, int lm, int lc, FmgrInfo *procinfo, Oid col
if (e->closer)
r = lappend(r, e);
else
wd[wdlen++] = e;
pairingheap_add(wd, &(CreatePairingHeapNode(e)->ph_node));
}
/* Cached value can only be used in future if sorted deterministically */
neighbors->closerSet = sortCandidates;
/* Keep pruned connections */
while (wdoff < wdlen && list_length(r) < lm)
r = lappend(r, wd[wdoff++]);
while (!pairingheap_is_empty(wd) && list_length(r) < lm)
r = lappend(r, ((HnswPairingHeapNode *) pairingheap_remove_first(wd))->inner);
/* Return pruned for update connections */
if (pruned != NULL)
{
if (wdoff < wdlen)
*pruned = wd[wdoff];
if (!pairingheap_is_empty(wd))
*pruned = ((HnswPairingHeapNode *) pairingheap_first(wd))->inner;
else
*pruned = linitial(w);
}

View File

@@ -904,6 +904,27 @@ IvfflatBeginParallel(IvfflatBuildState * buildstate, bool isconcurrent, int requ
WaitForParallelWorkersToAttach(pcxt);
}
/*
* Compute parallel workers
*/
static int
ComputeParallelWorkers(Relation heap, Relation index)
{
int parallel_workers;
/* Make sure it's safe to use parallel workers */
parallel_workers = plan_create_index_workers(RelationGetRelid(heap), RelationGetRelid(index));
if (parallel_workers == 0)
return 0;
/* Use parallel_workers storage parameter on table if set */
parallel_workers = RelationGetParallelWorkers(heap, -1);
if (parallel_workers != -1)
return Min(parallel_workers, max_parallel_maintenance_workers);
return max_parallel_maintenance_workers;
}
/*
* Scan table for tuples to index
*/
@@ -923,7 +944,7 @@ AssignTuples(IvfflatBuildState * buildstate)
/* Calculate parallel workers */
if (buildstate->heap != NULL)
parallel_workers = plan_create_index_workers(RelationGetRelid(buildstate->heap), RelationGetRelid(buildstate->index));
parallel_workers = ComputeParallelWorkers(buildstate->heap, buildstate->index);
/* Attempt to launch parallel worker scan when required */
if (parallel_workers > 0)

View File

@@ -1,4 +1,4 @@
comment = 'vector data type and ivfflat and hnsw access methods'
default_version = '0.6.1'
default_version = '0.6.0'
module_pathname = '$libdir/vector'
relocatable = true