Increase hash_mem_multiplier default to 2.0.

Double the default setting for hash_mem_multiplier, from 1.0 to 2.0.
This setting makes hash-based executor nodes use twice the usual
work_mem limit.

The PostgreSQL 15 release notes should have a compatibility note about
this change.

Author: Peter Geoghegan <[email protected]>
Discussion: https://postgr.es/m/CAH2-Wzndc_ROk6CY-bC6p9O53q974Y0Ey4WX8jcPbuTZYM4Q3A@mail.gmail.com

diff --git a/doc/src/sgml/config.sgml b/doc/src/sgml/config.sgml
index 53b361e7a92de0ebb3c93f1e54d163bf7f2455d5..d99bf38e677926ee64efc9315c7c9f478f3b0cf7 100644 (file)
--- a/doc/src/sgml/config.sgml
+++ b/doc/src/sgml/config.sgml
@@ -1849,9 +1849,8 @@ include_dir 'conf.d'
         operations can use.  The final limit is determined by
         multiplying <varname>work_mem</varname> by
         <varname>hash_mem_multiplier</varname>.  The default value is
-        1.0, which makes hash-based operations subject to the same
-        simple <varname>work_mem</varname> maximum as sort-based
-        operations.
+        2.0, which makes hash-based operations use twice the usual
+        <varname>work_mem</varname> base amount.
        </para>
        <para>
         Consider increasing <varname>hash_mem_multiplier</varname> in
@@ -1859,7 +1858,7 @@ include_dir 'conf.d'
         occurrence, especially when simply increasing
         <varname>work_mem</varname> results in memory pressure (memory
         pressure typically takes the form of intermittent out of
-        memory errors).  A setting of 1.5 or 2.0 may be effective with
+        memory errors).  The default setting of 2.0 is often effective with
         mixed workloads.  Higher settings in the range of 2.0 - 8.0 or
         more may be effective in environments where
         <varname>work_mem</varname> has already been increased to 40MB

diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index c26a1a73df81e11ae775694ec3d8abd38b79c7e3..3419c099b2828e2ba3908d76850f875e45ed38d0 100644 (file)
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -122,7 +122,7 @@ int         IntervalStyle = INTSTYLE_POSTGRES;
 bool       enableFsync = true;
 bool       allowSystemTableMods = false;
 int            work_mem = 4096;
-double     hash_mem_multiplier = 1.0;
+double     hash_mem_multiplier = 2.0;
 int            maintenance_work_mem = 65536;
 int            max_parallel_maintenance_workers = 2;
 

diff --git a/src/backend/utils/misc/guc.c b/src/backend/utils/misc/guc.c
index 9d0208ec98a43647dc86658a774e3bdeae5da2a1..01f373815e0a465f298890dc5b5b4118096e7fd9 100644 (file)
--- a/src/backend/utils/misc/guc.c
+++ b/src/backend/utils/misc/guc.c
@@ -3762,7 +3762,7 @@ static struct config_real ConfigureNamesReal[] =
            GUC_EXPLAIN
        },
        &hash_mem_multiplier,
-       1.0, 1.0, 1000.0,
+       2.0, 1.0, 1000.0,
        NULL, NULL, NULL
    },
 

diff --git a/src/backend/utils/misc/postgresql.conf.sample b/src/backend/utils/misc/postgresql.conf.sample
index 56d0bee6d9be6756c2b043c5fa230eff12d74ae7..4a094bb38be5415239165cf42fff7238eeedad23 100644 (file)
--- a/src/backend/utils/misc/postgresql.conf.sample
+++ b/src/backend/utils/misc/postgresql.conf.sample
@@ -136,7 +136,7 @@
 # Caution: it is not advisable to set max_prepared_transactions nonzero unless
 # you actively intend to use prepared transactions.
 #work_mem = 4MB                # min 64kB
-#hash_mem_multiplier = 1.0     # 1-1000.0 multiplier on hash table work_mem
+#hash_mem_multiplier = 2.0     # 1-1000.0 multiplier on hash table work_mem
 #maintenance_work_mem = 64MB       # min 1MB
 #autovacuum_work_mem = -1      # min 1MB, or -1 to use maintenance_work_mem
 #logical_decoding_work_mem = 64MB  # min 64kB

diff --git a/src/test/regress/expected/groupingsets.out b/src/test/regress/expected/groupingsets.out
index 4c467c1b15e69d33fa34d4f99559ba97bd9ae311..58a25b691acb7d4d59b82749b528418d4e47e8a8 100644 (file)
--- a/src/test/regress/expected/groupingsets.out
+++ b/src/test/regress/expected/groupingsets.out
@@ -1574,6 +1574,7 @@ select array(select row(v.a,s1.*) from (select two,four, count(*) from onek grou
 
 -- test the knapsack
 set enable_indexscan = false;
+set hash_mem_multiplier = 1.0;
 set work_mem = '64kB';
 explain (costs off)
   select unique1,
@@ -1919,6 +1920,7 @@ select g100, g10, sum(g::numeric), count(*), max(g::text)
 from gs_data_1 group by cube (g1000, g100,g10);
 set enable_sort = true;
 set work_mem to default;
+set hash_mem_multiplier to default;
 -- Compare results
 (select * from gs_hash_1 except select * from gs_group_1)
   union all

diff --git a/src/test/regress/expected/join_hash.out b/src/test/regress/expected/join_hash.out
index 3a91c144a27fc8cfe4594d784f73349525141fcd..3ec07bc1af4d52022e1b6b52c5ccced7d2962f7f 100644 (file)
--- a/src/test/regress/expected/join_hash.out
+++ b/src/test/regress/expected/join_hash.out
@@ -86,6 +86,7 @@ alter table wide set (parallel_workers = 2);
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
                QUERY PLAN               
@@ -119,6 +120,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -156,6 +158,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -196,6 +199,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
                QUERY PLAN               
@@ -229,6 +233,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -266,6 +271,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '192kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -307,6 +313,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) FROM simple r JOIN bigger_than_it_looks s USING (id);
                       QUERY PLAN                      
@@ -340,6 +347,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join bigger_than_it_looks s using (id);
@@ -377,6 +385,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 1;
 set local work_mem = '192kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join bigger_than_it_looks s using (id);
@@ -419,6 +428,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from simple r join extremely_skewed s using (id);
                     QUERY PLAN                    
@@ -451,6 +461,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join extremely_skewed s using (id);
@@ -486,6 +497,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 1;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join extremely_skewed s using (id);
@@ -523,6 +535,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 set local parallel_leader_participation = off;
 select * from hash_join_batches(
 $$
@@ -551,6 +564,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '64kB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -602,6 +616,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '4MB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -653,6 +668,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '64kB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -704,6 +720,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '4MB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -843,6 +860,7 @@ savepoint settings;
 set max_parallel_workers_per_gather = 2;
 set enable_parallel_hash = on;
 set work_mem = '128kB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select length(max(s.t))
   from wide left join (select id, coalesce(t, '') || '' as t from wide) s using (id);

diff --git a/src/test/regress/expected/memoize.out b/src/test/regress/expected/memoize.out
index 4ca0bd1f1e1ef0ab103187ede3d35ea015040db5..00438eb1ea07a2ada626abd5738e7082c4e83e7d 100644 (file)
--- a/src/test/regress/expected/memoize.out
+++ b/src/test/regress/expected/memoize.out
@@ -90,8 +90,9 @@ WHERE t1.unique1 < 1000;
   1000 | 9.5000000000000000
 (1 row)
 
--- Reduce work_mem so that we see some cache evictions
+-- Reduce work_mem and hash_mem_multiplier so that we see some cache evictions
 SET work_mem TO '64kB';
+SET hash_mem_multiplier TO 1.0;
 SET enable_mergejoin TO off;
 -- Ensure we get some evictions.  We're unable to validate the hits and misses
 -- here as the number of entries that fit in the cache at once will vary
@@ -238,6 +239,7 @@ WHERE unique1 < 3
 RESET enable_seqscan;
 RESET enable_mergejoin;
 RESET work_mem;
+RESET hash_mem_multiplier;
 RESET enable_bitmapscan;
 RESET enable_hashjoin;
 -- Test parallel plans with Memoize

diff --git a/src/test/regress/sql/groupingsets.sql b/src/test/regress/sql/groupingsets.sql
index 394494470449a28fc3f2e475ea9ff6817ed18902..473d21f6b97a89a3d6cb41e13118f537c013a5b5 100644 (file)
--- a/src/test/regress/sql/groupingsets.sql
+++ b/src/test/regress/sql/groupingsets.sql
@@ -424,6 +424,7 @@ select array(select row(v.a,s1.*) from (select two,four, count(*) from onek grou
 -- test the knapsack
 
 set enable_indexscan = false;
+set hash_mem_multiplier = 1.0;
 set work_mem = '64kB';
 explain (costs off)
   select unique1,
@@ -519,6 +520,7 @@ from gs_data_1 group by cube (g1000, g100,g10);
 
 set enable_sort = true;
 set work_mem to default;
+set hash_mem_multiplier to default;
 
 -- Compare results
 

diff --git a/src/test/regress/sql/join_hash.sql b/src/test/regress/sql/join_hash.sql
index 68c1a8c7b65e48767790455f83d32082c89c32b1..77dbc182d5380e985ffcdc979a7a68013f67dc47 100644 (file)
--- a/src/test/regress/sql/join_hash.sql
+++ b/src/test/regress/sql/join_hash.sql
@@ -95,6 +95,7 @@ alter table wide set (parallel_workers = 2);
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
 select count(*) from simple r join simple s using (id);
@@ -109,6 +110,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -124,6 +126,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -143,6 +146,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
 select count(*) from simple r join simple s using (id);
@@ -157,6 +161,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -172,6 +177,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '192kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join simple s using (id);
@@ -192,6 +198,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) FROM simple r JOIN bigger_than_it_looks s USING (id);
 select count(*) FROM simple r JOIN bigger_than_it_looks s USING (id);
@@ -206,6 +213,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join bigger_than_it_looks s using (id);
@@ -221,6 +229,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 1;
 set local work_mem = '192kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join bigger_than_it_looks s using (id);
@@ -242,6 +251,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 0;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from simple r join extremely_skewed s using (id);
 select count(*) from simple r join extremely_skewed s using (id);
@@ -255,6 +265,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = off;
 explain (costs off)
   select count(*) from simple r join extremely_skewed s using (id);
@@ -269,6 +280,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 1;
 set local work_mem = '128kB';
+set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join extremely_skewed s using (id);
@@ -285,6 +297,7 @@ rollback to settings;
 savepoint settings;
 set local max_parallel_workers_per_gather = 2;
 set local work_mem = '4MB';
+set local hash_mem_multiplier = 1.0;
 set local parallel_leader_participation = off;
 select * from hash_join_batches(
 $$
@@ -311,6 +324,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '64kB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -338,6 +352,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '4MB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -365,6 +380,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '64kB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -392,6 +408,7 @@ set max_parallel_workers_per_gather = 2;
 set enable_material = off;
 set enable_mergejoin = off;
 set work_mem = '4MB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select count(*) from join_foo
     left join (select b1.id, b1.t from join_bar b1 join join_bar b2 using (id)) ss
@@ -454,6 +471,7 @@ savepoint settings;
 set max_parallel_workers_per_gather = 2;
 set enable_parallel_hash = on;
 set work_mem = '128kB';
+set hash_mem_multiplier = 1.0;
 explain (costs off)
   select length(max(s.t))
   from wide left join (select id, coalesce(t, '') || '' as t from wide) s using (id);

diff --git a/src/test/regress/sql/memoize.sql b/src/test/regress/sql/memoize.sql
index c6ed5a2aa6617b19da68b516f1ffbf00b63ff1fb..0979bcdf768ca4c43a10ace0ff33eab6b57e1413 100644 (file)
--- a/src/test/regress/sql/memoize.sql
+++ b/src/test/regress/sql/memoize.sql
@@ -55,8 +55,9 @@ SELECT COUNT(*),AVG(t2.unique1) FROM tenk1 t1,
 LATERAL (SELECT t2.unique1 FROM tenk1 t2 WHERE t1.twenty = t2.unique1) t2
 WHERE t1.unique1 < 1000;
 
--- Reduce work_mem so that we see some cache evictions
+-- Reduce work_mem and hash_mem_multiplier so that we see some cache evictions
 SET work_mem TO '64kB';
+SET hash_mem_multiplier TO 1.0;
 SET enable_mergejoin TO off;
 -- Ensure we get some evictions.  We're unable to validate the hits and misses
 -- here as the number of entries that fit in the cache at once will vary
@@ -126,6 +127,7 @@ WHERE unique1 < 3
 RESET enable_seqscan;
 RESET enable_mergejoin;
 RESET work_mem;
+RESET hash_mem_multiplier;
 RESET enable_bitmapscan;
 RESET enable_hashjoin;