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 <pg@bowt.ie>
Discussion: https://postgr.es/m/CAH2-Wzndc_ROk6CY-bC6p9O53q974Y0Ey4WX8jcPbuTZYM4Q3A@mail.gmail.com

Author: petergeoghegan

doc/src/sgml/config.sgml

@@ -1849,17 +1849,16 @@ 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
         environments where spilling by query operations is a regular
         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

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;
 

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
 	},
 

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

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

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);

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

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
 

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);