Repair error apparently introduced in the initial coding of GUC: the

default value for geqo_effort is supposed to be 40, not 1.  The actual
'genetic' component of the GEQO algorithm has been practically disabled
since 7.1 because of this mistake.  Improve documentation while at it.

Author: tglsfdc

doc/src/sgml/runtime.sgml

@@ -1,5 +1,5 @@
 <!--
-$PostgreSQL: pgsql/doc/src/sgml/runtime.sgml,v 1.230 2004/01/10 23:28:43 neilc Exp $
+$PostgreSQL: pgsql/doc/src/sgml/runtime.sgml,v 1.231 2004/01/21 23:33:34 tgl Exp $
 -->
 
 <Chapter Id="runtime">
@@ -1404,23 +1404,25 @@ SET ENABLE_SEQSCAN TO OFF;
      </varlistentry>
 
      <varlistentry>
-      <term><varname>geqo_effort</varname> (<type>integer</type>)</term>
-      <term><varname>geqo_generations</varname> (<type>integer</type>)</term>
       <term><varname>geqo_pool_size</varname> (<type>integer</type>)</term>
+      <term><varname>geqo_generations</varname> (<type>integer</type>)</term>
+      <term><varname>geqo_effort</varname> (<type>integer</type>)</term>
       <term><varname>geqo_selection_bias</varname> (<type>floating point</type>)</term>
       <listitem>
        <para>
         Various tuning parameters for the genetic query optimization
-        algorithm: The pool size is the number of individuals in one
+        algorithm. The pool size is the number of individuals in one
         population. Valid values are between 128 and 1024. If it is set
         to 0 (the default) a pool size of 2^(QS+1), where QS is the
-        number of <literal>FROM</> items in the query, is taken. The effort is used
-        to calculate a default for generations. Valid values are between
-        1 and 80, 40 being the default. Generations specifies the number
-        of iterations in the algorithm. The number must be a positive
-        integer. If 0 is specified then <literal>Effort *
-        Log2(PoolSize)</literal> is used. The run time of the algorithm
-        is roughly proportional to the sum of pool size and generations.
+        number of <literal>FROM</> items in the query, is used.
+	Generations specifies the number of iterations of the algorithm.
+	The value must be a positive integer. If 0 is specified then
+	<literal>Effort * Log2(PoolSize)</literal> is used.
+	The run time of the algorithm is roughly proportional to the sum of
+	pool size and generations.
+	<varname>geqo_effort</varname> is only used in computing the default
+	generations setting, as just described.  The default value is 40,
+	and the allowed range 1 to 100.
         The selection bias is the selective pressure within the
         population. Values can be from 1.50 to 2.00; the latter is the
         default.

src/backend/optimizer/geqo/geqo_main.c

@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/backend/optimizer/geqo/geqo_main.c,v 1.41 2003/11/29 19:51:50 pgsql Exp $
+ * $PostgreSQL: pgsql/src/backend/optimizer/geqo/geqo_main.c,v 1.42 2004/01/21 23:33:34 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -38,13 +38,13 @@
  * Configuration options
  */
 int			Geqo_pool_size;
-int			Geqo_effort;
 int			Geqo_generations;
+int			Geqo_effort;
 double		Geqo_selection_bias;
 
 
 static int	gimme_pool_size(int nr_rel);
-static int	gimme_number_generations(int pool_size, int effort);
+static int	gimme_number_generations(int pool_size);
 
 /* define edge recombination crossover [ERX] per default */
 #if !defined(ERX) && \
@@ -92,7 +92,7 @@ geqo(Query *root, int number_of_rels, List *initial_rels)
 
 /* set GA parameters */
 	pool_size = gimme_pool_size(number_of_rels);
-	number_generations = gimme_number_generations(pool_size, Geqo_effort);
+	number_generations = gimme_number_generations(pool_size);
 	status_interval = 10;
 
 /* allocate genetic pool memory */
@@ -284,7 +284,7 @@ gimme_pool_size(int nr_rel)
 {
 	double		size;
 
-	if (Geqo_pool_size != 0)
+	if (Geqo_pool_size > 0)
 		return Geqo_pool_size;
 
 	size = pow(2.0, nr_rel + 1.0);
@@ -305,10 +305,20 @@ gimme_pool_size(int nr_rel)
  * = Effort * Log2(PoolSize)
  */
 static int
-gimme_number_generations(int pool_size, int effort)
+gimme_number_generations(int pool_size)
 {
-	if (Geqo_generations <= 0)
-		return effort * (int) ceil(log((double) pool_size) / log(2.0));
-	else
+	double		gens;
+
+	if (Geqo_generations > 0)
 		return Geqo_generations;
+
+	gens = Geqo_effort * log((double) pool_size) / log(2.0);
+
+	/* bound it to a sane range */
+	if (gens <= 0)
+		gens = 1;
+	else if (gens > 10000)
+		gens = 10000;
+
+	return (int) ceil(gens);
 }

src/backend/utils/misc/guc.c

@@ -10,7 +10,7 @@
  * Written by Peter Eisentraut <peter_e@gmx.net>.
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/utils/misc/guc.c,v 1.177 2004/01/19 19:04:40 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/utils/misc/guc.c,v 1.178 2004/01/21 23:33:34 tgl Exp $
  *
  *--------------------------------------------------------------------
  */
@@ -931,23 +931,23 @@ static struct config_int ConfigureNamesInt[] =
 		&Geqo_pool_size,
 		DEFAULT_GEQO_POOL_SIZE, 0, MAX_GEQO_POOL_SIZE, NULL, NULL
 	},
-	{
-		{"geqo_effort", PGC_USERSET, QUERY_TUNING_GEQO,
-			gettext_noop("GEQO: effort is used to calculate a default for generations."),
-			NULL
-		},
-		&Geqo_effort,
-		1, 1, INT_MAX, NULL, NULL
-	},
 	{
 		{"geqo_generations", PGC_USERSET, QUERY_TUNING_GEQO,
-			gettext_noop("GEQO: number of iterations in the algorithm."),
-			gettext_noop("The number must be a positive integer. If 0 is "
+			gettext_noop("GEQO: number of iterations of the algorithm."),
+			gettext_noop("The value must be a positive integer. If 0 is "
 						 "specified then effort * log2(poolsize) is used.")
 		},
 		&Geqo_generations,
 		0, 0, INT_MAX, NULL, NULL
 	},
+	{
+		{"geqo_effort", PGC_USERSET, QUERY_TUNING_GEQO,
+			gettext_noop("GEQO: effort is used to set the default for generations."),
+			NULL
+		},
+		&Geqo_effort,
+		DEFAULT_GEQO_EFFORT, MIN_GEQO_EFFORT, MAX_GEQO_EFFORT, NULL, NULL
+	},
 
 	{
 		{"deadlock_timeout", PGC_SIGHUP, LOCK_MANAGEMENT,

src/backend/utils/misc/postgresql.conf.sample

@@ -123,10 +123,10 @@
 
 #geqo = true
 #geqo_threshold = 11
-#geqo_effort = 1
-#geqo_generations = 0
 #geqo_pool_size = 0		# default based on tables in statement,
 				# range 128-1024
+#geqo_generations = 0		# use default: effort * log2(pool_size)
+#geqo_effort = 40		# range 1-100
 #geqo_selection_bias = 2.0	# range 1.5-2.0
 
 # - Other Planner Options -

src/include/optimizer/geqo.h

@@ -6,7 +6,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/optimizer/geqo.h,v 1.33 2003/11/29 22:41:07 pgsql Exp $
+ * $PostgreSQL: pgsql/src/include/optimizer/geqo.h,v 1.34 2004/01/21 23:33:34 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -44,22 +44,27 @@
 
 /*
  * Configuration options
+ *
+ * If you change these, update backend/utils/misc/postgresql.sample.conf
  */
-/* If you change these, update backend/utils/misc/postgresql.sample.conf */
 extern int	Geqo_pool_size;
 
 #define DEFAULT_GEQO_POOL_SIZE 0	/* = default based on no. of relations. */
 #define MIN_GEQO_POOL_SIZE 128
 #define MAX_GEQO_POOL_SIZE 1024
 
-extern int	Geqo_effort;		/* 1 .. inf, only used to calculate
-								 * generations default */
 extern int	Geqo_generations;	/* 1 .. inf, or 0 to use default based on
 								 * pool size */
 
+extern int	Geqo_effort;		/* only used to calculate default for
+								 * generations */
+
+#define DEFAULT_GEQO_EFFORT 40
+#define MIN_GEQO_EFFORT 1
+#define MAX_GEQO_EFFORT 100
+
 extern double Geqo_selection_bias;
 
-/* If you change these, update backend/utils/misc/postgresql.sample.conf */
 #define DEFAULT_GEQO_SELECTION_BIAS 2.0
 #define MIN_GEQO_SELECTION_BIAS 1.5
 #define MAX_GEQO_SELECTION_BIAS 2.0