Add basic support for using the POPCNT and SSE4.2s LZCNT opcodes

These opcodes have been around in the AMD world since 2007, and 2008 in
the case of intel.  They're supported in GCC and Clang via some __builtin
macros.  The opcodes may be unavailable during runtime, in which case we
fall back on a C-based implementation of the code.  In order to get the
POPCNT instruction we must pass the -mpopcnt option to the compiler.  We
do this only for the pg_bitutils.c file.

David Rowley (with fragments taken from a patch by Thomas Munro)

Discussion: https://postgr.es/m/CAKJS1f9WTAGG1tPeJnD18hiQW5gAk59fQ6WK-vfdAKEHyRg2RA@mail.gmail.com

config/c-compiler.m4

@@ -378,6 +378,122 @@ fi])# PGAC_C_BUILTIN_OP_OVERFLOW
 
 
 
+# PGAC_C_BUILTIN_POPCOUNT
+# -------------------------
+AC_DEFUN([PGAC_C_BUILTIN_POPCOUNT],
+[define([Ac_cachevar], [AS_TR_SH([pgac_cv_popcount])])dnl
+AC_CACHE_CHECK([for __builtin_popcount], [Ac_cachevar],
+[pgac_save_CFLAGS=$CFLAGS
+CFLAGS="$pgac_save_CFLAGS -mpopcnt"
+AC_COMPILE_IFELSE([AC_LANG_SOURCE(
+[static int x = __builtin_popcount(255);])],
+[Ac_cachevar=yes],
+[Ac_cachevar=no])
+CFLAGS="$pgac_save_CFLAGS"])
+if test x"$Ac_cachevar" = x"yes"; then
+  CFLAGS_POPCNT="-mpopcnt"
+AC_DEFINE(HAVE__BUILTIN_POPCOUNT, 1,
+          [Define to 1 if your compiler understands __builtin_popcount.])
+fi
+undefine([Ac_cachevar])dnl
+])# PGAC_C_BUILTIN_POPCOUNT
+
+
+
+# PGAC_C_BUILTIN_POPCOUNTL
+# -------------------------
+AC_DEFUN([PGAC_C_BUILTIN_POPCOUNTL],
+[define([Ac_cachevar], [AS_TR_SH([pgac_cv_popcountl])])dnl
+AC_CACHE_CHECK([for __builtin_popcountl], [Ac_cachevar],
+[pgac_save_CFLAGS=$CFLAGS
+CFLAGS="$pgac_save_CFLAGS -mpopcnt"
+AC_COMPILE_IFELSE([AC_LANG_SOURCE(
+[static int x = __builtin_popcountl(255);])],
+[Ac_cachevar=yes],
+[Ac_cachevar=no])
+CFLAGS="$pgac_save_CFLAGS"])
+if test x"$Ac_cachevar" = x"yes"; then
+  CFLAGS_POPCNT="-mpopcnt"
+AC_DEFINE(HAVE__BUILTIN_POPCOUNTL, 1,
+          [Define to 1 if your compiler understands __builtin_popcountl.])
+fi
+undefine([Ac_cachevar])dnl
+])# PGAC_C_BUILTIN_POPCOUNTL
+
+
+
+# PGAC_C_BUILTIN_CTZ
+# -------------------------
+# Check if the C compiler understands __builtin_ctz(),
+# and define HAVE__BUILTIN_CTZ if so.
+AC_DEFUN([PGAC_C_BUILTIN_CTZ],
+[AC_CACHE_CHECK(for __builtin_ctz, pgac_cv__builtin_ctz,
+[AC_COMPILE_IFELSE([AC_LANG_SOURCE(
+[static int x = __builtin_ctz(256);]
+)],
+[pgac_cv__builtin_ctz=yes],
+[pgac_cv__builtin_ctz=no])])
+if test x"$pgac_cv__builtin_ctz" = xyes ; then
+AC_DEFINE(HAVE__BUILTIN_CTZ, 1,
+          [Define to 1 if your compiler understands __builtin_ctz.])
+fi])# PGAC_C_BUILTIN_CTZ
+
+
+
+# PGAC_C_BUILTIN_CTZL
+# -------------------------
+# Check if the C compiler understands __builtin_ctzl(),
+# and define HAVE__BUILTIN_CTZL if so.
+AC_DEFUN([PGAC_C_BUILTIN_CTZL],
+[AC_CACHE_CHECK(for __builtin_ctzl, pgac_cv__builtin_ctzl,
+[AC_COMPILE_IFELSE([AC_LANG_SOURCE(
+[static int x = __builtin_ctzl(256);]
+)],
+[pgac_cv__builtin_ctzl=yes],
+[pgac_cv__builtin_ctzl=no])])
+if test x"$pgac_cv__builtin_ctzl" = xyes ; then
+AC_DEFINE(HAVE__BUILTIN_CTZL, 1,
+          [Define to 1 if your compiler understands __builtin_ctzl.])
+fi])# PGAC_C_BUILTIN_CTZL
+
+
+
+# PGAC_C_BUILTIN_CLZ
+# -------------------------
+# Check if the C compiler understands __builtin_clz(),
+# and define HAVE__BUILTIN_CLZ if so.
+AC_DEFUN([PGAC_C_BUILTIN_CLZ],
+[AC_CACHE_CHECK(for __builtin_clz, pgac_cv__builtin_clz,
+[AC_COMPILE_IFELSE([AC_LANG_SOURCE(
+[static int x = __builtin_clz(256);]
+)],
+[pgac_cv__builtin_clz=yes],
+[pgac_cv__builtin_clz=no])])
+if test x"$pgac_cv__builtin_clz" = xyes ; then
+AC_DEFINE(HAVE__BUILTIN_CLZ, 1,
+          [Define to 1 if your compiler understands __builtin_clz.])
+fi])# PGAC_C_BUILTIN_CLZ
+
+
+
+# PGAC_C_BUILTIN_CLZL
+# -------------------------
+# Check if the C compiler understands __builtin_clzl(),
+# and define HAVE__BUILTIN_CLZL if so.
+AC_DEFUN([PGAC_C_BUILTIN_CLZL],
+[AC_CACHE_CHECK(for __builtin_clzl, pgac_cv__builtin_clzl,
+[AC_COMPILE_IFELSE([AC_LANG_SOURCE(
+[static int x = __builtin_clzl(256);]
+)],
+[pgac_cv__builtin_clzl=yes],
+[pgac_cv__builtin_clzl=no])])
+if test x"$pgac_cv__builtin_clzl" = xyes ; then
+AC_DEFINE(HAVE__BUILTIN_CLZL, 1,
+          [Define to 1 if your compiler understands __builtin_clzl.])
+fi])# PGAC_C_BUILTIN_CLZL
+
+
+
 # PGAC_C_BUILTIN_UNREACHABLE
 # --------------------------
 # Check if the C compiler understands __builtin_unreachable(),

configure

@@ -651,6 +651,7 @@ CFLAGS_ARMV8_CRC32C
 CFLAGS_SSE42
 have_win32_dbghelp
 LIBOBJS
+CFLAGS_POPCNT
 UUID_LIBS
 LDAP_LIBS_BE
 LDAP_LIBS_FE
@@ -14059,6 +14060,158 @@ if test x"$pgac_cv__builtin_constant_p" = xyes ; then
 
 $as_echo "#define HAVE__BUILTIN_CONSTANT_P 1" >>confdefs.h
 
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __builtin_popcount" >&5
+$as_echo_n "checking for __builtin_popcount... " >&6; }
+if ${pgac_cv_popcount+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  pgac_save_CFLAGS=$CFLAGS
+CFLAGS="$pgac_save_CFLAGS -mpopcnt"
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+static int x = __builtin_popcount(255);
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  pgac_cv_popcount=yes
+else
+  pgac_cv_popcount=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+CFLAGS="$pgac_save_CFLAGS"
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv_popcount" >&5
+$as_echo "$pgac_cv_popcount" >&6; }
+if test x"$pgac_cv_popcount" = x"yes"; then
+  CFLAGS_POPCNT="-mpopcnt"
+
+$as_echo "#define HAVE__BUILTIN_POPCOUNT 1" >>confdefs.h
+
+fi
+
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __builtin_popcountl" >&5
+$as_echo_n "checking for __builtin_popcountl... " >&6; }
+if ${pgac_cv_popcountl+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  pgac_save_CFLAGS=$CFLAGS
+CFLAGS="$pgac_save_CFLAGS -mpopcnt"
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+static int x = __builtin_popcountl(255);
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  pgac_cv_popcountl=yes
+else
+  pgac_cv_popcountl=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+CFLAGS="$pgac_save_CFLAGS"
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv_popcountl" >&5
+$as_echo "$pgac_cv_popcountl" >&6; }
+if test x"$pgac_cv_popcountl" = x"yes"; then
+  CFLAGS_POPCNT="-mpopcnt"
+
+$as_echo "#define HAVE__BUILTIN_POPCOUNTL 1" >>confdefs.h
+
+fi
+
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __builtin_ctz" >&5
+$as_echo_n "checking for __builtin_ctz... " >&6; }
+if ${pgac_cv__builtin_ctz+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+static int x = __builtin_ctz(256);
+
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  pgac_cv__builtin_ctz=yes
+else
+  pgac_cv__builtin_ctz=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__builtin_ctz" >&5
+$as_echo "$pgac_cv__builtin_ctz" >&6; }
+if test x"$pgac_cv__builtin_ctz" = xyes ; then
+
+$as_echo "#define HAVE__BUILTIN_CTZ 1" >>confdefs.h
+
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __builtin_ctzl" >&5
+$as_echo_n "checking for __builtin_ctzl... " >&6; }
+if ${pgac_cv__builtin_ctzl+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+static int x = __builtin_ctzl(256);
+
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  pgac_cv__builtin_ctzl=yes
+else
+  pgac_cv__builtin_ctzl=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__builtin_ctzl" >&5
+$as_echo "$pgac_cv__builtin_ctzl" >&6; }
+if test x"$pgac_cv__builtin_ctzl" = xyes ; then
+
+$as_echo "#define HAVE__BUILTIN_CTZL 1" >>confdefs.h
+
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __builtin_clz" >&5
+$as_echo_n "checking for __builtin_clz... " >&6; }
+if ${pgac_cv__builtin_clz+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+static int x = __builtin_clz(256);
+
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  pgac_cv__builtin_clz=yes
+else
+  pgac_cv__builtin_clz=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__builtin_clz" >&5
+$as_echo "$pgac_cv__builtin_clz" >&6; }
+if test x"$pgac_cv__builtin_clz" = xyes ; then
+
+$as_echo "#define HAVE__BUILTIN_CLZ 1" >>confdefs.h
+
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __builtin_clzl" >&5
+$as_echo_n "checking for __builtin_clzl... " >&6; }
+if ${pgac_cv__builtin_clzl+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+static int x = __builtin_clzl(256);
+
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  pgac_cv__builtin_clzl=yes
+else
+  pgac_cv__builtin_clzl=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__builtin_clzl" >&5
+$as_echo "$pgac_cv__builtin_clzl" >&6; }
+if test x"$pgac_cv__builtin_clzl" = xyes ; then
+
+$as_echo "#define HAVE__BUILTIN_CLZL 1" >>confdefs.h
+
 fi
 { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __builtin_unreachable" >&5
 $as_echo_n "checking for __builtin_unreachable... " >&6; }
@@ -14577,6 +14730,8 @@ $as_echo "#define LOCALE_T_IN_XLOCALE 1" >>confdefs.h
 
 fi
 
+
+
 # MSVC doesn't cope well with defining restrict to __restrict, the
 # spelling it understands, because it conflicts with
 # __declspec(restrict). Therefore we define pg_restrict to the

configure.in

@@ -1489,6 +1489,12 @@ PGAC_C_BUILTIN_BSWAP16
 PGAC_C_BUILTIN_BSWAP32
 PGAC_C_BUILTIN_BSWAP64
 PGAC_C_BUILTIN_CONSTANT_P
+PGAC_C_BUILTIN_POPCOUNT
+PGAC_C_BUILTIN_POPCOUNTL
+PGAC_C_BUILTIN_CTZ
+PGAC_C_BUILTIN_CTZL
+PGAC_C_BUILTIN_CLZ
+PGAC_C_BUILTIN_CLZL
 PGAC_C_BUILTIN_UNREACHABLE
 PGAC_C_COMPUTED_GOTO
 PGAC_STRUCT_TIMEZONE
@@ -1503,6 +1509,8 @@ AC_TYPE_LONG_LONG_INT
 
 PGAC_TYPE_LOCALE_T
 
+AC_SUBST(CFLAGS_POPCNT)
+
 # MSVC doesn't cope well with defining restrict to __restrict, the
 # spelling it understands, because it conflicts with
 # __declspec(restrict). Therefore we define pg_restrict to the

src/Makefile.global.in

@@ -260,6 +260,7 @@ CXX = @CXX@
 CFLAGS = @CFLAGS@
 CFLAGS_VECTOR = @CFLAGS_VECTOR@
 CFLAGS_SSE42 = @CFLAGS_SSE42@
+CFLAGS_POPCNT = @CFLAGS_POPCNT@
 CFLAGS_ARMV8_CRC32C = @CFLAGS_ARMV8_CRC32C@
 PERMIT_DECLARATION_AFTER_STATEMENT = @PERMIT_DECLARATION_AFTER_STATEMENT@
 CXXFLAGS = @CXXFLAGS@

src/backend/access/heap/visibilitymap.c

@@ -89,12 +89,12 @@
 #include "access/visibilitymap.h"
 #include "access/xlog.h"
 #include "miscadmin.h"
+#include "port/pg_bitutils.h"
 #include "storage/bufmgr.h"
 #include "storage/lmgr.h"
 #include "storage/smgr.h"
 #include "utils/inval.h"
 
-
 /*#define TRACE_VISIBILITYMAP */
 
 /*
@@ -115,43 +115,9 @@
 #define HEAPBLK_TO_MAPBYTE(x) (((x) % HEAPBLOCKS_PER_PAGE) / HEAPBLOCKS_PER_BYTE)
 #define HEAPBLK_TO_OFFSET(x) (((x) % HEAPBLOCKS_PER_BYTE) * BITS_PER_HEAPBLOCK)
 
-/* tables for fast counting of set bits for visible and frozen */
-static const uint8 number_of_ones_for_visible[256] = {
-	0, 1, 0, 1, 1, 2, 1, 2, 0, 1, 0, 1, 1, 2, 1, 2,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	0, 1, 0, 1, 1, 2, 1, 2, 0, 1, 0, 1, 1, 2, 1, 2,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	2, 3, 2, 3, 3, 4, 3, 4, 2, 3, 2, 3, 3, 4, 3, 4,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	2, 3, 2, 3, 3, 4, 3, 4, 2, 3, 2, 3, 3, 4, 3, 4,
-	0, 1, 0, 1, 1, 2, 1, 2, 0, 1, 0, 1, 1, 2, 1, 2,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	0, 1, 0, 1, 1, 2, 1, 2, 0, 1, 0, 1, 1, 2, 1, 2,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	2, 3, 2, 3, 3, 4, 3, 4, 2, 3, 2, 3, 3, 4, 3, 4,
-	1, 2, 1, 2, 2, 3, 2, 3, 1, 2, 1, 2, 2, 3, 2, 3,
-	2, 3, 2, 3, 3, 4, 3, 4, 2, 3, 2, 3, 3, 4, 3, 4
-};
-static const uint8 number_of_ones_for_frozen[256] = {
-	0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 2, 2, 1, 1, 2, 2,
-	0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 2, 2, 1, 1, 2, 2,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 2, 2, 1, 1, 2, 2,
-	0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 2, 2, 1, 1, 2, 2,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	2, 2, 3, 3, 2, 2, 3, 3, 3, 3, 4, 4, 3, 3, 4, 4,
-	2, 2, 3, 3, 2, 2, 3, 3, 3, 3, 4, 4, 3, 3, 4, 4,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	1, 1, 2, 2, 1, 1, 2, 2, 2, 2, 3, 3, 2, 2, 3, 3,
-	2, 2, 3, 3, 2, 2, 3, 3, 3, 3, 4, 4, 3, 3, 4, 4,
-	2, 2, 3, 3, 2, 2, 3, 3, 3, 3, 4, 4, 3, 3, 4, 4
-};
+/* Masks for bit counting bits in the visibility map. */
+#define VISIBLE_MASK64 0x5555555555555555 /* The lower bit of each bit pair */
+#define FROZEN_MASK64 0xaaaaaaaaaaaaaaaa /* The upper bit of each bit pair */
 
 /* prototypes for internal routines */
 static Buffer vm_readbuf(Relation rel, BlockNumber blkno, bool extend);
@@ -408,18 +374,16 @@ void
 visibilitymap_count(Relation rel, BlockNumber *all_visible, BlockNumber *all_frozen)
 {
 	BlockNumber mapBlock;
+	BlockNumber nvisible = 0;
+	BlockNumber nfrozen = 0;
 
 	/* all_visible must be specified */
 	Assert(all_visible);
 
-	*all_visible = 0;
-	if (all_frozen)
-		*all_frozen = 0;
-
 	for (mapBlock = 0;; mapBlock++)
 	{
 		Buffer		mapBuffer;
-		unsigned char *map;
+		uint64	   *map;
 		int			i;
 
 		/*
@@ -436,17 +400,30 @@ visibilitymap_count(Relation rel, BlockNumber *all_visible, BlockNumber *all_fro
 		 * immediately stale anyway if anyone is concurrently setting or
 		 * clearing bits, and we only really need an approximate value.
 		 */
-		map = (unsigned char *) PageGetContents(BufferGetPage(mapBuffer));
+		map = (uint64 *) PageGetContents(BufferGetPage(mapBuffer));
 
-		for (i = 0; i < MAPSIZE; i++)
+		StaticAssertStmt(MAPSIZE % sizeof(uint64) == 0,
+						 "unsupported MAPSIZE");
+		if (all_frozen == NULL)
+		{
+			for (i = 0; i < MAPSIZE / sizeof(uint64); i++)
+				nvisible += pg_popcount64(map[i] & VISIBLE_MASK64);
+		}
+		else
 		{
-			*all_visible += number_of_ones_for_visible[map[i]];
-			if (all_frozen)
-				*all_frozen += number_of_ones_for_frozen[map[i]];
+			for (i = 0; i < MAPSIZE / sizeof(uint64); i++)
+			{
+				nvisible += pg_popcount64(map[i] & VISIBLE_MASK64);
+				nfrozen += pg_popcount64(map[i] & FROZEN_MASK64);
+			}
 		}
 
 		ReleaseBuffer(mapBuffer);
 	}
+
+	*all_visible = nvisible;
+	if (all_frozen)
+		*all_frozen = nfrozen;
 }
 
 /*

src/backend/lib/bloomfilter.c

@@ -37,6 +37,7 @@
 
 #include "access/hash.h"
 #include "lib/bloomfilter.h"
+#include "port/pg_bitutils.h"
 
 #define MAX_HASH_FUNCS		10
 
@@ -187,19 +188,7 @@ double
 bloom_prop_bits_set(bloom_filter *filter)
 {
 	int			bitset_bytes = filter->m / BITS_PER_BYTE;
-	uint64		bits_set = 0;
-	int			i;
-
-	for (i = 0; i < bitset_bytes; i++)
-	{
-		unsigned char byte = filter->bitset[i];
-
-		while (byte)
-		{
-			bits_set++;
-			byte &= (byte - 1);
-		}
-	}
+	uint64		bits_set = pg_popcount((char *) filter->bitset, bitset_bytes);
 
 	return bits_set / (double) filter->m;
 }

src/backend/nodes/bitmapset.c

@@ -22,6 +22,7 @@
 
 #include "access/hash.h"
 #include "nodes/pg_list.h"
+#include "port/pg_bitutils.h"
 
 
 #define WORDNUM(x)	((x) / BITS_PER_BITMAPWORD)
@@ -51,79 +52,23 @@
 
 #define HAS_MULTIPLE_ONES(x)	((bitmapword) RIGHTMOST_ONE(x) != (x))
 
+/* Set the bitwise macro version we must use based on the bitmapword size */
+#if BITS_PER_BITMAPWORD == 32
 
-/*
- * Lookup tables to avoid need for bit-by-bit groveling
- *
- * rightmost_one_pos[x] gives the bit number (0-7) of the rightmost one bit
- * in a nonzero byte value x.  The entry for x=0 is never used.
- *
- * leftmost_one_pos[x] gives the bit number (0-7) of the leftmost one bit in a
- * nonzero byte value x.  The entry for x=0 is never used.
- *
- * number_of_ones[x] gives the number of one-bits (0-8) in a byte value x.
- *
- * We could make these tables larger and reduce the number of iterations
- * in the functions that use them, but bytewise shifts and masks are
- * especially fast on many machines, so working a byte at a time seems best.
- */
+#define bmw_popcount(w)			pg_popcount32(w)
+#define bmw_rightmost_one(w)	pg_rightmost_one32(w)
+#define bmw_leftmost_one(w)		pg_leftmost_one32(w)
+
+#elif BITS_PER_BITMAPWORD == 64
+
+#define bmw_popcount(w)			pg_popcount64(w)
+#define bmw_rightmost_one(w)	pg_rightmost_one64(w)
+#define bmw_leftmost_one(w)		pg_leftmost_one64(w)
+
+#else
+#error "invalid BITS_PER_BITMAPWORD"
+#endif
 
-static const uint8 rightmost_one_pos[256] = {
-	0, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	5, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	6, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	5, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	7, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	5, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	6, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	5, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0,
-	4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0
-};
-
-static const uint8 leftmost_one_pos[256] = {
-	0, 0, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3,
-	4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
-	5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
-	5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,
-	6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6,
-	6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6,
-	6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6,
-	6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7,
-	7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7
-};
-
-static const uint8 number_of_ones[256] = {
-	0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4,
-	1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
-	1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
-	2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
-	1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
-	2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
-	2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
-	3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
-	1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
-	2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
-	2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
-	3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
-	2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
-	3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
-	3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
-	4, 5, 5, 6, 5, 6, 6, 7, 5, 6, 6, 7, 6, 7, 7, 8
-};
 
 
 /*
@@ -607,12 +552,7 @@ bms_singleton_member(const Bitmapset *a)
 			if (result >= 0 || HAS_MULTIPLE_ONES(w))
 				elog(ERROR, "bitmapset has multiple members");
 			result = wordnum * BITS_PER_BITMAPWORD;
-			while ((w & 255) == 0)
-			{
-				w >>= 8;
-				result += 8;
-			}
-			result += rightmost_one_pos[w & 255];
+			result += bmw_rightmost_one(w);
 		}
 	}
 	if (result < 0)
@@ -650,12 +590,7 @@ bms_get_singleton_member(const Bitmapset *a, int *member)
 			if (result >= 0 || HAS_MULTIPLE_ONES(w))
 				return false;
 			result = wordnum * BITS_PER_BITMAPWORD;
-			while ((w & 255) == 0)
-			{
-				w >>= 8;
-				result += 8;
-			}
-			result += rightmost_one_pos[w & 255];
+			result += bmw_rightmost_one(w);
 		}
 	}
 	if (result < 0)
@@ -681,12 +616,9 @@ bms_num_members(const Bitmapset *a)
 	{
 		bitmapword	w = a->words[wordnum];
 
-		/* we assume here that bitmapword is an unsigned type */
-		while (w != 0)
-		{
-			result += number_of_ones[w & 255];
-			w >>= 8;
-		}
+		/* No need to count the bits in a zero word */
+		if (w != 0)
+			result += bmw_popcount(w);
 	}
 	return result;
 }
@@ -1041,12 +973,7 @@ bms_first_member(Bitmapset *a)
 			a->words[wordnum] &= ~w;
 
 			result = wordnum * BITS_PER_BITMAPWORD;
-			while ((w & 255) == 0)
-			{
-				w >>= 8;
-				result += 8;
-			}
-			result += rightmost_one_pos[w & 255];
+			result += bmw_rightmost_one(w);
 			return result;
 		}
 	}
@@ -1096,12 +1023,7 @@ bms_next_member(const Bitmapset *a, int prevbit)
 			int			result;
 
 			result = wordnum * BITS_PER_BITMAPWORD;
-			while ((w & 255) == 0)
-			{
-				w >>= 8;
-				result += 8;
-			}
-			result += rightmost_one_pos[w & 255];
+			result += bmw_rightmost_one(w);
 			return result;
 		}
 
@@ -1168,14 +1090,9 @@ bms_prev_member(const Bitmapset *a, int prevbit)
 		if (w != 0)
 		{
 			int			result;
-			int			shift = BITS_PER_BITMAPWORD - 8;
 
 			result = wordnum * BITS_PER_BITMAPWORD;
-
-			while ((w >> shift) == 0)
-				shift -= 8;
-
-			result += shift + leftmost_one_pos[(w >> shift) & 255];
+			result += bmw_leftmost_one(w);
 			return result;
 		}
 

src/include/pg_config.h.in

@@ -754,6 +754,24 @@
 /* Define to 1 if your compiler understands __builtin_$op_overflow. */
 #undef HAVE__BUILTIN_OP_OVERFLOW
 
+/* Define to 1 if your compiler understands __builtin_popcount. */
+#undef HAVE__BUILTIN_POPCOUNT
+
+/* Define to 1 if your compiler understands __builtin_popcountl. */
+#undef HAVE__BUILTIN_POPCOUNTL
+
+/* Define to 1 if your compiler understands __builtin_ctz. */
+#undef HAVE__BUILTIN_CTZ
+
+/* Define to 1 if your compiler understands __builtin_ctzl. */
+#undef HAVE__BUILTIN_CTZL
+
+/* Define to 1 if your compiler understands __builtin_clz. */
+#undef HAVE__BUILTIN_CLZ
+
+/* Define to 1 if your compiler understands __builtin_clzl. */
+#undef HAVE__BUILTIN_CLZL
+
 /* Define to 1 if your compiler understands __builtin_types_compatible_p. */
 #undef HAVE__BUILTIN_TYPES_COMPATIBLE_P
 

src/include/pg_config.h.win32

@@ -593,6 +593,24 @@
 /* Define to 1 if your compiler understands __builtin_$op_overflow. */
 /* #undef HAVE__BUILTIN_OP_OVERFLOW */
 
+/* Define to 1 if your compiler understands __builtin_popcount. */
+/* #undef HAVE__BUILTIN_POPCOUNT */
+
+/* Define to 1 if your compiler understands __builtin_popcountl. */
+/* #undef HAVE__BUILTIN_POPCOUNTL */
+
+/* Define to 1 if your compiler understands __builtin_ctz. */
+/* #undef HAVE__BUILTIN_CTZ */
+
+/* Define to 1 if your compiler understands __builtin_ctzl. */
+/* #undef HAVE__BUILTIN_CTZL */
+
+/* Define to 1 if your compiler understands __builtin_clz. */
+/* #undef HAVE__BUILTIN_CLZ */
+
+/* Define to 1 if your compiler understands __builtin_clzl. */
+/* #undef HAVE__BUILTIN_CLZL */
+
 /* Define to 1 if your compiler understands __builtin_types_compatible_p. */
 /* #undef HAVE__BUILTIN_TYPES_COMPATIBLE_P */
 

src/include/port/pg_bitutils.h

+/*------------------------------------------------------------------------ -
+ *
+ * pg_bitutils.h
+ *	  miscellaneous functions for bit-wise operations.
+  *
+ *
+ * Portions Copyright(c) 2019, PostgreSQL Global Development Group
+ *
+ * src/include/port/pg_bitutils.h
+ *
+ *------------------------------------------------------------------------ -
+ */
+
+#ifndef PG_BITUTILS_H
+#define PG_BITUTILS_H
+
+extern int (*pg_popcount32) (uint32 word);
+extern int (*pg_popcount64) (uint64 word);
+extern int (*pg_rightmost_one32) (uint32 word);
+extern int (*pg_rightmost_one64) (uint64 word);
+extern int (*pg_leftmost_one32) (uint32 word);
+extern int (*pg_leftmost_one64) (uint64 word);
+
+extern uint64 pg_popcount(const char *buf, int bytes);
+
+#endif							/* PG_BITUTILS_H */

src/port/Makefile

@@ -36,7 +36,7 @@ override CPPFLAGS := -I$(top_builddir)/src/port -DFRONTEND $(CPPFLAGS)
 LIBS += $(PTHREAD_LIBS)
 
 OBJS = $(LIBOBJS) $(PG_CRC32C_OBJS) chklocale.o erand48.o inet_net_ntop.o \
-	noblock.o path.o pgcheckdir.o pgmkdirp.o pgsleep.o \
+	noblock.o path.o pg_bitutils.o pgcheckdir.o pgmkdirp.o pgsleep.o \
 	pg_strong_random.o pgstrcasecmp.o pgstrsignal.o pqsignal.o \
 	qsort.o qsort_arg.o quotes.o snprintf.o sprompt.o strerror.o \
 	tar.o thread.o
@@ -78,6 +78,9 @@ pg_crc32c_armv8.o: CFLAGS+=$(CFLAGS_ARMV8_CRC32C)
 pg_crc32c_armv8_shlib.o: CFLAGS+=$(CFLAGS_ARMV8_CRC32C)
 pg_crc32c_armv8_srv.o: CFLAGS+=$(CFLAGS_ARMV8_CRC32C)
 
+# pg_bitutils.c needs CFLAGS_POPCNT
+pg_bitutils.o: CFLAGS+=$(CFLAGS_POPCNT)
+
 #
 # Shared library versions of object files
 #

src/tools/msvc/Mkvcbuild.pm

@@ -112,6 +112,7 @@ sub mkvcbuild
 		push(@pgportfiles, 'pg_crc32c_sse42_choose.c');
 		push(@pgportfiles, 'pg_crc32c_sse42.c');
 		push(@pgportfiles, 'pg_crc32c_sb8.c');
+		push(@pgportfiles, 'pg_bitutils.c');
 	}
 	else
 	{