mirrors/postgres
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Micro-optimize AllocSetFreeIndex() by reference to pg_bitutils code.

Browse Source 
Use __builtin_clz() where available.  Where it isn't, we can still win
a little by using the pg_leftmost_one_pos[] lookup table instead of
having a private table.

Also drop the initial right shift by ALLOC_MINBITS in favor of
subtracting ALLOC_MINBITS from the leftmost-one-pos result.  This
is a win because the compiler can fold that adjustment into other
constants it'd have to add anyway, making the shift-removal free.

Also, we can explain this coding as an unrolled form of
pg_leftmost_one_pos32(), even though that's a bit ahistorical
since it long predates pg_bitutils.h.

John Naylor, with some cosmetic adjustments by me

Discussion: https://postgr.es/m/CACPNZCuNUGMxjK7WTn_=WZnRbfASDdBxmjsVf2+m9MdmeNw_sg@mail.gmail.com
This commit is contained in:
Tom Lane 2019-12-28 17:21:17 -05:00
parent 27a3b2ad83
commit 0ce38730ac
1 changed files with 31 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

56
src/backend/utils/mmgr/aset.c
View File

@ -46,6 +46,7 @@
#include "postgres.h"
#include "port/pg_bitutils.h"
#include "utils/memdebug.h"
#include "utils/memutils.h"
@ -297,18 +298,6 @@ static const MemoryContextMethods AllocSetMethods = {
#endif
};
/*
* Table for AllocSetFreeIndex
*/
#define LT16(n) n, n, n, n, n, n, n, n, n, n, n, n, n, n, n, n
static const unsigned char LogTable256[256] =
{
0, 1, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4,
LT16(5), LT16(6), LT16(6), LT16(7), LT16(7), LT16(7), LT16(7),
LT16(8), LT16(8), LT16(8), LT16(8), LT16(8), LT16(8), LT16(8), LT16(8)
};
/* ----------
* Debug macros
* ----------
@ -337,24 +326,41 @@ static inline int
AllocSetFreeIndex(Size size)
{
int idx;
unsigned int t,
tsize;
if (size > (1 << ALLOC_MINBITS))
{
tsize = (size - 1) >> ALLOC_MINBITS;
/*
* At this point we need to obtain log2(tsize)+1, ie, the number of
* not-all-zero bits at the right. We used to do this with a
* shift-and-count loop, but this function is enough of a hotspot to
* justify micro-optimization effort. The best approach seems to be
* to use a lookup table. Note that this code assumes that
* ALLOCSET_NUM_FREELISTS <= 17, since we only cope with two bytes of
* the tsize value.
/*----------
* At this point we must compute ceil(log2(size >> ALLOC_MINBITS)).
* This is the same as
* pg_leftmost_one_pos32((size - 1) >> ALLOC_MINBITS) + 1
* or equivalently
* pg_leftmost_one_pos32(size - 1) - ALLOC_MINBITS + 1
*
* However, rather than just calling that function, we duplicate the
* logic here, allowing an additional optimization. It's reasonable
* to assume that ALLOC_CHUNK_LIMIT fits in 16 bits, so we can unroll
* the byte-at-a-time loop in pg_leftmost_one_pos32 and just handle
* the last two bytes.
*
* Yes, this function is enough of a hot-spot to make it worth this
* much trouble.
*----------
*/
#ifdef HAVE__BUILTIN_CLZ
idx = 31 - __builtin_clz((uint32) size - 1) - ALLOC_MINBITS + 1;
#else
uint32 t,
tsize;
/* Statically assert that we only have a 16-bit input value. */
StaticAssertStmt(ALLOC_CHUNK_LIMIT < (1 << 16),
"ALLOC_CHUNK_LIMIT must be less than 64kB");
tsize = size - 1;
t = tsize >> 8;
idx = t ? LogTable256[t] + 8 : LogTable256[tsize];
idx = t ? pg_leftmost_one_pos[t] + 8 : pg_leftmost_one_pos[tsize];
idx -= ALLOC_MINBITS - 1;
#endif
Assert(idx < ALLOCSET_NUM_FREELISTS);
}