diff options
| author | Tom Lane <tgl@sss.pgh.pa.us> | 2005-11-25 04:24:48 +0000 |
|---|---|---|
| committer | Tom Lane <tgl@sss.pgh.pa.us> | 2005-11-25 04:24:48 +0000 |
| commit | dab52ab13d3d3cce26e9bcc3193eb285c195d430 (patch) | |
| tree | d7b4ebc7af6d4e09341637179e46d212c69070af /src/backend/executor/nodeTidscan.c | |
| parent | c0a2f8cc4decba933f0ebb6d5b4ffd9eca036a78 (diff) | |
| download | postgresql-dab52ab13d3d3cce26e9bcc3193eb285c195d430.tar.gz postgresql-dab52ab13d3d3cce26e9bcc3193eb285c195d430.zip | |
Improve ExecStoreTuple to be smarter about replacing the contents of
a TupleTableSlot: instead of calling ExecClearTuple, inline the needed
operations, so that we can avoid redundant steps. In particular, when
the old and new tuples are both on the same disk page, avoid releasing
and re-acquiring the buffer pin --- this saves work in both the bufmgr
and ResourceOwner modules. To make this improvement actually useful,
partially revert a change I made on 2004-04-21 that caused SeqNext
et al to call ExecClearTuple before ExecStoreTuple. The motivation
for that, to avoid grabbing the BufMgrLock separately for releasing
the old buffer and grabbing the new one, no longer applies. My
profiling says that this saves about 5% of the CPU time for an
all-in-memory seqscan.
Diffstat (limited to 'src/backend/executor/nodeTidscan.c')
| -rw-r--r-- | src/backend/executor/nodeTidscan.c | 13 |
1 files changed, 3 insertions, 10 deletions
diff --git a/src/backend/executor/nodeTidscan.c b/src/backend/executor/nodeTidscan.c index c8708f58311..4b0775719e5 100644 --- a/src/backend/executor/nodeTidscan.c +++ b/src/backend/executor/nodeTidscan.c @@ -8,7 +8,7 @@ * * * IDENTIFICATION - * $PostgreSQL: pgsql/src/backend/executor/nodeTidscan.c,v 1.43 2005/10/15 02:49:17 momjian Exp $ + * $PostgreSQL: pgsql/src/backend/executor/nodeTidscan.c,v 1.44 2005/11/25 04:24:48 tgl Exp $ * *------------------------------------------------------------------------- */ @@ -107,13 +107,6 @@ TidNext(TidScanState *node) scanrelid = ((TidScan *) node->ss.ps.plan)->scan.scanrelid; /* - * Clear any reference to the previously returned tuple. This doesn't - * offer any great performance benefit, but it keeps this code in sync - * with SeqNext and IndexNext. - */ - ExecClearTuple(slot); - - /* * Check if we are evaluating PlanQual for tuple of this relation. * Additional checking is not good, but no other way for now. We could * introduce new nodes for this case and handle TidScan --> NewNode @@ -123,7 +116,7 @@ TidNext(TidScanState *node) estate->es_evTuple[scanrelid - 1] != NULL) { if (estate->es_evTupleNull[scanrelid - 1]) - return slot; /* return empty slot */ + return ExecClearTuple(slot); /* * XXX shouldn't we check here to make sure tuple matches TID list? In @@ -135,7 +128,7 @@ TidNext(TidScanState *node) /* Flag for the next call that no more tuples */ estate->es_evTupleNull[scanrelid - 1] = true; - return (slot); + return slot; } /* |