v1-0002-Reverse-loop-to-match-counterpart-and-optimize-fo.patch
application/octet-stream
Filename: v1-0002-Reverse-loop-to-match-counterpart-and-optimize-fo.patch
Type: application/octet-stream
Part: 2
Patch
Same data as JSON:
GET /api/v1/attachments/:id/patch
the parsed metadata as JSON — format, series position, per-file stats; never the diff bytes.
API reference →
Format: format-patch
Series: patch v1-0002
Subject: Reverse loop to match counterpart and optimize for CPU prefetching.
| File | + | − |
|---|---|---|
| src/backend/access/heap/vacuumlazy.c | 8 | 3 |
From dff9c509856223db78178a0f6fdd1a8b3f0affd7 Mon Sep 17 00:00:00 2001
From: Greg Burd <greg@burd.me>
Date: Tue, 6 May 2025 11:14:19 -0400
Subject: [PATCH v1 2/3] Reverse loop to match counterpart and optimize for CPU
prefetching.
Small change to ensure that the loop in heap_page_is_all_visible()
takes advantage of the CPU prefetching optimization found in its
counterpart heap_page_prune_and_freeze() and is generally more in
sync.
Author: Greg Burd <greg@burd.me>
---
src/backend/access/heap/vacuumlazy.c | 11 ++++++++---
1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/src/backend/access/heap/vacuumlazy.c b/src/backend/access/heap/vacuumlazy.c
index e450911424f..14cf880340b 100644
--- a/src/backend/access/heap/vacuumlazy.c
+++ b/src/backend/access/heap/vacuumlazy.c
@@ -3625,10 +3625,15 @@ heap_page_is_all_visible(LVRelState *vacrel, Buffer buf,
*visibility_cutoff_xid = InvalidTransactionId;
*all_frozen = true;
+ /*
+ * Processing the items in reverse order (and thus the tuples in
+ * increasing order) increases prefetching efficiency significantly /
+ * decreases the number of cache misses.
+ */
maxoff = PageGetMaxOffsetNumber(page);
- for (offnum = FirstOffsetNumber;
- offnum <= maxoff && all_visible;
- offnum = OffsetNumberNext(offnum))
+ for (offnum = maxoff;
+ offnum >= FirstOffsetNumber && all_visible;
+ offnum = OffsetNumberPrev(offnum))
{
ItemId itemid;
HeapTupleData tuple;
--
2.39.5 (Apple Git-154)