Horton

Re: Introduce new multi insert Table AM and improve performance of various SQL commands with it for Heap AM

Daniil Davydov <3danissimo@gmail.com>

View thread
From: Daniil Davydov <3danissimo@gmail.com>
To: Jingtang Zhang <mrdrivingduck@gmail.com>
Cc: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>, pgsql-hackers@lists.postgresql.org, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2025-03-17T04:49:45Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. libpq: Fix some issues in TAP tests for service files

  2. Multiple revisions to the GROUP BY reordering tests

  3. Explore alternative orderings of group-by pathkeys during optimization.

Hi,
Recently I took more careful measurements of the performance. I
compared three branches with each other: HEAD, Patched and Patched
with tuplestore.
Here are the results :

1)
Test case : matview creation test attached in the email from Jingtang Zhang.
10 measurements for each branch.
Result in wall clock execution time :

HEAD
30.532 +- 0.59 seconds elapsed
Patched
20.454 +- 0.114 seconds elapsed
Patched with tuplestore
19.653 +- 0.111 seconds elapsed

2)
-- init.sql
drop table test_insert;
vacuum;
checkpoint;
create table test_insert(i int, f float);

-- iowrite.sql
insert into test_insert select g, (g % 100) / 100.0 from
generate_series(1, 1000000) as g;

Test case :
pgbench -f iowrite.sql -n -j 4 -c 10 -T 40
5 measurements for each branch.
Result in tps :

HEAD
1.025 +- 0.009
Patched
2.923 +- 0.032
Patched with tuplestore
2.987 +- 0.011

P.S.
I cannot find a commitfest entry for this patch. Should we add it there?

--
Best regards,
Daniil Davydov