Horton

Re: Flushing large data immediately in pqcomm

David Rowley <dgrowleyml@gmail.com>

View thread View on postgr.es
From: David Rowley <dgrowleyml@gmail.com>
To: Jelte Fennema-Nio <postgres@jeltef.nl>
Cc: Melih Mutlu <m.melihmutlu@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Andres Freund <andres@anarazel.de>, Heikki Linnakangas <hlinnaka@iki.fi>, Robert Haas <robertmhaas@gmail.com>, Thomas Munro <thomas.munro@gmail.com>
Date: 2024-04-06T12:51:03Z
Lists: pgsql-hackers

Attachments

On Sat, 6 Apr 2024 at 23:17, Jelte Fennema-Nio <postgres@jeltef.nl> wrote:
> Weird that on your machines you don't see a difference. Are you sure
> you didn't make a silly mistake, like not restarting postgres or
> something?

I'm sure. I spent quite a long time between the AMD and an Apple m2 trying.

I did see the same regression as you on the smaller numbers.  I
experimented with the attached which macro'ifies internal_flush() and
pg_noinlines internal_flush_buffer.

Can you try that to see if it gets rid of the regression on the first two tests?

David

Commits

  1. Avoid needless large memcpys in libpq socket writing