Content-Type: text/plain;
	charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3696.120.41.1.1\))
Subject: Re: Postgresql 14 performance
From: Scott Ribe <scott_ribe@elevated-dev.com>
In-Reply-To: 
 <CAC0w7L+ezaB8tw-kh6rZC0RhG8ZwhLG_EkUErGEiUQ=f8uAC6g@mail.gmail.com>
Date: Sun, 21 Aug 2022 12:57:54 -0600
Cc: pgsql-admin <pgsql-admin@lists.postgresql.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <3FA5ECDA-A99E-4ACC-B431-54A5BC7FC4D4@elevated-dev.com>
References: 
 <CAC0w7LL+uFgPvctiHMxrVZV-DnNVg2K_wFaeZwM96XQFkjk1uA@mail.gmail.com>
 <62c9a8d8-8c0f-3c11-5cca-e05e126c25d8@gmail.com>
 <CAC0w7L+ezaB8tw-kh6rZC0RhG8ZwhLG_EkUErGEiUQ=f8uAC6g@mail.gmail.com>
To: Kenny Bachman <kenny.bachman17@gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/3FA5ECDA-A99E-4ACC-B431-54A5BC7FC4D4%40elevated-dev.com>
Precedence: bulk

> On Aug 21, 2022, at 12:55 PM, Kenny Bachman =
<kenny.bachman17@gmail.com> wrote:
>=20
> How can we explain the cache doesn't work as expected? I mean, I =
expect more performant the second or third time running because of the =
cache. However, every time I run it I get the same exec time value.

Take a look at system paging activity, and at pg IO waits. (For the =
latter, I think you have to enable tracking in postgresql.conf.) Also =
take a peek at CPU usage--if cache is doing what you expect here, then =
CPU should go to 100% on one or more cores.