Horton

Re: Use read streams in pg_visibility

Nazir Bilal Yavuz <byavuz81@gmail.com>

View thread View on postgr.es
From: Nazir Bilal Yavuz <byavuz81@gmail.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2024-08-19T11:01:00Z
Lists: pgsql-hackers

Attachments

Hi,

On Mon, 19 Aug 2024 at 09:30, Michael Paquier <michael@paquier.xyz> wrote:
>
> On Tue, Aug 13, 2024 at 03:22:27PM +0300, Nazir Bilal Yavuz wrote:
> > Hi,
> >
> > I am working on using the read stream in pg_visibility. There are two
> > places to use it:
> >
> > 1- collect_visibility_data()
> >
> > This one is straightforward. I created a read stream object if
> > 'include_pd' is true because read I/O is done when 'include_pd' is
> > true. There is ~4% timing improvement with this change. I started the
> > server with the default settings and created a 6 GB table. Then run
> > 100 times pg_visibility() by clearing the OS cache between each run.
> > ----------
>
> Mind sharing a script for reproducibility?  Except for the drop_caches
> part, of course..

Sure, the script is attached.

-- 
Regards,
Nazir Bilal Yavuz
Microsoft

Commits

  1. Optimize pg_visibility with read streams.

  2. Revert "Optimize pg_visibility with read streams."

  3. Fix stack variable scope from previous commit.

  4. Add block_range_read_stream_cb(), to deduplicate code.