Message-ID: <9d984638e80cd99a699ed85b3cebcc855d1cf8d1.camel@cybertec.at>
Subject: Re: How to set parallel_tuple_cost
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: Jeff Janes <jeff.janes@gmail.com>, "pgsql-performa."
	 <pgsql-performance@postgresql.org>
Date: Fri, 20 Dec 2019 19:42:19 +0100
In-Reply-To: 
 <CAMkU=1ziWTO_OBmFYkwhdtpc_JrM2Eroip=8vS3CNtgJs6KcoQ@mail.gmail.com>
References: 
	<CAMkU=1ziWTO_OBmFYkwhdtpc_JrM2Eroip=8vS3CNtgJs6KcoQ@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.32.5 (3.32.5-1.fc30) 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Fri, 2019-12-20 at 13:03 -0500, Jeff Janes wrote:
> The docs for parallel_tuple_cost are quite terse, as the reference section of the docs usually are:
> 
> "Sets the planner's estimate of the cost of transferring one tuple from a parallel worker process to another process. The default is 0.1."
> 
> Usually you can find more extensive discussion of such settings in informal resources like blog posts or mailing lists,
> but Googling the name I don't find much for this setting.  Is there good information out there somewhere?
> 
> If you were take the doc description literally, then the default value seems much too high, as it doesn't take
> 10x the (default) cpu_tuple_cost to transfer a tuple up from a parallel worker.  On the other hand, you probably
> don't want a query which consumes 8x the CPU resources just to finish only 5% faster (on an otherwise idle server with 8 CPUs).
> Maybe this Amdahl factor is what inspired the high default value?

Hmm.  The parameter was introduced into the discussion here:
https://www.postgresql.org/message-id/CAA4eK1L0dk9D3hARoAb84v2pGvUw4B5YoS4x18ORQREwR%2B1VCg%40mail.gmail.com
and while the name was changed from "cpu_tuple_comm_cost" to "parallel_tuple_cost"
later, the default value seems not to have been the subject of discussion.

Yours,
Laurenz Albe
-- 
Cybertec | https://www.cybertec-postgresql.com