From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
cc: Alvaro Herrera <alvherre@alvh.no-ip.org>,
        Robert Haas <robertmhaas@gmail.com>,
        Thomas Munro <thomas.munro@enterprisedb.com>,
        "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Subject: Re: pgsql: Add parallel-aware hash joins.
In-reply-to: <20180124204359.ognamsfdqbrfbazd@alap3.anarazel.de>
References: <20180124182018.jx2h4wp6twqo7xix@alap3.anarazel.de>
 <20180124201826.g6vhuewpnycpt3ta@alvherre.pgsql>
 <20180124204359.ognamsfdqbrfbazd@alap3.anarazel.de>
Comments: In-reply-to Andres Freund <andres@anarazel.de>
	message dated "Wed, 24 Jan 2018 12:43:59 -0800"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <19264.1516827496.1@sss.pgh.pa.us>
Date: Wed, 24 Jan 2018 15:58:16 -0500
Message-ID: <19265.1516827496@sss.pgh.pa.us>
Precedence: bulk

Andres Freund <andres@anarazel.de> writes:
> On 2018-01-24 17:18:26 -0300, Alvaro Herrera wrote:
>> Yeah, I proposed this a decade ago but never had the wits to write the
>> code.

> It shouldn't be too hard, right? Leaving defining the file format,
> parsing it, creating the new schedule with depencencies and adapting
> tests aside (hah), it mostly seems a relatively simple graph ordering /
> topological sort problem, right?

Yeah.  We already have topo sort code in pg_dump, maybe we could push that
into someplace like src/common or src/fe_utils?  Although pg_dump hasn't
got any need for edge weights, so maybe sharing code isn't worth it.

We could flush the existing schedule files and use a simple format like
	testname: list of earlier tests it depends on
(I guess there would be more properties than just the dependencies,
but still not hard to parse.)

> If we keep the timings from an earlier
> run somwhere, we can use the timing of runs as edge weights, making the
> schedule better.

I think we could just use constant values hand-coded in the schedule file.
It might occasionally be worth updating them, but realistically it's not
going to matter that they be very accurate.  Probably weights like 1, 2,
and 3 would be plenty ;-)

			regards, tom lane