Re: bad estimation together with large work_mem generates terrible slow hash joins

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: Tomas Vondra <tv(at)fuzzy(dot)cz>
Cc: "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: bad estimation together with large work_mem generates terrible slow hash joins
Date: 2014-09-12 12:28:49
Message-ID: CA+TgmoYt9gtGc+-oCprBV7OKwAvTwK24ZF_G7ZHrFq6VnsSAwg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thu, Sep 11, 2014 at 5:57 PM, Tomas Vondra <tv(at)fuzzy(dot)cz> wrote:
> Attached is the patch split as suggested:
>
> (a) hashjoin-nbuckets-v14a-size.patch
>
> * NTUP_PER_BUCKET=1
> * counting buckets towards work_mem
> * changes in ExecChooseHashTableSize (due to the other changes)

OK, I'm going to work on this one now. I have some ideas about how to
simplify this a bit and will post an update in a few hours once I see
how that pans out.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Amit Kapila 2014-09-12 12:42:56 Re: pg_basebackup vs. Windows and tablespaces
Previous Message Michael Paquier 2014-09-12 12:16:42 Re: Incorrect initialization of sentPtr in walsender.c