Re: using custom scan nodes to prototype parallel sequential scan - Mailing list pgsql-hackers

From David Rowley
Subject Re: using custom scan nodes to prototype parallel sequential scan
Date
Msg-id CAHoyFK-4kb-OR2ukvibvi=X2GQZ=9_VsPTLrL2bE+B3r=n4a=w@mail.gmail.com
Whole thread Raw
In response to Re: using custom scan nodes to prototype parallel sequential scan  (Simon Riggs <simon@2ndQuadrant.com>)
Responses Re: using custom scan nodes to prototype parallel sequential scan  (Kouhei Kaigai <kaigai@ak.jp.nec.com>)
List pgsql-hackers
On 18 November 2014 05:19, Simon Riggs <simon@2ndquadrant.com> wrote:
On 14 November 2014 11:02, Kouhei Kaigai <kaigai@ak.jp.nec.com> wrote:

> I'd like to throw community folks a question.
> Did someone have a discussion to the challenge of aggregate push-down across
> relations join in the past? It potentially reduces number of rows to be joined.
> If we already had, I'd like to check up the discussion at that time.

Yes, I was looking at aggregate pushdown. I think it needs the same
changes to aggregates discussed upthread.


I have something that I've been working on locally. It's far from ready, but it does work in very simple cases, and shows a nice performance boost. I'll start another thread soon and copy you both in. Perhaps we can share some ideas.

Regards

David Rowley 

pgsql-hackers by date:

Previous
From: Andres Freund
Date:
Subject: Re: 9.5: Better memory accounting, towards memory-bounded HashAgg
Next
From: Robert Haas
Date:
Subject: Re: [GENERAL] Performance issue with libpq prepared queries on 9.3 and 9.4