Horton

Re: Keep elog(ERROR) and ereport(ERROR) calls in the cold path

David Rowley <dgrowleyml@gmail.com>

View thread View on postgr.es
From: David Rowley <dgrowleyml@gmail.com>
To: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
Cc: Andres Freund <andres@anarazel.de>, PostgreSQL Developers <pgsql-hackers@lists.postgresql.org>
Date: 2020-11-23T20:36:55Z
Lists: pgsql-hackers 
On Sat, 21 Nov 2020 at 03:26, Peter Eisentraut
<peter.eisentraut@2ndquadrant.com> wrote:
>
> I did tests of elog_ereport_attribute_cold_v4.patch on an oldish Mac
> Intel laptop with pgbench scale 1 (default), and then:
>
> pgbench -S -T 60
>
> master:  tps = 8251.883229 (excluding connections establishing)
> patched: tps = 9556.836232 (excluding connections establishing)
>
> pgbench -S -T 60 -M prepared
>
> master:  tps = 14713.821837 (excluding connections establishing)
> patched: tps = 16200.066185 (excluding connections establishing)
>
> So from that this seems like an easy win.

Well, that makes it look pretty good.  If we can get 10-15% on some
machines without making things slower on any other machines, then that
seems like a good win to me.

Thanks for testing that.

David

Commits

  1. Stop gap fix for __attribute__((cold)) compiler bug in MinGW 8.1

  2. Tidy up definitions of pg_attribute_hot and pg_attribute_cold

  3. Fix unportable usage of __has_attribute

  4. Improve compiler code layout in elog/ereport ERROR calls

  5. Define pg_attribute_cold and pg_attribute_hot macros