On 5/22/09 2:36 PM, Tom Lane wrote:
> andrew@dunslane.net writes:
>> Wouldn't he just need to rerun the tests with default_stats_target set to
>> the old value? I presume he has actually done this already in order to
>> come to the conclusion he did about the cause of the regression.
>
> Yeah, he did, so we know it's slower that way. But exactly *why* it's
> slower is not proven. It could be an artifact rather than something
> we really ought to react to.
It appears (right now) to be an artifact.
The drop in performance happens with queries which are called using C
stored procedures exclusively. It doesn't show up on other benchmarks
which call similar queries directly.
Jignesh and I will be testing some stuff next week to get a better idea
of what exactly makes the drop happen, but for not this appears to be a
corner case.
--
Josh Berkus
PostgreSQL Experts Inc.
www.pgexperts.com