Your TPC-H test

From: Josh Berkus <josh(at)postgresql(dot)org>
To: zgomez(at)ac(dot)upc(dot)es
Cc: pgsql-advocacy(at)postgresql(dot)org
Subject: Your TPC-H test
Date: 2004-04-21 17:41:22
Message-ID: 200404211041.22711.josh@postgresql.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-advocacy

Mr. Gomez,

Hi! I just read your TPC-H test restult posting. Thank you for taking the
time to do such a significant job reporting the results.

Unfortunately, your setup for the test could use some help from us. Mainly:

1) For any reasonable results, PostgreSQL must be tuned to non-default
parameters. See in particular:
http://www.varlena.com/varlena/GeneralBits/Tidbits/perf.html
We do realize that it imposes a burden on such tests to need manual
configuration of the database server. However, that's a problem we're not
going to solve for PostgreSQL this week, so please let us help you. Your
load time of days for the data in particular indicates a tuning problem; when
I ran TPC-H, data loaded in under an hour.

2) Testing MyISAM tables on MySQL without Foriegn keys or transactions
invalidates your test; these are requirements of the TPC and not optional.
You should test on an installation of MySQL which supports these features.

3) I would also like to see you test the current CVS for PostgreSQL. We've
been working with the OSDL to improve our TPC-H test results, and the queries
which timed out now return correct results.

--
Josh Berkus
PostgreSQL Advocacy Volunteer
http://advocacy.postgresql.org
San Francisco

Browse pgsql-advocacy by date

  From Date Subject
Next Message Josh Berkus 2004-04-21 17:47:03 Re: [PERFORM] MySQL vs PG TPC-H benchmarks
Previous Message Josh Berkus 2004-04-21 17:31:31 Fwd: Re: [PERFORM] MySQL vs PG TPC-H benchmarks