This is part of The Pile, a partial archive of some open source mailing lists and newsgroups.
A piece of flamebait in progress: To: "Dbi-Users" <dbi-users@isc.org> Subject: Has anyone done any database benchmarking? -------- http://slashdot.org/article.pl?sid=00/08/14/2128237&mode=nested http://www.phpbuilder.com/columns/tim20000705.php3?page=1 === Subject: Re: [HACKERS] benchmarks - anyone can make em From: Jeff MacDonald <jeff@pgsql.com> Date: Fri, 18 Aug 2000 13:17:12 -0300 (ADT) On Fri, 18 Aug 2000, Jeff MacDonald wrote: > I have a friend who works at IBM on DB2, i showed > her "the benchmark" just for kicks.. (we teach each > other over who's is better etc.. all good fun) > > anyway she sends me back this article about how DB2 > blows the lid off of TPC-C.. just proof that any > benchmark can mean anything.. > > at the same time i talked to a friend, who's a strong > postgres proponent, he was curious as to what size > the databases the postgres benchmarks were done on.. > > ned maybe you can field this second question. course i could definatly include the url http://www.zdnet.com/eweek/stories/general/0,11011,2599679,00.html === Subject: Re: [HACKERS] benchmarks - anyone can make em From: Don Baccus <dhogaza@pacifier.com> Date: Fri, 18 Aug 2000 11:14:08 -0700 At 01:17 PM 8/18/00 -0300, Jeff MacDonald wrote: >course i could definatly include the url > >http://www.zdnet.com/eweek/stories/general/0,11011,2599679,00.html OK, so they weren't running on a large mainframe. They were running on a cluster of 32 servers. I should HOPE they'd beat Postgres on a single server! This is FAR more "apples to oranges" than comparing engines on the same machine, same Linux release ... === Subject: Re: [HACKERS] benchmarks - anyone can make em From: Don Baccus <dhogaza@pacifier.com> Date: Fri, 18 Aug 2000 11:11:37 -0700 At 01:11 PM 8/18/00 -0300, Jeff MacDonald wrote: >I have a friend who works at IBM on DB2, i showed >her "the benchmark" just for kicks.. (we teach each >other over who's is better etc.. all good fun) > >anyway she sends me back this article about how DB2 >blows the lid off of TPC-C.. just proof that any >benchmark can mean anything.. Sure they do, IBM still makes mainframes far more powerful than puny dual-processor Linux boxes... The "proprietary 1", mySQL and Postgres tests were all run on the same operating system/hardware combination, at least. === Subject: Re: [HACKERS] benchmarks - anyone can make em From: Ned Lilly <ned@greatbridge.com> Date: Fri, 18 Aug 2000 14:58:55 -0400 What size? According to the engineer who ran the test: > We set it up for 100 users. There is a default loadset for that > number of users, which loads up the tables to various levels. It's described > in the docs. Some tables had 100,000 rows. More info on the AS3AP test can be found at http://www.benchmarkresources.com/handbook/5-3.html More info on the TPC-C at http://www.tpc.org/faq_TPCC.html. ===