pg.ddx.io pgsql-sql@postgresql.org mailing list archivehelp / color / mirror / Atom feed
Benchmark 2+ messages / 2 participants [nested] [flat]
* Benchmark @ 2000-06-02 18:35 Nikolaj Lundsgaard <Nikolaj@konfus.dk> 0 siblings, 1 reply; 2+ messages in thread From: Nikolaj Lundsgaard @ 2000-06-02 18:35 UTC (permalink / raw) To: pgsql-sql Is there anywhere I can find some benchmark test for postgres. The reason is that I want to use it for a site that could have several simultanous users querying the database and I want to be sure that it is stable (and fast) enough to handle that kind of (ab)use. Also which is fastest ? 1. A query (using index of course) for a text in the database. or 2. Reading from a file on the harddisk. If the answer is 1 will the answer still be 1 if I do it several times (100.000+ a day) Regards Nikolaj ^ permalink raw reply [nested|flat] 2+ messages in thread
* Re: Benchmark @ 2000-06-02 20:58 Jan Wieck <JanWieck@t-online.de> parent: Nikolaj Lundsgaard <Nikolaj@konfus.dk> 0 siblings, 0 replies; 2+ messages in thread From: JanWieck@t-online.de @ 2000-06-02 20:58 UTC (permalink / raw) To: Nikolaj Lundsgaard <Nikolaj@konfus.dk>; +Cc: pgsql-sql Nikolaj Lundsgaard wrote: > Is there anywhere I can find some benchmark test for postgres. The reason is that I want to use it for a site that could have several simultanous users querying the database and I want to be sure that it is stable (and fast) enough to handle that kind of (ab)use. There is actually no (DB-independant) benchmark suite publicly available. > > Also which is fastest ? > > 1. A query (using index of course) for a text in the database. > > or > > 2. Reading from a file on the harddisk. > > > If the answer is 1 will the answer still be 1 if I do it several times (100.000+ a day) Since any DB access boils down somehow to file access, it could never be 1 (as long as your solution doesn't need to sequentially read big files to find small pieces of information). The question you have to ask is, what would happen in a concurrent environment? If there's at least one writing process, must the entire information be blocked from reading? Postgres uses MVCC (Multi Version Concurrency Control) for it. So no reader will ever be blocked by a writer - it simply sees a snapshot of the entire data as it was when he executed the first DML statement. If you have total static data, that never changes during uptime of your application, use flat files or some higher level file interface like GDB or MySQL. If you have data in motion, use a database. Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #================================================== JanWieck@Yahoo.com # ^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2000-06-02 20:58 UTC | newest] Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2000-06-02 18:35 Benchmark Nikolaj Lundsgaard <Nikolaj@konfus.dk> 2000-06-02 20:58 ` Jan Wieck <JanWieck@t-online.de>
This inbox is served by DDX for PostgreSQL; see mirroring instructions for how to clone and mirror all data and code used for this inbox