Re: Is there a meaningful benchmark?

I've had lots of experience fighting with the LOAD FROM command with an
old application using an old version of Informix. When the data to be
loaded is large enough, Informix (dbaccess) hits a wall. Tracking it
down with tools generally shows that zillions of locks are accumulating.
The system thrashes.

So far I haven't found this problem with Postgres.

Incidentally, in case people are interested in hearing the follow-up to
my original message, my home-cooked benchmark with lots of processes
doing tons of updates seemed to initially show MyISAM winning over
Postgres.

However I spent some time reading what I could from the manual on
performance improvement and realized that commits might be an issue. I
moved a commit() call out of the main loop, and bang! Postgres showed a
substantial performance advantage over MyISAM.

-Will

-----Original Message-----
From: Scott Marlowe [mailto:scott(dot)marlowe(at)gmail(dot)com]
Sent: 27 March 2009 02:25
To: Scott Ribe
Cc: Will Rutherdale (rutherw); pgsql-general(at)postgresql(dot)org
Subject: Re: [GENERAL] Is there a meaningful benchmark?

To be fare, some database apps have a few rows they update in a near
continuous stream, and they row lock. These databases are often
better served by db2 or some other row locking database than pgsql
where you may or may not have problems with bloating.