On Mon, 2005-03-14 at 05:52, Rick Schumeyer wrote:
> Below are some PRELIMINARY results in comparing the performance of pgsql and
> mysql.
>
> These results are for a single process populating a table with 934k rows,
> and then performing some selects. I also compared the effect of creating
> indexes on some of the columns.
>
> I have not yet done any testing of transactions, multiple concurrent
> processes, etc.
>
> I did not make any changes to the default config settings. I can do
> so if someone has some suggestions.
>
> My machine is a 3.0 GHz P4 with 1 GB ram, running FC 3.
>
> I used pg 8.0.1 and mysql 5.0.2 alpha.
Why are all the tests here select count(*) tests? Surely your
application does something more interesting than counting rows...
For a more interesting test, try setting up three or four streaming
writers that write information continuously into the database, and then
run the select count(*) queries against both and see what happens.
My guess is that the table level locking of myisam tables means the
MySQL database will slow to a crawl or throw error messages, while the
postgresql system will slow down somewhat but keep right on running.