Tony Caduto wrote:
http://newsvac.newsforge.com/newsvac/06/08/28/1738259.shtml
Don't know the validity of this dvd order test they did, but the article
claims Postgresql only did 120 OPM.
Seems a little fishy to me.
Now, this article really s**ks! First of all, the original contest was
specifically not only about performance. And the MySQL team did a whole
lot of dirty tricks (i.e. using memcached) to push their solution.
I am the one who has written he only PostgreSQL entry, for which I'm
still sorry and ashamed, because it performs so poorly. I just didn't
have much spare time to spend, but thought I'd send it in anyway. One of
the reasons it did not perform well was, that I simply have forgotten to
enable connection pooling.
Another entry using Perl and PostgreSQL from somebody who put a lot more
time and effort into it was disqualified because the benchmarker claimed
the files came in to late. While the author of the PostgreSQL solution
claims the wrong ZIP archive has been used.
Besides, was quite a simple database test, not too complicated queries,
very few stored procedures. No views, no triggers, no 2PC, no nothing.
So IMHO it was not even a good test of database extensibility.
So this comparison is really just... ah! forget about it! (Just please,
next time, let's put a little more effort into such a thing.)
Regards
Markus
