This is really directed at the MySQL testing team butI would be
interested in anyone who cares to comment:
When you give performance statistics, is it fair or honest to compare an
SQL database running on NT with one running on LINUX?
(or indeed ANY UNIX)
I have been an SQL database dba and/or freelance contractor in the
testing and trouble-shooting business for 10 years and in the same type
of business full-time for eight years before that and have worked with
ORACLE, SYBASE, INFORMIX and INGRES and in general they are all slower
than MySQL - some considerably so. It depends on how much functionality
is built in to the server, SYBASE the least therefore the fastest,
ORACLE the most therefore the slowest.
However, I have SYBASE 18.104.22.168 running on identical x86 boxes, one
under NT 4.00(SP3) and one under LINUX 5.02 and on average the LINUX
version runs between 2-3 times faster. Whether a particular
test/function runs 2 or 3 times faster depends on disk i/o, faster if
there is more i/o.
Hands-on-heart, honest, comments anyone?