On Wed, 2006-03-29 at 02:45, Antimon wrote:
> I know this is not even a test but i thought it might be a client
> library issue, not server itself.
> Well it is not, i tried it on .net with npgsql which is a .net client
> library (Not using libpq). Results are same. Connect time does not have
> much effect by the way.
Just so you understand, there was a period of time when MySQL AB put out
benchmark after benchmark like this to "prove" that MySQL was faster
than PostgreSQL. Of course, they never compared anything moderately
complex, since they knew they'd lose.
Is it a fair comparison of say, a mainframe and a workstation to compare
the boot up times? Not really, unless your particular application is
gonna be rebooting the mainframe a lot. It's a small, narrow channel to
compare on, and most the time it doesn't mean a lot.
That's all. No personal attack meant against you, sorry if it came
across that way. It's just many pgsql folks still smart from that era,
and it seems to come back to haunt us every so often.