Benchmark: NetBSD 2.0 beats FreeBSD 5.3
Martin P. Hellwig
mhellwig at xs4all.nl
Thu Jan 6 05:41:09 PST 2005
Nguyen Tam Chinh wrote:
> Please don't treat this seriously. Benchmarks are just benchmarks. But
> the benchmarks and comparison, widespreaded through sites like
> slashdot or osnews, sometimes affect the interest and view point of
> some new and potential users.
> May be we should do some full benchmarks as an answer and to review
> the true status of our 5.x, 4.x and others?
<cut>
True, I don't take it very seriously but it does say something, like all
things you measure it shoul be put into perspective.
I have already contacted the author of the benchmark, in short I've
asked him if he could do the test with latest stable DragonFly too.
I don't see the logic in testing FBSD4 as this is a "Legacy" branch,
just as stated on www.FreeBSD.org
But perhaps the test should be redone on multiple popular role based
hardware configuration, including OpenBSD, DragonFly and any other OS
you wish to test.
Roles based in the meaning of benchmarking typical firewall, webserver,
database, file and print servers roles.
But there other things that must be benchmarked too if you want a near
objective view, like stability, hardware support, security and design.
Personally I don't give a * about performance as long as it doesn't hold
me back, but what I do find irritating is that when there is a security
issue in a port I should have to rebuild all my ports because of some
libthreading issue.
Now when talking about a few home boxes this is not a problem, but in a
productivity environment with dozen machines having all its specific
adaption on configuration and ports, thing get to start ugly. Of course
this problem is not a real challenge it is just an inconvenience if you
did not expected it.
Just like installing a MS patch on the server and finding out that all
shared HP printer don't work anymore ;-)
Aah well keeps me off the street and out of trouble.
--
mph
More information about the freebsd-current
mailing list