Performance 4.x vs. 6.x
chrcoluk at gmail.com
Mon Oct 16 18:10:50 UTC 2006
On 16/10/06, Mark Kirkwood <markir at paradise.net.nz> wrote:
> Mark Linimon wrote:
> > On Sun, Oct 15, 2006 at 02:01:08PM -0400, Michael Butler wrote:
> >> For everyone's benefit then, please feel free to submit your patches
> >> along with your technical analysis.
> > I think his best bet is a fork, instead. Then he can tell all the people
> > that volunteer to work on _his_ project exactly what to do, and see how
> > far he gets with that approach.
> He might have got further by volunteering to create and supply profiles
> for those specific workloads that were faster in 4.x than 6.x on UP
> machinery etc... i.e. help make 6.x better rather than discourage the
> development team (whose efforts are much appreciated by the rest of us
> that are happily using 6.x...)
I recently ordered some servers from a datacentre on lease, specs were
UP p4 2.8ghz gig of ddr2 ram and sata hd, intel lan card. None of the
servers would boot in freebsd 6.x, they booted in freebsd 4.x but
needed a pata controller, they only worked properly in freebsd 5.x.
It seems their are 2 major problems with freebsd at the moment (1) is
the hardware support is still way behind both linux and windows and
its very frustrating in the amount of datacentres that dont support
freebsd. and (2) the uniprocessor performance remains below par.
The freebsd team it would help to realise not everyone can pick and
choose their hardware and not everyone has the budget for state of the
art hardware, certianly if you go around browsing datacentre websites
for dedicated servers the dominant spec is x86 single processor, dual
cpu is growing but still not dominent and I think 32bit UP wont be
dead for at least half a decade. freebsd 4.x their is nothing to
dispute, its leaner and smoother on pretty much every UP setup and I
think it would do freebsd's reputation some good if a 4.12 was to come
just my 2 pence worth.
More information about the freebsd-performance