Polling For 100 mbps Connections? (Was Re: Freebsd Theme Song)

[Danial Thom]

--- Ted Mittelstaedt <tedm at toybox.placo.com>
wrote:

> 
> 
> >-----Original Message-----
> >From: Danial Thom
> [mailto:danial_thom at yahoo.com]
> >Sent: Monday, December 12, 2005 1:35 PM
> >To: Drew Tomlinson; Ted Mittelstaedt
> >Cc: Michael Vince; danial_thom at yahoo.com;
> freebsd-questions at freebsd.org;
> >Kris Kennaway
> >Subject: Re: Polling For 100 mbps Connections?
> (Was Re: Freebsd Theme
> >Song)
> >
> >
> >
> >
> >--- Drew Tomlinson <drew at mykitchentable.net>
> >wrote:
> >
> >> On 12/12/2005 8:13 AM Ted Mittelstaedt
> wrote:
> >>
> 
> > >Danial is claiming the slowness is in the
> main
> > >ram section of
> > >things, not in the ethernet driver code.
> 
> >I don't think I'm claiming that at all.
> 
> Oh, really, do tell then:
> 
> >The
> >slowness is in the latency and inefficiencies
> of
> >the scheduler and whatever other kernel
> "stuff"
> >(locking, general overheads).
> 
> Which runs in main ram...
> 
> >The entire point of
> >the tests are that the managing of the packets
> is
> >a constant, in that its the same hardware and
> >mostly the same code.
> 
> What I said...
> 
> >Now I suppose its possible
> >that the em driver could just be slower in 5.4
> >and 6.0, but the code is fundamentally the
> same,
> >so it should be a constant. So since the
> >processing of the packets is a constant, then
> if
> >you can process less packets on the same
> machine
> >the overhead of the OS must be the culprit.
> 
> And, where again does the OS do it's
> processing...
> 
> >It
> >could be the code,
> 
> Well, if it's not, then your explanation and
> everything
> you have said up to this point sure strongly
> implies it.
> 
> What's wrong Danial, now that you have actually
> had to
> think about it, now realizing you have some
> holes in
> your bitching?  Scared that I'm about ready to
> start
> punching holes in your flimsy inferences?
>


Not really, because its still a FreeBSD release,
so whether its the driver or the scheduler or the
code generated by the compiler, it still
substantially worse than FreeBSD 4.x. And MP is
SLOWER than UP for many functions. So
specifically WHAT it is doesn't change my claim
the FreeBSD 5.x and 6.x suck, at least relative
to what you started with. If you take something
and make it worse, and seem to have no ability to
figure out WHY, then you're incompetent. Its as
simple as that.

I have posted a reasonable test and results, and
there are countless complaints about performance.

I think the fact that every time someone
complains Robert Watson tells them to "wait for
6.0, or wait for 7.0" is a pretty good indication
that things aren't what the Teds and Krises
claim. 

DT

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com