Date: Mon, 12 Jul 1999 19:28:50 -0500 From: "Mike Avery" <mavery@mail.otherwhen.com> To: Doug <Doug@gorean.org> Cc: kris@airnet.net, freebsd-chat@freebsd.org Subject: Re: 3C905 versus Intel Etherexpress PRO/100?! Message-ID: <199907130032.TAA23620@hostigos.otherwhen.com> In-Reply-To: <Pine.BSF.4.05.9907121539280.8174-100000@dt054n86.san.rr.com> References: <199907122230.AAA25713@p.funk.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 12 Jul 99, at 15:41, Doug wrote:
> On Tue, 13 Jul 1999, Alex Le Heux wrote:
>
> > We're talking about a 2% difference in cpu utilisation here. Is that
> > even statistically significant?
>
> Yes. I have more than one environment where every cpu cycle is
> precious, either due to long-term load or due to the need for fast
> recovery from load bursts.
The question wasn't if 2% is significant, but if the measurement was
statistically significant. Sure, it was measured. But was it a real
measurement, or a stopwatch error.
In truth, Novell's Perform3 is not the best of tests.... I've seen a
number of nets performing at 110 to 120% of their available
bandwidth....
> A fundamental design element for a server OS (as opposed to a
> desktop OS) is to always assume that *every* cpu cycle is valuable.
Okies... if you say so... still, 2% change can be caused by a driver
change. And the results are only valid on the OS where they were
tested. With the patches in place. etc, etc, etc.
Mike
======================================================================
Mike Avery MAvery@mail.otherwhen.com
(409)-842-2942 (work)
ICQ: 16241692
* Spam is for lusers who can't get business any other way *
A Randomly Selected Thought For The Day:
Confession is good for the soul, but bad for your career.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199907130032.TAA23620>