Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 24 Apr 1998 19:18:38 +0200 (CEST)
From:      Remy NONNENMACHER <remy@synx.com>
To:        dwilde1@ibm.net
Cc:        don@partsnow.com, freebsd-advocacy@FreeBSD.ORG, freebsd-hardware@FreeBSD.ORG
Subject:   Re: *** Real Action Item: SPECweb
Message-ID:   <9804241908.aa27782@s3.synx.com>
In-Reply-To: <3540B774.CAB68C0C@partsnow.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On 24 Apr, Don Wilde wrote:
> I have spoken to several of my lower level salesman types at various companies,
> and they all could _immediately_ see the benefit to their companies of such a
> media circus as I'm proposing. (For those on -hardware, seek this thread in the
> -advocacy archives, I won't repeat it here)
> 
> With that in mind, I'd like to solicit comments on what the best hardware
> (single CPU) we can get to give us the highest TCPIP network throughput. If
> you'll look at http://www.SPECbench.org, you'll see the disclosures for the
> systems Novell and others have used to achieve their performance. The key to
> this thing is maximum bandwidth, disk performance, and maximum intelligence on
> the IO channels. Even if we do it by using a gigabit card set and fiber, we're
> after exposure for FreeBSD and Apache, not necessarily long-lasting benchmark
> numbers.
> 
> Once we have the system spec'd, I have no doubt I can get hardware, because it's
> very obvious that this will be an enormous bonanza for all major corporate
> sponsors.

Interesting points about the specs :


     "Cache volume: NetWare volume spanning six drives, 
     each with a single 204MB partition"

This means they used stripping and only the beginning of disks (where
the media rate is the highest). Hard to beat without a well tuned ccd
and less than 2 SCSI controllers. (note they also used 10K rpm disks).

(PS : 6x4.3Gig for creating a 1.2G filesystem !! what an efficiency !!
4% efficiency is far from a realistic system).

Other parameters tend to indicate that all goes to memory and stick
there. Is there any specification about a cool-down stage between tests
?

Network: "All nets at half duplex". Why not full dup ? they can't ?
something hidden with the BT350 ?



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hardware" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi? <9804241908.aa27782>