*BSD News Article 39947


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!agate!howland.reston.ans.net!swiss.ans.net!news.dfn.de!rrz.uni-koeln.de!RRZ.Uni-Koeln.DE!RRZ.Uni-Koeln.DE!news
From: se@fileserv1.MI.Uni-Koeln.DE (Stefan Esser)
Newsgroups: comp.os.386bsd.questions
Subject: Re: AMD DX2/80 w/ NetBSD?
Date: 27 Dec 1994 16:13:37 GMT
Organization: Institute of Nuclear Physics, University of Cologne, Germany
Lines: 50
Distribution: world
Message-ID: <3dpefhINN3u8a@rs1.rrz.Uni-Koeln.DE>
References: <3d71n3$2hp@solaris.cc.vt.edu> <MICHAELV.94Dec21161813@MindBender.HeadCandy.com>
NNTP-Posting-Host: fileserv1.mi.uni-koeln.de
Keywords: PCI

In article <MICHAELV.94Dec21161813@MindBender.HeadCandy.com>, michaelv@MindBender.HeadCandy.com (Michael L. VanLoon) writes:
|> In article <3da3m5$6bb@news.cloud9.net> tls@cloud9.net (Thor Lancelot Simon) writes:
|> 
|>    If you find a 486 PCI motherboard that can clock the CPU at 40MHz,
|>    let me know.
|> 
|> One problem is that the PCI spec specifically states that the
|> supported speeds for the PCI bus are 25MHz to 33MHz.  The PCI bus will
|> not run at 40MHz.  So, you would have to uncouple the CPU clock from
|> the PCI bus clock.  I don't see any real problem with doing this
|> (after all, the PCI bus was designed to be uncoupled from the CPU),
|> but in the interest of cost, it appears that nobody had designed a
|> motherboard this way.  As long as the motherboard designer keeps the
|> bus and CPU clocks coupled, you won't be able to go over 33MHz with
|> PCI.

There was a review of PCI + VLB motherboards 
in c't magazine, issue 8'94, if I remember right.
(Since not everybody has access to that particular
computer magazine, written in german, to make things
worse, it might be a good idea to post short summary
of their results, as far as they might be useful for
*BSD users ... Guess I'll do that next time, they
print an interesting review ...) 

There was one board that had the PCI bus running
asynchronously, resulting in PCI throughput nearly
the same as good old ISA ... :(

I.e. the synchronisation overhead seems to be 
anything but negligable (surprising, since most
PCI accesses will take a number of 30ns cycles, 
and synchronisation should require at most one 
cycle).

Seems that the only currently available 486 with 
PCI clock uncoupled from the CPU (running at 40MHz
external clock) is no faster than a well designed 
standard PCI motherboard with a 33MHz (ext.) CPU.

And then, I recently saw first offers of complete 
systems with AMD DX4/100 ...

STefan
-- 
 Stefan Esser				Internet:	<se@ZPR.Uni-Koeln.DE>
 Zentrum fuer Paralleles Rechnen	Tel:		+49 221 4706010
 Universitaet zu Koeln			FAX:		+49 221 4705160
 Weyertal 80
 50931 Koeln