Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!swrinde!elroy.jpl.nasa.gov!decwrl!sgi!vjs@rhyolite.wpd.sgi.com
From: vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
Newsgroups: comp.dcom.lans
Subject: Re: FDDI cards
Message-ID: <105582@sgi.sgi.com>
Date: 22 May 91 17:44:42 GMT
References: <1991May21.190618.2247@cbnewsl.att.com>
Sender: guest@sgi.sgi.com
Organization: Silicon Graphics, Inc., Mountain View, CA
Lines: 25


 
>                     Does it make sense from a through-put stand
> point to use an FDDI card in an ISA box with a 386 processor?  

It makes sense only if the 386 system (hardware and software including
operating system) needs more than ethernet speeds.  The bottleneck when
you're actually using FDDI to move data instead of cook numbers with a
bogus, toy benchmark is never the bus.  It's the softwar.  Yeah, "never" is
no doubt too strong; someone probably has a real peach of bus.  While
"real" benchmarks like ttcp are only distantly related to real
applications, the bogus benchmarks I mean are most of those that claim to
show a system can transmit over FDDI at 100Mb/s.  Most of those are blarney
designed to demonstrate that the hardware is not completely misdesigned.
People often write a little code to convince the AMD DPC-RBC chipset to
babble the same frames in a loop, and then claim their system runs at
100Mb/s.  These people try the same frauds with ethernet chips.

High performance workstations can push several MByte/sec thru TCP/FDDI.
They might feel the pinch of a bus as slow as ISA.  However, most 386
systems seem far slower.  They're not likely to notice whether they are
using 20% of the ethernet or 2% of FDDI.


Vernon Schryver,   vjs@sgi.com