Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!uunet!iphase!floydf
From: floydf@iphase.UUCP (Floyd Ferguson ENG)
Newsgroups: comp.unix.sysv386
Subject: Re: ESIX and MCA
Message-ID: <297@iphase.UUCP>
Date: 12 Nov 90 18:40:16 GMT
References: <5531@crash.cts.com> <1990Nov10.125312.10030@virtech.uucp>
Reply-To: floydf@iphase.UUCP (Floyd Ferguson ENG)
Organization: Interphase Corp. Dallas TX
Lines: 31

In article <1990Nov10.125312.10030@virtech.uucp> cpcahil@virtech.UUCP (Conor P. Cahill) writes:
>An article in the current issue of Personal Workstation discusses the rumour
>of non-compatible EISA stuff and states that all of PWs testing with different
>EISA parts has shown them to be compatible.

I am writing device drivers for our EISA SCSI adapter for UNIX and Novell,
so I am very familiar with the compatibility issues of 1 controller across
N systems, but not so much with the 1 system across N controller issue.

The vast majority of "compatibility" issues I have seen have been related
to the configuration process. EISA provides an extremely powerful software
based configuration mechanism which must be correctly used in order for
peripheral boards to function. Sometimes the correct usage is not the most
obvious usage.

Most of these problems arise from the increased power EISA 
provides to bus-master peripherals, and the resulting increase in
system complexity.

Overall, once the system is properly configured, I have seen no more
compatibility problems  with EISA than exist with ISA, and I suspect that
the majority of the "rumors" stem from difficulties in configuring 
products into an EISA system.

Incidentally, the EISA bus is _very_ fast, and outperforms the VME
backplane found on not a few other systems. I would be very suprised
if EISA did not displace quite a bit of VME, particularly as the Intel
based systems begin encroaching on the low- and mid-end mini's.

Floyd Ferguson
uunet!iphase!floydf