Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site wivax.UUCP
Path: utzoo!linus!decvax!wivax!dyer
From: dyer@wivax.UUCP
Newsgroups: net.micro.16k
Subject: Re: LMC Mega-Micros - (nf)
Message-ID: <18448@wivax.UUCP>
Date: Sun, 17-Jul-83 02:05:55 EDT
Article-I.D.: wivax.18448
Posted: Sun Jul 17 02:05:55 1983
Date-Received: Sun, 17-Jul-83 04:18:00 EDT
References: <2379@uiucdcs.UUCP> <855@umcp-cs.UUCP>
Organization: Wang Institute, Tyngsboro, Ma.
Lines: 21

LMC had their "MegaMicro" 16032 system at the Toronto USENIX on display
at the HCR booth (HCR having done the 4.1BSD port.)  To say that the system
was slow is an understatement.  Just about every vendor's machine had the
Sieve of Eratosthenes program on-line.  Having just come from Coherent's
demonstration of their IBM PC/XT UNIX lookalike, and seen a
"cc -O sieve.c" finish in about 30 seconds, it was quite sobering to
see the same operation take almost 90 seconds on the LMC with only myself
logged in!  Otherwise, it was a faithful VAX/4.1BSD clone; just too slow.

The LMC technical representative blamed the system's speed problems
to a Winchester with an 85ms seek time, and secondarily to the fact
that the 16032 was running at 4mhz rather than 6 or 10.  Apparently,
the slow Winchester was all that was available in time for the show.
Right now, I am willing to believe him.  I think that the LMC is certainly
something to keep watching as it matures.  With a faster disk and CPU,
it should be quite nice.

By the way, NSC also had its own 4.1 system on display.  It was much
faster, probably for the reasons above.  Unfortunately, it was described
by a technical rep as having a proprietary bus; the LMC uses the
Multibus for non-memory transfers.