Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!rochester!ur-tut!ur-valhalla!micropen!dave
From: dave@micropen (David F. Carlson)
Newsgroups: comp.sys.ibm.pc
Subject: 80287 bugs
Message-ID: <379@micropen>
Date: Wed, 23-Sep-87 11:27:35 EDT
Article-I.D.: micropen.379
Posted: Wed Sep 23 11:27:35 1987
Date-Received: Sat, 26-Sep-87 03:37:00 EDT
Organization: Micropen Direct Writing Systems, Pittsford, NY
Lines: 21
Keywords: math 80287 resellers

I have several AT clones running Microport.  There have been published
problems wth the FPE bug in Microport's UNIX but I am bothered by the fact
that only *some* of my machines exhibit the problem even though they are
identical *except* the source of the 80287 chips!  Could it be (like the
unscrupulous reselling of faulty 80386 chips that Intel admitted to in
Infoworld Sept. 21st ed.) that this particular reseller has sold faulty 
80287s that meet specs on math but don't handle exceptions correctly?
Obviously, my MS-DOS test suites pass as they use good data and floating
point exceptions in DOS cause a crash--like DOS is prone to do anyway.
But under UNIX, the improper FPE handling causes UNIX to hang.

Normally I'd suspect Microport, but why only on some of my machines?
And more suspiciously, why are the "problem" machines all retro 80287s
from the same vendor?  Hmmm...

Any similar experiences or confirmation of Intel selling junk parts?
-- 
David F. Carlson, Micropen, Inc.
...!{seismo}!rochester!ur-valhalla!micropen!dave

"The faster I go, the behinder I get." --Lewis Carroll