From owner-freebsd-hackers  Fri Jul 18 08:37:04 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id IAA28381
          for hackers-outgoing; Fri, 18 Jul 1997 08:37:04 -0700 (PDT)
Received: from beowulf.utmb.edu (beowulf.utmb.edu [129.109.59.83])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id IAA28375
          for <hackers@freebsd.org>; Fri, 18 Jul 1997 08:36:59 -0700 (PDT)
Received: (from bdodson@localhost)
	by beowulf.utmb.edu (8.8.5/8.8.5) id KAA13685
	for hackers@freebsd.org; Fri, 18 Jul 1997 10:34:30 -0500 (CDT)
Date: Fri, 18 Jul 1997 10:34:30 -0500 (CDT)
From: "M. L. Dodson" <bdodson@beowulf.utmb.edu>
Message-Id: <199707181534.KAA13685@beowulf.utmb.edu>
To: hackers@freebsd.org
Subject: PentiumPro/II floating point "errata" impact
X-Sun-Charset: US-ASCII
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Hi,
Does anyone know if the floating point bug in PentiumPros and PentiumIIs
is likely to be manifest in libm?  I'm talking about the unflagged
overflow when converting from floating point to integer (FIST? or some 
such).  If it is, does anyone know whether this can be handled by code 
generation in the gcc backend?  (The Intel web site talks about software 
workarounds.)

I use g77 to compile chemical kinetics simulation code, just got a new 
PentiumPro workstation, and would like to know how likely this is to 
affect my results.  Would not be too happy to have to compile it all 
to assembler, and grep out the the instruction codes :)

TIA,
Bud Dodson

--
M. L. Dodson                                bdodson@scms.utmb.edu
409-772-2178                                FAX: 409-772-1790