From mboxrd@z Thu Jan  1 00:00:00 1970
From: caszonyi@hotmail.com
To: gcc-gnats@gcc.gnu.org
Subject: c/4175: floating point wrong code generation
Date: Thu, 30 Aug 2001 01:26:00 -0000
Message-id: <20010830081635.9070.qmail@sourceware.cygnus.com>
X-SW-Source: 2001-08/msg00725.html
List-Id: <gcc-prs.sourceware.org>

>Number:         4175
>Category:       c
>Synopsis:       floating point wrong code generation
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Thu Aug 30 01:26:01 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     caszonyi@hotmail.com
>Release:        from 2.95.2 - 2.95.3 gcc 3.0 not tested
>Organization:
>Environment:
LInux 2.2.x -2.4.x Cyrix 486 66MHz 12Meg Ram 
gcc 2.95.2.1 (patch for compilation with glibc2.2)
also tested with gcc 2.95.3
>Description:
gcc 2.95.x generate incorrect code for floating point 
intensive programssuch as spice3f5 
( http://www.ibiblio.org/pub/Linux/apps/circuits/spice3f5sfix.tar.gz )
and xanim 2.8 (i don't remember the url)

egcs 1.1.2 (egcs 2.91.66) generates correct code (the program works if 
compiled with egcs 1.1.2) 
>How-To-Repeat:
compile xanim (-O2 )and open a known mpeg file
it will say it is not an mpeg1 stream(or something like this)

or compile spice3f5 (with or without optimizations)
and try to simulate a simple circuit.
>Fix:
i don,t know any fix
>Release-Note:
>Audit-Trail:
>Unformatted: