Bug 1645 - cygwin optimization problem
Summary: cygwin optimization problem
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: fortran (show other bugs)
Version: 2.97
: P3 normal
Target Milestone: ---
Assignee: Not yet assigned to anyone
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2001-01-14 04:06 UTC by david.billinghurst
Modified: 2003-07-25 17:33 UTC (History)
3 users (show)

See Also:
Host:
Target:
Build:
Known to work:
Known to fail:
Last reconfirmed:


Attachments
20010114.f (618 bytes, application/octet-stream)
2003-05-21 15:17 UTC, david.billinghurst
Details

Note You need to log in before you can comment on or make changes to this bug.
Description david.billinghurst 2001-01-14 04:06:01 UTC
There are a number of LAPACK 3.0 failures on cygwin with current gcc snapshots.  This test case is derived from routine chgeqz.f and should a bug that causes failures in the cgg tests.

Bug is an optimization problem.  It is not present at -O0 or -O1, but is present at -O2 and higher.  The problem is in the DO loop in subroutine CHGEQZ.  The variable I is either not initialized to J (=4), or the value is lost.

This case is will be submitted to gcc-patches for inclusion in the fortran testsuite.

Release:
g77 version 2.97 20010102

Environment:
cygwin 1.1.7 & NT4 SP5

How-To-Repeat:
g77 -O2 20010114.f
./a.exe

No output == success.
Comment 1 david.billinghurst 2001-01-14 04:06:01 UTC
Fix:
http://gcc.gnu.org/ml/gcc-patches/2001-04/msg00938.html
Comment 2 Toon Moene 2001-03-03 03:21:41 UTC
State-Changed-From-To: open->analyzed
State-Changed-Why: Really an optimization problem.
    Fortran test suite item 20010116.f
Comment 3 Toon Moene 2001-03-03 11:21:41 UTC
From: toon@gcc.gnu.org
To: David.Billinghurst@riotinto.com, gcc-gnats@gcc.gnu.org, nobody@gcc.gnu.org
Cc:  
Subject: Re: fortran/1645
Date: 3 Mar 2001 11:21:41 -0000

 Synopsis: cygwin optimization problem
 
 State-Changed-From-To: open->analyzed
 State-Changed-By: toon
 State-Changed-When: Sat Mar  3 03:21:41 2001
 State-Changed-Why:
     Really an optimization problem.
     Fortran test suite item 20010116.f
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=1645&database=gcc

Comment 4 david.billinghurst 2001-03-23 01:46:07 UTC
From: "Billinghurst, David (CRTS)" <David.Billinghurst@riotinto.com>
To: "'gcc-gnats@gcc.gnu.org'" <gcc-gnats@gcc.gnu.org>,
        "Billinghurst, David (CRTS)" <David.Billinghurst@riotinto.com>,
        "'nobody@gcc.gnu.org'" <nobody@gcc.gnu.org>
Cc:  
Subject: re: fortran/1645
Date: Fri, 23 Mar 2001 01:46:07 -0000

 PR fortran/1645 look similar to PR c++/2350: constant propagation problem
 across function call
Comment 5 Mark Mitchell 2001-04-18 19:14:08 UTC
State-Changed-From-To: analyzed->closed
State-Changed-Why: Fixed.
Comment 6 Mark Mitchell 2001-04-19 02:14:08 UTC
From: mmitchel@gcc.gnu.org
To: David.Billinghurst@riotinto.com, gcc-gnats@gcc.gnu.org, nobody@gcc.gnu.org
Cc:  
Subject: Re: fortran/1645
Date: 19 Apr 2001 02:14:08 -0000

 Synopsis: cygwin optimization problem
 
 State-Changed-From-To: analyzed->closed
 State-Changed-By: mmitchel
 State-Changed-When: Wed Apr 18 19:14:08 2001
 State-Changed-Why:
     Fixed.
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=1645&database=gcc