public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron
@ 2007-04-25 10:06 burnus at gcc dot gnu dot org
2007-04-25 12:28 ` [Bug middle-end/31697] " burnus at gcc dot gnu dot org
` (7 more replies)
0 siblings, 8 replies; 9+ messages in thread
From: burnus at gcc dot gnu dot org @ 2007-04-25 10:06 UTC (permalink / raw)
To: gcc-bugs
This is with the Polyhedron aermod.f90 test case
(http://www.polyhedron.co.uk/pb05/polyhedron_benchmark_suite.html).
The program crashes (in << 1 sec) with
gfortran -funroll-loops -march=opteron -O2 -g aermod.f90
It does not crash:
- without: -march=opteron
- without: -funroll-loops
- with: -O1
It was working on 2007-04-23 (should be r124055) and started to fail since
2007-04-24 (should be r124093).
gdb shows:
Program received signal SIGSEGV, Segmentation fault.
#0 resini_ () at aermod.f90:23766
#1 0x00000000004c7357 in MAIN__ () at aermod.f90:2189
And valgrind:
==6590== Invalid write of size 4
==6590== at 0x4066D1: resini_ (aermod.f90:23766)
==6590== by 0x4C8BD6: MAIN__ (aermod.f90:2189)
==6590== by 0x4C95AB: main (fmain.c:22)
==6590== Address 0x1BE65FD0 is not stack'd, malloc'd or (recently) free'd
23757 DO K = 1 , NUMGRP
23758 DO J = 1 , NUMREC
23759 AVEVAL(J,K,L,M) = 0.0
23760 DO I = 1 , NHIVAL
23761 HIVALU(J,I,K,L,M) = 0.0
23762 NHIDAT(J,I,K,L,M) = 0
23763 HCLMSG(J,I,K,L,M) = ' '
23764 HMAX(I,K,L,M) = 0.0
23765 HMDATE(I,K,L,M) = 0
23766 HMLOC(I,K,L,M) = 0
23767 HMCLM(I,K,L,M) = ' '
23768 ENDDO
23769 ENDDO
--
Summary: [Regression 4.3] Crash of gen. prog. when using -
funroll-loops -march=opteron
Product: gcc
Version: 4.3.0
Status: UNCONFIRMED
Keywords: wrong-code
Severity: normal
Priority: P3
Component: middle-end
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: burnus at gcc dot gnu dot org
GCC target triplet: x86_64-unknown-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
@ 2007-04-25 12:28 ` burnus at gcc dot gnu dot org
2007-04-25 14:28 ` [Bug middle-end/31697] [4.3 Regression] " rguenth at gcc dot gnu dot org
` (6 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: burnus at gcc dot gnu dot org @ 2007-04-25 12:28 UTC (permalink / raw)
To: gcc-bugs
------- Comment #1 from burnus at gcc dot gnu dot org 2007-04-25 13:28 -------
Might be related to the similar PR 31699.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [4.3 Regression] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
2007-04-25 12:28 ` [Bug middle-end/31697] " burnus at gcc dot gnu dot org
@ 2007-04-25 14:28 ` rguenth at gcc dot gnu dot org
2007-04-26 9:48 ` jb at gcc dot gnu dot org
` (5 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2007-04-25 14:28 UTC (permalink / raw)
To: gcc-bugs
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|[Regression 4.3] Crash of |[4.3 Regression] Crash of
|gen. prog. when using - |gen. prog. when using -
|funroll-loops -march=opteron|funroll-loops -march=opteron
Target Milestone|--- |4.3.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [4.3 Regression] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
2007-04-25 12:28 ` [Bug middle-end/31697] " burnus at gcc dot gnu dot org
2007-04-25 14:28 ` [Bug middle-end/31697] [4.3 Regression] " rguenth at gcc dot gnu dot org
@ 2007-04-26 9:48 ` jb at gcc dot gnu dot org
2007-04-26 10:23 ` ubizjak at gmail dot com
` (4 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: jb at gcc dot gnu dot org @ 2007-04-26 9:48 UTC (permalink / raw)
To: gcc-bugs
------- Comment #2 from jb at gcc dot gnu dot org 2007-04-26 10:47 -------
FYI, I am unable to confirm this on i686-pc-linux-gnu. aermod seems to work no
matter what switches I throw at the compiler.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [4.3 Regression] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
` (2 preceding siblings ...)
2007-04-26 9:48 ` jb at gcc dot gnu dot org
@ 2007-04-26 10:23 ` ubizjak at gmail dot com
2007-04-26 13:53 ` burnus at gcc dot gnu dot org
` (3 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: ubizjak at gmail dot com @ 2007-04-26 10:23 UTC (permalink / raw)
To: gcc-bugs
------- Comment #3 from ubizjak at gmail dot com 2007-04-26 11:23 -------
(In reply to comment #1)
> Might be related to the similar PR 31699.
No, because PR 31699 is triggered by -ftree-vectorize.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [4.3 Regression] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
` (3 preceding siblings ...)
2007-04-26 10:23 ` ubizjak at gmail dot com
@ 2007-04-26 13:53 ` burnus at gcc dot gnu dot org
2007-04-27 4:04 ` bangerth at dealii dot org
` (2 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: burnus at gcc dot gnu dot org @ 2007-04-26 13:53 UTC (permalink / raw)
To: gcc-bugs
------- Comment #4 from burnus at gcc dot gnu dot org 2007-04-26 14:53 -------
> FYI, I am unable to confirm this on i686-pc-linux-gnu.
> aermod seems to work no matter what switches I throw at the compiler.
--
> > Might be related to the similar PR 31699.
> No, because PR 31699 is triggered by -ftree-vectorize.
I'm really interested why this happens only here and not Janne's or SUSE's
system. Anyway, I reduced the test case to:
37 aermod2_RESINI.f90 - reduced RESINI subroutine
14 aermod2_prog.f90 - reduced PROGRAM aermod
720 aermod2_main.f90 - the full MAIN module
15410 aermod2.f90 - the remaining part of aermod
Compared with initially 26993 lines that's not too bad, but still way too big
:-(
If I compile everything with the previous options except of aermod2_RESINI.f90
it does not crash. Removing the SAVE in RESINI also gets rid of the crash. Any
idea how to proceed except of trying to reduce the 15410 lines?
Can anyone else reproduce this (maybe it is only visible using valgrind on
other systems)? My "reduced" version is available from:
http://physik.fu-berlin.de/~tburnus/tmp/aermod-debug.tar.gz
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [4.3 Regression] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
` (4 preceding siblings ...)
2007-04-26 13:53 ` burnus at gcc dot gnu dot org
@ 2007-04-27 4:04 ` bangerth at dealii dot org
2007-04-27 8:49 ` burnus at gcc dot gnu dot org
2007-05-03 13:16 ` burnus at gcc dot gnu dot org
7 siblings, 0 replies; 9+ messages in thread
From: bangerth at dealii dot org @ 2007-04-27 4:04 UTC (permalink / raw)
To: gcc-bugs
------- Comment #5 from bangerth at dealii dot org 2007-04-27 05:04 -------
Can you post disassembled code from gdb around the location where it
crashes?
--
bangerth at dealii dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |bangerth at dealii dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [4.3 Regression] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
` (5 preceding siblings ...)
2007-04-27 4:04 ` bangerth at dealii dot org
@ 2007-04-27 8:49 ` burnus at gcc dot gnu dot org
2007-05-03 13:16 ` burnus at gcc dot gnu dot org
7 siblings, 0 replies; 9+ messages in thread
From: burnus at gcc dot gnu dot org @ 2007-04-27 8:49 UTC (permalink / raw)
To: gcc-bugs
------- Comment #6 from burnus at gcc dot gnu dot org 2007-04-27 09:49 -------
> No, because PR 31699 is triggered by -ftree-vectorize.
Well, at least with today's GCC plus the patch in PR 31699, it does not crash
anymore. I will do some more checks to make sure I don't miss something.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug middle-end/31697] [4.3 Regression] Crash of gen. prog. when using -funroll-loops -march=opteron
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
` (6 preceding siblings ...)
2007-04-27 8:49 ` burnus at gcc dot gnu dot org
@ 2007-05-03 13:16 ` burnus at gcc dot gnu dot org
7 siblings, 0 replies; 9+ messages in thread
From: burnus at gcc dot gnu dot org @ 2007-05-03 13:16 UTC (permalink / raw)
To: gcc-bugs
------- Comment #7 from burnus at gcc dot gnu dot org 2007-05-03 14:16 -------
Mark as duplicate as the fix in PR 31699 fixes this bug.
*** This bug has been marked as a duplicate of 31699 ***
--
burnus at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution| |DUPLICATE
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31697
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2007-05-03 13:16 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-04-25 10:06 [Bug middle-end/31697] New: [Regression 4.3] Crash of gen. prog. when using -funroll-loops -march=opteron burnus at gcc dot gnu dot org
2007-04-25 12:28 ` [Bug middle-end/31697] " burnus at gcc dot gnu dot org
2007-04-25 14:28 ` [Bug middle-end/31697] [4.3 Regression] " rguenth at gcc dot gnu dot org
2007-04-26 9:48 ` jb at gcc dot gnu dot org
2007-04-26 10:23 ` ubizjak at gmail dot com
2007-04-26 13:53 ` burnus at gcc dot gnu dot org
2007-04-27 4:04 ` bangerth at dealii dot org
2007-04-27 8:49 ` burnus at gcc dot gnu dot org
2007-05-03 13:16 ` burnus at gcc dot gnu dot org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).