Linux Archive

Linux Archive (http://www.linux-archive.org/)
-   Debian User (http://www.linux-archive.org/debian-user/)
-   -   Octave crashed (http://www.linux-archive.org/debian-user/499542-octave-crashed.html)

Zbigniew Komarnicki 03-10-2011 08:57 AM

Octave crashed
 
Hello,

I have a problem with octave, see the following session from octave, it is
very simple (I want to compute eig(a) and then crash):

-------------------------------------------------------
$ octave
GNU Octave, version 3.2.4
Copyright (C) 2009 John W. Eaton and others.
This is free software; see the source code for copying conditions.
There is ABSOLUTELY NO WARRANTY; not even for MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. For details, type `warranty'.

Octave was configured for "i486-pc-linux-gnu".

Additional information about Octave is available at http://www.octave.org.

Please contribute if you find this software useful.
For more information, visit http://www.octave.org/help-wanted.html

Report bugs to <bug@octave.org> (but first, please read
http://www.octave.org/bugs.html to learn how to write a helpful report).

For information about changes from previous versions, type `news'.

warning: mark_as_command is obsolete and will be removed from a future version
of Octave
octave:1> a=rand(3,3)
a =

0.189941 0.570931 0.517541
0.092924 0.449368 0.680880
0.146931 0.266066 0.706786

octave:2> eig(a)
panic: Illegal instruction -- stopping myself...
attempting to save variables to `octave-core'...
save to `octave-core' complete
Illegal instruction
-------------------------------------------------------


I use Debian Squueze 6.0. I have installed every package with octave-*

$ uname -a
Linux S1 2.6.32-5-686 #1 SMP Wed Jan 12 04:01:41 UTC 2011 i686 GNU/Linux

Do you have also such problem or it is only specific to me on my hardware.

Zbigniew


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201103101057.03168.cblasius@gmail.com">http://lists.debian.org/201103101057.03168.cblasius@gmail.com

George 03-10-2011 09:35 AM

Octave crashed
 
On 3/10/11, Zbigniew Komarnicki <cblasius@gmail.com> wrote:

> octave:1> a=rand(3,3)
> a =
>
> 0.189941 0.570931 0.517541
> 0.092924 0.449368 0.680880
> 0.146931 0.266066 0.706786
>
> octave:2> eig(a)

I had the same problem, but on a non-Debian system. On my system the problem
was caused by using the SSE3 atlas libraries and was fixed by installing with
SSE2 libraries.


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: AANLkTimOFuh0x=RUq4pC16VoMoDEyJ24UNkw+bu8OrNn@mail .gmail.com">http://lists.debian.org/AANLkTimOFuh0x=RUq4pC16VoMoDEyJ24UNkw+bu8OrNn@mail .gmail.com

Zbigniew Komarnicki 03-10-2011 04:56 PM

Octave crashed
 
On Thursday 10 of March 2011 11:35:14 George wrote:
> I had the same problem, but on a non-Debian system. On my system the
> problem was caused by using the SSE3 atlas libraries and was fixed by
> installing with SSE2 libraries.

Thank you for your help.

I try to search atlas with sse, because my computer is old
--------------------------------------------------------------------------------
$ cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 6
model : 8
model name : AMD Athlon(tm) XP 2000+
stepping : 1
cpu MHz : 1660.266
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up
bogomips : 3320.53
clflush size : 32
cache_alignment : 32
address sizes : 34 bits physical, 32 bits virtual
power management: ts
--------------------------------------------------------------------------------

and haven't sse2 or sse3. I do

------------------------------------------------------------------------------------
$ dpkg -l *atlas*
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version
Description
+++-==================-=============-
======================================
un atlas-doc <none> (no description available)
un atlas2 <none> (no description available)
un atlas2-dev <none> (no description available)
un atlas2-headers <none> (no description available)
un atlas3-base <none> (no description available)
un atlas3-doc <none> (no description available)
un libatlas-base-dev <none> (no description available)
ii libatlas-doc 3.8.3-27 Automatically Tuned Linear Algebra
Software, documentation
un libatlas.so.3gf <none> (no description available)
un libatlas3gf-2.0 <none> (no description available)
un libatlas3gf-altivec <none> (no description available)
un libatlas3gf-amd64sse3 <none> (no description available)
ii libatlas3gf-base 3.8.3-27 Automatically Tuned Linear Algebra
Software, generic shared
un libatlas3gf-core2sse3 <none> (no description available)
un libatlas3gf-corei7sse3 <none> (no description available)
un libatlas3gf-ev6 <none> (no description available)
un libatlas3gf-sse <none> (no description available)
un libatlas3gf-sse2 <none> (no description available)
un libatlas3gf-sse3 <none> (no description available)
un libatlas3gf-v9 <none> (no description available)
----------------------------------------------------------------------------

I found this
un libatlas3gf-sse <none> (no description available)

but after:
------------------------------------------------------------------------------------
# aptitude install libatlas3gf-sse
No candidate version found for libatlas3gf-sse
No candidate version found for libatlas3gf-sse
No packages will be installed, upgraded, or removed.
0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 0 B will be used.
------------------------------------------------------------------------------------

This package is unavailable so what to do to solve this problem ?
Any help will be very appreciated.

Zbigniew


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201103101856.16274.cblasius@gmail.com">http://lists.debian.org/201103101856.16274.cblasius@gmail.com

George 03-11-2011 06:12 AM

Octave crashed
 
On 3/10/11, Zbigniew Komarnicki <cblasius@gmail.com> wrote:

> This package is unavailable so what to do to solve this problem ?
> Any help will be very appreciated.

I tried to find packages that would help you choose different
optimizations, but I was unable to find any. Maybe you can try posting
to the debian maintaners' development list at
http://pkg-octave.alioth.debian.org/#lists . They should be able to
help.


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: AANLkTikoDrUi6DYaj26ZXbMt6QwCpv9x+DEtYOXxO3mA@mail .gmail.com">http://lists.debian.org/AANLkTikoDrUi6DYaj26ZXbMt6QwCpv9x+DEtYOXxO3mA@mail .gmail.com

"Pedro A. Tonelli" 03-12-2011 02:20 PM

Octave crashed
 
Hi

2011/3/10 Zbigniew Komarnicki <cblasius@gmail.com>

On Thursday 10 of March 2011 11:35:14 George wrote:

> I had the same problem, but on a non-Debian system. On my system the

> problem was caused by using the SSE3 atlas libraries and was fixed by

> installing with SSE2 libraries.



Thank you for your help.



I try to search atlas with sse, because my computer is old

--------------------------------------------------------------------------------

What is the answer to the command
sudo update-alternatives --list liblapack.so.3gf

*In my case I have to adjust this library to /usr/lib/lapack/liblapack.so.3gf. I did it with

sudo update-alternatives --config liblapack.so.3gf

I think the package you need is libatlas3gf-base.

hope this helps
--
-- Pedro Aladar Tonelli --

Zbigniew Komarnicki 03-13-2011 05:29 PM

Octave crashed
 
On Saturday 12 of March 2011 22:56:48 you wrote:
> 2011/3/12 Zbigniew Komarnicki <cblasius@gmail.com>
>
> > I do this but with no success. Thank you anyway.
> >
> > Zbigniew
>
> Hmm. But i dont think the problem has to do with the processor, since I am
> running octave in a very
> similar one (athlon xp 2600). I am attaching the result of "ldd
> /usr/bin/octave" so you can inspect if we have the same
> libraries. Good luck
>
> ldd /usr/bin/octave
>
> linux-gate.so.1 => (0xb78c2000)
> liboctinterp.so => /usr/lib/octave-3.2.4/liboctinterp.so (0xb6c0d000)
> liboctave.so => /usr/lib/octave-3.2.4/liboctave.so (0xb614d000)
> libcruft.so => /usr/lib/octave-3.2.4/libcruft.so (0xb60d6000)
> libumfpack.so.5.4.0 => /usr/lib/libumfpack.so.5.4.0 (0xb6011000)
> libamd.so.2.2.0 => /usr/lib/libamd.so.2.2.0 (0xb6009000)
> libcamd.so.2.2.0 => /usr/lib/libcamd.so.2.2.0 (0xb6000000)
> libcolamd.so.2.7.1 => /usr/lib/libcolamd.so.2.7.1 (0xb5ff9000)
> libcholmod.so.1.7.1 => /usr/lib/libcholmod.so.1.7.1 (0xb5f38000)
> libccolamd.so.2.7.1 => /usr/lib/libccolamd.so.2.7.1 (0xb5f2e000)
> libcxsparse.so.2.2.3 => /usr/lib/libcxsparse.so.2.2.3 (0xb5f02000)
> liblapack.so.3gf => /usr/lib/liblapack.so.3gf (0xb56fd000)
> libblas.so.3gf => /usr/lib/libblas.so.3gf (0xb5681000)
> libfftw3.so.3 => /usr/lib/libfftw3.so.3 (0xb553c000)
> libfftw3f.so.3 => /usr/lib/libfftw3f.so.3 (0xb53ff000)
> libqrupdate.so.1 => /usr/lib/libqrupdate.so.1 (0xb53e1000)
> libarpack.so.2 => /usr/lib/libarpack.so.2 (0xb5394000)
> libftgl.so.2 => /usr/lib/libftgl.so.2 (0xb5367000)
> libfreetype.so.6 => /usr/lib/libfreetype.so.6 (0xb52e0000)
> libz.so.1 => /usr/lib/libz.so.1 (0xb52cc000)
> libGL.so.1 => /usr/lib/libGL.so.1 (0xb5267000)
> libGLU.so.1 => /usr/lib/libGLU.so.1 (0xb51f7000)
> libX11.so.6 => /usr/lib/libX11.so.6 (0xb50d6000)
> libreadline.so.6 => /lib/libreadline.so.6 (0xb50a1000)
> libncurses.so.5 => /lib/libncurses.so.5 (0xb5067000)
> libdl.so.2 => /lib/i686/cmov/libdl.so.2 (0xb5063000)
> libhdf5.so.6 => /usr/lib/libhdf5.so.6 (0xb4cdb000)
> libgfortran.so.3 => /usr/lib/libgfortran.so.3 (0xb4c1a000)
> libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0xb4b2b000)
> libm.so.6 => /lib/i686/cmov/libm.so.6 (0xb4b05000)
> libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xb4ae9000)
> libpthread.so.0 => /lib/i686/cmov/libpthread.so.0 (0xb4ad0000)
> libc.so.6 => /lib/i686/cmov/libc.so.6 (0xb498a000)
> libpcre.so.3 => /lib/libpcre.so.3 (0xb494c000)
> libXext.so.6 => /usr/lib/libXext.so.6 (0xb493d000)
> libXxf86vm.so.1 => /usr/lib/libXxf86vm.so.1 (0xb4938000)
> libXdamage.so.1 => /usr/lib/libXdamage.so.1 (0xb4935000)
> libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0xb492f000)
> libdrm.so.2 => /usr/lib/libdrm.so.2 (0xb4925000)
> libxcb.so.1 => /usr/lib/libxcb.so.1 (0xb490c000)
> /lib/ld-linux.so.2 (0xb78c3000)
> librt.so.1 => /lib/i686/cmov/librt.so.1 (0xb4903000)
> libXau.so.6 => /usr/lib/libXau.so.6 (0xb48ff000)
> libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0xb48fa000)

Thank you for your help. I have the same but with different address allocation
(oxZZZZZZZZ).


Plase, see

$ ldd /usr/bin/octave
linux-gate.so.1 => (0xb7876000)
liboctinterp.so => /usr/lib/octave-3.2.4/liboctinterp.so (0xb6bc1000)
liboctave.so => /usr/lib/octave-3.2.4/liboctave.so (0xb6101000)
libcruft.so => /usr/lib/octave-3.2.4/libcruft.so (0xb608a000)
libumfpack.so.5.4.0 => /usr/lib/libumfpack.so.5.4.0 (0xb5fd1000)
libamd.so.2.2.0 => /usr/lib/libamd.so.2.2.0 (0xb5fc9000)
libcamd.so.2.2.0 => /usr/lib/libcamd.so.2.2.0 (0xb5fc0000)
libcolamd.so.2.7.1 => /usr/lib/libcolamd.so.2.7.1 (0xb5fb9000)
libcholmod.so.1.7.1 => /usr/lib/libcholmod.so.1.7.1 (0xb5ef8000)
libccolamd.so.2.7.1 => /usr/lib/libccolamd.so.2.7.1 (0xb5eee000)
libcxsparse.so.2.2.3 => /usr/lib/libcxsparse.so.2.2.3 (0xb5ec2000)
liblapack.so.3gf => /usr/lib/liblapack.so.3gf (0xb568e000)
libblas.so.3gf => /usr/lib/libblas.so.3gf (0xb5385000)
libfftw3.so.3 => /usr/lib/libfftw3.so.3 (0xb5240000)
libfftw3f.so.3 => /usr/lib/libfftw3f.so.3 (0xb5103000)
libqrupdate.so.1 => /usr/lib/libqrupdate.so.1 (0xb50e5000)
libarpack.so.2 => /usr/lib/libarpack.so.2 (0xb5098000)
libftgl.so.2 => /usr/lib/libftgl.so.2 (0xb506b000)
libfreetype.so.6 => /usr/lib/libfreetype.so.6 (0xb4ff4000)
libz.so.1 => /usr/lib/libz.so.1 (0xb4fe0000)
libGL.so.1 => /usr/lib/libGL.so.1 (0xb4f7b000)
libGLU.so.1 => /usr/lib/libGLU.so.1 (0xb4f0b000)
libX11.so.6 => /usr/lib/libX11.so.6 (0xb4ded000)
libreadline.so.6 => /lib/libreadline.so.6 (0xb4db8000)
libncurses.so.5 => /lib/libncurses.so.5 (0xb4d7e000)
libdl.so.2 => /lib/i686/cmov/libdl.so.2 (0xb4d7a000)
libhdf5.so.6 => /usr/lib/libhdf5.so.6 (0xb49f2000)
libgfortran.so.3 => /usr/lib/libgfortran.so.3 (0xb492c000)
libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0xb4837000)
libm.so.6 => /lib/i686/cmov/libm.so.6 (0xb4811000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xb47f3000)
libpthread.so.0 => /lib/i686/cmov/libpthread.so.0 (0xb47da000)
libc.so.6 => /lib/i686/cmov/libc.so.6 (0xb4694000)
libpcre.so.3 => /lib/libpcre.so.3 (0xb4660000)
libXext.so.6 => /usr/lib/libXext.so.6 (0xb4651000)
libXxf86vm.so.1 => /usr/lib/libXxf86vm.so.1 (0xb464c000)
libXdamage.so.1 => /usr/lib/libXdamage.so.1 (0xb4649000)
libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0xb4643000)
libdrm.so.2 => /usr/lib/libdrm.so.2 (0xb4639000)
libxcb.so.1 => /usr/lib/libxcb.so.1 (0xb4620000)
/lib/ld-linux.so.2 (0xb7877000)
librt.so.1 => /lib/i686/cmov/librt.so.1 (0xb4617000)
libXau.so.6 => /usr/lib/libXau.so.6 (0xb4613000)
libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0xb460e000)



Zbigniew


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201103131929.47444.cblasius@gmail.com">http://lists.debian.org/201103131929.47444.cblasius@gmail.com

Thomas Weber 03-16-2011 06:19 PM

Octave crashed
 
Hi,

guys, if you forward a mail from a thread to a new recipient (in this
case the pkg-octave-devel list), please include *all* relevant
information in the first mail. I actually thought that we received that
mail in error, because there was no mention of Debian *at all* in the
mail - and even then, there's no mention of any problem in the forwarded
mail, just some ldd outputs.


On Sun, Mar 13, 2011 at 07:29:46PM +0100, Zbigniew Komarnicki wrote:
> On Saturday 12 of March 2011 22:56:48 you wrote:
> > 2011/3/12 Zbigniew Komarnicki <cblasius@gmail.com>
> >
> > > I do this but with no success. Thank you anyway.
<snip>
> > ldd /usr/bin/octave
<snip>.

Zbigniew, please remove all packages with 'atlas' in their name from the
system, then try again. If it still crashes, open a bug against
octave3.2, using reportbug. If it doesn't crash anymore, try with the
atlas base package (libatlas3gf-base).

Thanks
Thomas


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110316191930.GA26343@bully">http://lists.debian.org/20110316191930.GA26343@bully

Zbigniew Komarnicki 03-17-2011 08:41 AM

Octave crashed
 
On Wednesday 16 of March 2011 20:19:30 Thomas Weber wrote:
> Hi,
>
> guys, if you forward a mail from a thread to a new recipient (in this
> case the pkg-octave-devel list), please include *all* relevant
> information in the first mail. I actually thought that we received that
> mail in error, because there was no mention of Debian *at all* in the
> mail - and even then, there's no mention of any problem in the forwarded
> mail, just some ldd outputs.

> Zbigniew, please remove all packages with 'atlas' in their name from the
> system, then try again. If it still crashes, open a bug against
> octave3.2, using reportbug. If it doesn't crash anymore, try with the
> atlas base package (libatlas3gf-base).

Thank you for your help.

After removing all atlas package octave works correctly.

If I install "libatlas3gf-base" the octave crashed as before.

I'm very sorry for the forward mail from a thread to a new recipient.
I was thinking, if such mail reach to pkg-octave-devel then you will be
interested to check debian-user list for this thread, because it refer to
octave. I'm sorry once again and in the future I will start new thread with
full information.

I have one more question: where first to signal if given program crashed?
To debian-user list or to specific list e.g. *-devel ? Because also "scilab"
crased by atlas package. After removing all atlas packages I check that scilab
now works.
http://lists.debian.org/debian-user/2011/03/msg01112.html


> Thanks
> Thomas

Thank you,
Zbigniew


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201103171041.06924.cblasius@gmail.com">http://lists.debian.org/201103171041.06924.cblasius@gmail.com

Thomas Weber 03-17-2011 07:44 PM

Octave crashed
 
On Thu, Mar 17, 2011 at 10:41:06AM +0100, Zbigniew Komarnicki wrote:
> On Wednesday 16 of March 2011 20:19:30 Thomas Weber wrote:
> > Hi,
> >
> > guys, if you forward a mail from a thread to a new recipient (in this
> > case the pkg-octave-devel list), please include *all* relevant
> > information in the first mail. I actually thought that we received that
> > mail in error, because there was no mention of Debian *at all* in the
> > mail - and even then, there's no mention of any problem in the forwarded
> > mail, just some ldd outputs.
>
> > Zbigniew, please remove all packages with 'atlas' in their name from the
> > system, then try again. If it still crashes, open a bug against
> > octave3.2, using reportbug. If it doesn't crash anymore, try with the
> > atlas base package (libatlas3gf-base).
>
> Thank you for your help.
>
> After removing all atlas package octave works correctly.
>
> If I install "libatlas3gf-base" the octave crashed as before.
>
> I'm very sorry for the forward mail from a thread to a new recipient.
> I was thinking, if such mail reach to pkg-octave-devel then you will be
> interested to check debian-user list for this thread, because it refer to
> octave. I'm sorry once again and in the future I will start new thread with
> full information.

It's not necessary to start a new thread, just give a short summary.
I'm offline most of the time, though. So at the time I read your mail, I
usually don't have net access.

> I have one more question: where first to signal if given program crashed?
That depends. If you think you are doing something wrong, asking on a
user list is probably the best start. If you are somewhat sure that
there's a bug, file a bug with reportbug. There's no black and white
here, but lots of shades of grey.

Now, please file a bug against Atlas using reportbug (which includes
some helpful information for the atlas maintainer by default).

Thomas


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20110317204420.GB7494@bully">http://lists.debian.org/20110317204420.GB7494@bully

Zbigniew Komarnicki 03-19-2011 02:51 AM

Octave crashed
 
On Thursday 17 of March 2011 21:44:20 Thomas Weber wrote:

> It's not necessary to start a new thread, just give a short summary.
> I'm offline most of the time, though. So at the time I read your mail, I
> usually don't have net access.
>
> > I have one more question: where first to signal if given program crashed?
>
> That depends. If you think you are doing something wrong, asking on a
> user list is probably the best start. If you are somewhat sure that
> there's a bug, file a bug with reportbug. There's no black and white
> here, but lots of shades of grey.
>
> Now, please file a bug against Atlas using reportbug (which includes
> some helpful information for the atlas maintainer by default).
>
> Thomas


Thank you.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=576972

Zbigniew


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201103190451.54535.cblasius@gmail.com">http://lists.debian.org/201103190451.54535.cblasius@gmail.com


All times are GMT. The time now is 06:14 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.