Another segfault with ATLAS and large matrices

32 views
Skip to first unread message

leif

unread,
Aug 10, 2011, 1:42:53 AM8/10/11
to sage-devel
I don't know yet if this is again with a *binary* Sage distribution:

http://trac.sagemath.org/sage_trac/ticket/11674

See also

http://trac.sagemath.org/sage_trac/ticket/11581

for a previous one, and also

http://groups.google.com/group/sage-support/browse_thread/thread/4b421d04fb2d2297


[cc'ing anyone unfortunately currently doesn't make much sense...]


-leif

leif

unread,
Aug 10, 2011, 2:45:41 AM8/10/11
to sage-devel
On 10 Aug., 07:42, leif <not.rea...@online.de> wrote:
> I don't know yet if this is again with a *binary* Sage distribution:
>
> http://trac.sagemath.org/sage_trac/ticket/11674

I can reproduce this error with the *binary* Sage 4.7 distribution on
Ubuntu 10.04.3 x86_64 (Core2, Penryn).

Note that it was built on redhawk, an AMD Opteron 8439 SE; I btw.
don't know why the outdated 10.04.1 release is used on that.


-leif

>
> See also
>
> http://trac.sagemath.org/sage_trac/ticket/11581
>
> for a previous one, and also
>
> http://groups.google.com/group/sage-support/browse_thread/thread/4b42...

leif

unread,
Aug 11, 2011, 6:52:11 AM8/11/11
to sage-devel
On 10 Aug., 08:45, leif <not.rea...@online.de> wrote:
> On 10 Aug., 07:42, leif <not.rea...@online.de> wrote:
>
> > I don't know yet if this is again with a *binary* Sage distribution:

It is.


> >http://trac.sagemath.org/sage_trac/ticket/11674
>
> I can reproduce this error with the *binary* Sage 4.7 distribution on
> Ubuntu 10.04.3 x86_64 (Core2, Penryn).
>
> Note that it was built on redhawk, an AMD Opteron 8439 SE; I btw.
> don't know why the outdated 10.04.1 release is used on that.


Could someone perhaps try building a binary distribution (with
SAGE_FAT_BINARY=yes and) with the new ATLAS spkg from #10226 [1]
(which isn't merged into 4.7.1) on redhawk to see if that fixes the
problem?

I think it's quite odd to just tell people "compile from source" when
they get segfaults with the binary distribution, and in case the new
ATLAS spkg should fix this, we IMHO shouldn't continue to provide
broken binary dists for 4.7.1.

Another try would be to prepare (at least) the Ubuntu 10.04.* 64-bit
binary distribution on a different machine.


-leif

[1] http://trac.sagemath.org/sage_trac/ticket/10226

Volker Braun

unread,
Aug 11, 2011, 5:21:47 PM8/11/11
to sage-...@googlegroups.com
I built a Sage binary distribution on redhawk with atlas-3.8.4.spkg and SAGE_FAT_BINARY=yes. It is at:

http://boxen.math.washington.edu/home/vbraun/Sage/sage-4.7.1.rc2/dist/sage-4.7.1.rc2-fat-x86_64-Linux.tar.gz

Please let me know if it fixes the segfaults. Or post your backtrace and processor info if not ;-)

Volker

leif

unread,
Aug 11, 2011, 8:15:15 PM8/11/11
to sage-devel
On 11 Aug., 23:21, Volker Braun <vbraun.n...@gmail.com> wrote:
> I built a Sage binary distribution on redhawk with atlas-3.8.4.spkg and
> SAGE_FAT_BINARY=yes. It is at:
>
> http://boxen.math.washington.edu/home/vbraun/Sage/sage-4.7.1.rc2/dist...
>
> Please let me know if it fixes the segfaults. Or post your backtrace and
> processor info if not ;-)

Unfortunately it doesn't, see #11674 for details.


-leif

leif

unread,
Aug 11, 2011, 8:48:01 PM8/11/11
to sage-devel
Perhaps disabling multithreading in ATLAS when building binary dists
helps.

I wonder if redhawk's libpthread is somehow incompatible; does it
really run 10.04.*1*?

I have (Ubuntu 10.04.3):

-rwxr-xr-x 1 root root 135745 Jan 21 2011 /lib/libpthread-2.11.1.so
lrwxrwxrwx 1 root root 20 Feb 2 2011 /lib/libpthread.so.0 ->
libpthread-2.11.1.so

-rw-r--r-- 1 root root 40298 Jan 21 2011 /usr/include/pthread.h

So it is clearly younger than that originally shipped with 10.04(.1);
alternatively, we would have to include redhawk's, but that's IMHO a
rather bad idea. Better upgrade redhawk...


-leif

P.S.: I'll attach my pthread.h to #11674.

leif

unread,
Aug 21, 2011, 12:34:29 AM8/21/11
to sage-devel
The new Sage 4.7.1 binary, again built on redhawk (apparently still
running Ubuntu 10.04.1) fails in exactly the same way:

sage: import numpy as np
sage: ms = np.load("/home/leif/Sage/tmp/matrices.npy")
sage: time np.dot(ms[0],ms[1])
/tmp/sage-4.7.1-linux-64bit-ubuntu_10.04.1_lts-x86_64-Linux/local/lib/
libcsage.so(print_backtrace+0x31)[0x7f3b480f3817]
/tmp/sage-4.7.1-linux-64bit-ubuntu_10.04.1_lts-x86_64-Linux/local/lib/
libcsage.so(sigdie+0x14)[0x7f3b480f3849]
/tmp/sage-4.7.1-linux-64bit-ubuntu_10.04.1_lts-x86_64-Linux/local/lib/
libcsage.so(sage_signal_handler+0x20e)[0x7f3b480f3474]
/lib/libpthread.so.0(+0xf8f0)[0x7f3b4d35a8f0]
/tmp/sage-4.7.1-linux-64bit-ubuntu_10.04.1_lts-x86_64-Linux/local/lib/
libatlas.so(+0x6b70d)[0x7f3b41f7470d]

------------------------------------------------------------------------
Unhandled SIGSEGV: A segmentation fault occurred in Sage.
This probably occurred because a *compiled* component of Sage has a
bug
in it and is not properly wrapped with sig_on(), sig_off(). You might
want to run Sage under gdb with 'sage -gdb' to debug this.
Sage will now terminate.
------------------------------------------------------------------------
/tmp/sage-4.7.1-linux-64bit-ubuntu_10.04.1_lts-x86_64-Linux/local/bin/
sage-sage: line 301: 22150 Segmentation fault sage-ipython "$@" -
i


-leif
Reply all
Reply to author
Forward
0 new messages