public inbox for cygwin@cygwin.com
 help / color / mirror / Atom feed
From: Marco Atzeri <marco.atzeri@gmail.com>
To: cygwin@cygwin.com
Subject: Re: Octave 4.0.0-3 x86_64 segfault (suspect openblas)
Date: Mon, 23 Nov 2015 06:43:00 -0000	[thread overview]
Message-ID: <5652B582.20405@gmail.com> (raw)
In-Reply-To: <BAY169-DS260B02A7A24CFA816E17BA7070@phx.gbl>

On 23/11/2015 04:33, Tony Kelman wrote:
> The following works fine on 32 bit, but segfaults reliably on 64 bit:
>
> $ octave
[cut]
>
> octave:1> D = 1000*(rand(1000,1000)-0.5);
> octave:2> [E,F] = eig(D)
> panic: Segmentation fault -- stopping myself...
> attempting to save variables to 'octave-workspace'...
> save to 'octave-workspace' complete
> Segmentation fault (core dumped)

likely openblas or its iteration with lapack

To test without openblas , you can renable netlib blas
$ mv /usr/bin/cygblas-0.dll  /usr/bin/cygblas-0.dll_bk


> I see the same when I call a mingw-w64 build of openblas 0.2.15 through
> Julia (which isn't packaged in Cygwin but shares many common
> dependencies with Octave), so this is mostly to let Marco know about the
> issue. I'll have to reduce this to the corresponding Lapack call and
> come up with a standalone C or Fortran test case to report upstream. My
> processor is a Sandy Bridge i7-2630QM.
>
> -Tony
>

Same with i5-3320M
The bug seems not present in openblas 0.2.14

As latest octave and openblas now have a debuginfo package:

Program received signal SIGSEGV, Segmentation fault.
0x00000003c866ad50 in cygblas-0!blas_memory_free () from 
/usr/bin/cygblas-0.dll
(gdb) bt
#0  0x00000003c866ad50 in cygblas-0!blas_memory_free () from 
/usr/bin/cygblas-0.dll
#1  0x00000003c841eff8 in dgemv_ () from /usr/bin/cygblas-0.dll
#2  0x00000003aa785fae in dlahr2_ () from /usr/lib/lapack/cyglapack-0.dll
#3  0x00000003aa73f8ec in dgehrd_ () from /usr/lib/lapack/cyglapack-0.dll
#4  0x00000003aa73c1f3 in dgeev_ () from /usr/lib/lapack/cyglapack-0.dll


Regards
Marco

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

  reply	other threads:[~2015-11-23  6:43 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-11-23  3:34 Tony Kelman
2015-11-23  6:43 ` Marco Atzeri [this message]
2015-11-23  9:00   ` Tony Kelman
     [not found]     ` <5652D754.1040405@gmail.com>
2015-11-23 10:01       ` Tony Kelman
2015-11-23 14:22         ` Marco Atzeri

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=5652B582.20405@gmail.com \
    --to=marco.atzeri@gmail.com \
    --cc=cygwin@cygwin.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).