octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: classes.tst hangs octave


From: John W. Eaton
Subject: Re: classes.tst hangs octave
Date: Mon, 30 Oct 2017 20:20:49 -0400
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.2.1

On 10/30/2017 05:16 PM, Dmitri A. Sergatskov wrote:
On Mon, Oct 30, 2017 at 11:45 AM, John W. Eaton <address@hidden <mailto:address@hidden>> wrote:



    If it is hanging, then you could also try to attach to the
    octave-gui process with "gdb -p PID", interrupt the process, and
then get a stack trace to see where it is looping.

​The backtrace is attached (this version compiled w/o Qt, otherwise i endup in some Qt thread).

In that case, you can try

  thread apply all where

to see the stack traces for all threads.

> (gdb) bt
> #0  0x00007f67a678830c in dgemm_ () from /lib64/libopenblas.so.0
> #1 0x00007f67ad0d1836 in xgemm (a=..., b=..., address@hidden, address@hidden) at ../liboctave/array/dMatrix.cc:3008 > #2 0x00007f67ad0d1e24 in operator* (a=..., b=...) at ../liboctave/array/dMatrix.cc:3023 > #3 0x00007f67ae32230b in oct_binop_mul (a1=..., a2=...) at ../libinterp/operators/op-m-m.cc:63


So it is hanging in dgemm from openblas? Hmm. I tried installing openblas on my Debian system but it is still not hanging for me.

Can you try with reference blas and see whether that changes the result for you?

At this point, I'd try to figure out what code Octave is executing. You could try

  call show_octave_backtrace()

from the (gdb) prompt to maybe narrow it down a bit. But you may want to insert some print statements in the tests to really find out what's happening and what values are being multiplied here.

jwe



reply via email to

[Prev in Thread] Current Thread [Next in Thread]