mlab segmentation fault

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

mlab segmentation fault

MBR MBR
Dear all,

I encounter a segmentation fault using Mayavi (from the EPD distribution, v 7.2.2 rh 5). When I use the image_plane_widget, I get a segfault, with the following backtrace:

Program received signal SIGSEGV, Segmentation fault.
0x00007fffc35e48c9 in vtkUpload3D<unsigned char>::Upload(void*, unsigned char*, unsigned int*, int, long long*, int, int*) () from /opt/epd/lib/vtk-5.6/libvtkRendering.so.5.6

What puzzle me is that I use the same piece of code on another computer under the same system (Ubuntu 12.04.1 LTS) and with the same EPD version, with no segmentation fault at all.

Any help or solution will be greatly appreciated!

Best,

MBR

_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev
Reply | Threaded
Open this post in threaded view
|

Re: mlab segmentation fault

Gael Varoquaux
On Fri, Feb 15, 2013 at 03:17:58PM +0100, MBR MBR wrote:
> What puzzle me is that I use the same piece of code on another computer under
> the same system (Ubuntu 12.04.1 LTS) and with the same EPD version, with no
> segmentation fault at all.

It's probably a problem related to your graphics card :(. You could check
this by running a C debugger and inspecting the stack.

If it's the case, there isn't much that we can do about it :$.

Gaël
_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev
Reply | Threaded
Open this post in threaded view
|

Re: mlab segmentation fault

MBR MBR
I forgot to say that it's a remote machine, and the problem seems indeed to be related to problems with my graphic cards.

I found this old discussion in the mailing list though:
https://mail.enthought.com/pipermail/enthought-dev/2010-April/026084.html
where you suggested a workaround using xvfb to do some offscreen rendering. I tried it (since it's a remote machine, it would be anyway interesting to do offscreen rendering, and if the segfault is related to my graphic card, I guessed this workaround should do it for me).

It works (almost) fine, funny thing being that I encounter the problem Roan LaPlante just mentionned with the scrambled saved figures. I will follow this discussion with great interest!

Thanks for your time,

Best,

MBR


On Fri, Feb 15, 2013 at 3:34 PM, Gael Varoquaux <[hidden email]> wrote:
On Fri, Feb 15, 2013 at 03:17:58PM +0100, MBR MBR wrote:
> What puzzle me is that I use the same piece of code on another computer under
> the same system (Ubuntu 12.04.1 LTS) and with the same EPD version, with no
> segmentation fault at all.

It's probably a problem related to your graphics card :(. You could check
this by running a C debugger and inspecting the stack.

If it's the case, there isn't much that we can do about it :$.

Gaël
_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev


_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev
Reply | Threaded
Open this post in threaded view
|

Re: mlab segmentation fault

MBR MBR
By the way, defining the size of the figure instead of using the magnification parameter does not give a better result (on the contrary...)!


On Fri, Feb 15, 2013 at 4:38 PM, MBR MBR <[hidden email]> wrote:
I forgot to say that it's a remote machine, and the problem seems indeed to be related to problems with my graphic cards.

I found this old discussion in the mailing list though:
https://mail.enthought.com/pipermail/enthought-dev/2010-April/026084.html
where you suggested a workaround using xvfb to do some offscreen rendering. I tried it (since it's a remote machine, it would be anyway interesting to do offscreen rendering, and if the segfault is related to my graphic card, I guessed this workaround should do it for me).

It works (almost) fine, funny thing being that I encounter the problem Roan LaPlante just mentionned with the scrambled saved figures. I will follow this discussion with great interest!

Thanks for your time,

Best,

MBR


On Fri, Feb 15, 2013 at 3:34 PM, Gael Varoquaux <[hidden email]> wrote:
On Fri, Feb 15, 2013 at 03:17:58PM +0100, MBR MBR wrote:
> What puzzle me is that I use the same piece of code on another computer under
> the same system (Ubuntu 12.04.1 LTS) and with the same EPD version, with no
> segmentation fault at all.

It's probably a problem related to your graphics card :(. You could check
this by running a C debugger and inspecting the stack.

If it's the case, there isn't much that we can do about it :$.

Gaël
_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev



_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev