Hi all,

After a fair bit of tutorial related pain on the mailist recently, I've
just forked the tutorial-devel repo to
https://github.com/jack-h/tutorials_devel and recompiled tutorials 1-3 with
Sysgen 11.5 and the libraries in the main casper github repo.

The tutorials in the 2011 directory *should* work (along with their python
counterparts).

Feel free to give those a go and give me a shout if you still have problems.

Cheers,
Jack

On 9 May 2012 20:49, John Ford <[email protected]> wrote:

> Hi all.
>
> Rich Lacasse at NRAO in Charlottesville has the same problem.  I also
> tried it on our machine in Green Bank, with the same  bad results.  That's
> three separate machines and users with the same problem.  Something's
> wrong with this tutorial, or it is incompatible with the latest
> mlib_devel.  Could someone have a look at it?
>
> Thanks!
>
> John
>
>
> > Hey Gopal,
> >
> > One thing you can try is to start matlab without loading the yellow
> > blocks,
> > open the model file and add them in fresh from the library, save and
> > return
> > to normal usage.
> >
> > This will require you comment out the load xps_library portion of your
> > matlab startup script.  Start matlab, and open the desired model file.
> >  Doing this, you may be able to see the yellow blocks show up as "bad
> > links" with red dashes around them.  From there you can delete them, and
> > manually add them back in from the xps library.  You may need to consult
> > the tutorial, or look at the original model .mdl file in a text editor,
> to
> > make sure you set the parameters appropriately.  From there you should be
> > able to save the .mdl file, close matlab, go back to the startup file and
> > uncomment back to the regular usage.  Hopefully it opens up fine this
> > time.
> >
> >
> > If this solution works, perhaps I can write something more detailed to
> the
> > wiki page for the Tutorials.  Would that be an appropriate edit for the
> > wiki?  I think something stated on that page would help - we have seen
> > several new roach users encounter this issue.
> >
> > Gopal, I was also strongly advised when I started out to use 11.5 and the
> > stable release of the tools.  This solution fixed the problem on that
> > setup, hopefully it does on 13 as well.
> >
> > --Laura
> >
> > On Mon, May 7, 2012 at 12:44 PM, Gopal Narayanan
> > <[email protected]>wrote:
> >
> >> Hello All,
> >>
> >> I am running Matlab R2011A on a 64-bit Debian Linux System. I also have
> >> tested this issue in two other machines, one running Ubuntu (64 bit as
> >> well), and the other running Windows XP-64 with similar results. The
> >> Xilinx System Generator version in all cases is 13.3.4175.
> >>
> >> I have successfully created, compiled and programmed our ROACH board
> >> with tutorials 1 and 2. I am able to download the tutorial model files
> >> from the CASPER website and open Tutorials 1 and 2, and compile/simulate
> >> them. However, when I download the r_spec_2048_103.mdl or the
> >> r_spec_2048_105.mdl (both are tutorial 3 files) from the CASPER git
> >> site, and open within simulink, immediately I get a segfault crash in
> >> matlab.
> >>
> >> I attach the crash dump for my ubuntu machine with this. I contacted
> >> Matlab technical support, but after looking at it, they blame Xilinx
> >> System Generator for this problem even indicating an url that apparently
> >> show memory leaks in sysgen:
> >>
> >> http://www.xilinx.com/support/answers/34287.htm
> >>
> >> I thought I would write to you all to find if anyone has had issues like
> >> mine. And if you have some helpful tips to fix the issue. Alternatively,
> >> if someone could send me an updated saved version of the tutorial 3 file
> >> saved with a more recent system generator version, I would appreciate
> >> it! Maybe I could try with this updated version.
> >>
> >> Thanks a bunch.
> >>
> >> Gopal
> >> --
> >> Gopal Narayanan                          Ph #: (413) 545 0925
> >> Department of Astronomy                  e-mail: [email protected]
> >> University of Massachusetts              Amherst MA 01003
> >>
> >
>
>
>
>

Reply via email to