Simulink crash--bus size limitation?

[Jeff Copeland]

I've never used simulink in this way before, so any ideas
as to attack the problem would be appreciated.

I've got this simulation where I'm trying to track
information on hundreds of different objects at the same
time. At each iteration of the simulation I provide a new
report from various sensors that needs to be used to update
the track information. The best way I could see to track
that information was to use a bus that stores the
information in a memory block for feedback at the next
iteration. I've also messed around with the 'Data Store
Memory' blocks and have ran into the exact same issue.

Here is the issue. I've found through trial and error that
when I increase the dimensions of my past a certain point,
when I run the simulation everything MATLAB crashes without
any information whatsoever. All windows just immediately
disappear. This makes me think that there is some low
level memory issue that is corrupting matlab stack space
and causing a violation that doesn't get caught. Is there
any way to get some visibility here?

The bus consists of 16 or so signals of type double, each
of which has dimension of 500. I've found when the
dimension is decreased to 250 I don't have any issues
running.

Most of the model is embedded matlab code. I placed
breakpoints throughout to try and pin down the issue and I
found that blocks that only read from the track bus, didn't
seem to cause crashing. Crashes only occur in blocks that
both read and write that bus information to an output.

Infact, in blocks that both read and write bus info, if I
merely hover over the input bus name while stopped in the
debuger, everything crashes. Again, only an issue with the
larger bus dimension.

Any thoughts as to what may cause this issue? Is there a
bus size limitation? Is there a better way to handle the
track data?

mem: 2Gb
plat: pc
ver: R2007a

[Krishna Balasubramanian]

Hi Jeff,
There are no size limitations for bus elements. As you suspect, this may
be due to a memory corruption that is causing MATLAB to abort. To debug this
issue further, I suggest that you contact The MathWorks technical support
at:

http://www.mathworks.com/support/contact_us/

When you contact technical support, it will be very useful if you can share
your model with support - this will help us identify the problem faster.

Regards,
Krishna Balasubramanian
The MathWorks

"Jeff Copeland" <dont...@abc.com> wrote in message
news:feitg7$pvk$1...@fred.mathworks.com...