Veristand Crash

[E-man-O]

Hello,

I'm working on a Veristand Project which has an ample number of inports and outports. I am able to compile the model successfully but when undeploying this particular model, Veristand always crashes. It's a simulation model.  Interesting enough, a coworker has worked with the same model but doesn't have the same crashing problem that I have when undeploying.

I have uninstalled and reinstalled Veristand hoping that will rectify the situation but the issue is still persistent.

Please advice. Thanks

[Justin_P]

Hello,

Have you been able to reproduce this with other models? Can you deploy and undeploy the Sinewave Delay example without a crash?  And how have we narrowed the crash down to the model?  Is there a third development system that we can test on?

Those should be some pretty quick tests.  I look forward to hearing the results.

[E-man-O]

Hello, 

I have been able to deploy other models without a crash (including the Sinewave Delay). It's worth mentioning that Veristand has been a bit laggy on my present machine.

With this particular model, i am able to deploy but Veristand always crashes upon undeployment. It's a Veristand project which loads a simulink model.

This very Veristand project works perfectly on another coworkers system so i doubt it's the simulink mdl being called by Veristand.

I have attached the error log and a screen shot of the error message. Maybe that can throw more light on it.

[Justin_P]

Hello,

Is there anything different between the two systems?  Can you do a MAX report and check the driver versions.  Is there anything different?  If you take the model out of the system defination file does it stop crashing?  If you make a new project with just the model, does it crash as well?

Thanks,

[E-man-O]

Hello,

As per the two systems i dont know of anything differernt. we both have Veristand 2011 installed along with the same versions of LabView runtime. I was able to perform

the MAX report on my machine but We weren't able to perfom it on my coworker's machine. Whenever he tried to, the Measurement and Automation Explorer crashes. The

model works perfectly by itself. The crashing always occurs when i'm undeploying the project. I did make a new project with the model but i still have Veristand crashing.

Emmanuel

[ClaudioH]

Hi,

Are you and your co-worker using the same model dll or are you using the same mdl to generate the dll? The problem may be in the compilation process.

How many inports and outports do it have?

Are you importing all signals and parameters?

Cheers,

[E-man-O]

Hello Claudio,

My co-worker and I are essentially using the same mdl. We aren't generating a dll, just adding the Simulink model in the System Definition File of the Project. The model compiles just fine in Simulink but we have been having an issue with Veristand not loading all the ports into the project. (though his doesn't crash). I believe NI is looking into his problem as at now.

To further investigate what the issue with mine might be, i reduced the number of inport/ouports of my model by taking out some subsystems (and thus inports and outports) and it seemed  everything worked fine then with all ports being loaded into the project. I then added another subsystem to my Simulink mdl to again increase the number of inports/outport. This time around all the ports weren't loaded again and the crashing effect reproduced itself.

I'm beginning to think it might have something to do with Veristand not being able to handle the model inports and outports.

It's a mere 178 inports and ouports as at now...we are only in the nascent stages of our project.

Thanks,

Emmanuel

[Kevin_F]

Hey Emmanuel,

I looked at the crash dump from the error and it says:The thread tried to read from or write to a virtual address for which it does not have the appropriate access. If the error goes away with less channels then it may be that your computer needs more virtual memory than it is allowed to alocate to run the model. Would you be able to increase the amount of virtual memory allocated on this machine? I think you may also run into a similar error if not running as an administrator on the PC.  I also wanted to confirm that all of these models are running on the Veristand Development machine (not a PXI controller).

Regards,

Kevin Fort

Applications Engineering | National Instruments

www.ni.com/support

[NestorG]

Hello E-man-O,

Is this forum post from your coworker? If so, we have indeed identified the issue and it has been reported to R&D (#335859). Technically, the issue is not necessarily with the number of ports, but rather with the combined length of the ports' paths; that is the total length of the string starting at the root MDL node down through each final subsystem name. Arun was informed to reduce the length of each subsystem name in an effort to reduce the overall length of the string as much as possible. Your coworker has more details on this issue. Please inform us, via the service request, if this workaround was insufficient so that the fix may be addressed accordingly.

[E-man-O]

Hello Kevin,

Sorry for the wait. The models are running on my machine (Laptop) and not a PXI controller. I did increase the VM on the machine and run as an admin ( I have local admin privileges ) but unfortunately the same error and crashing effect was reproduced. Any more thoughts?

Emmanuel.