Maximum waveform size for a networked shared variable

[Jon_S.]

Hello Rick,

I ran a test for over an hour this morning.  I had a waveform variable that was hosted and updated on an RT PC.  I then updated a chart from the network published shared variable on a host VI on my development machine.  I didn't receive any errors or crashes.  I'll do an overnight test today.  I need a little more info about your variable.  Is it hosted on the RT target or on the Host VI?  Are you using RT FIFOing?  If so, what are you doing for buffering?  How often are you writing to and reading from the shared variable?

In my test I did not have RT FIFO enabled for the variable that was published over the network.

[Rick_F_Foster]

The waveform variable is currently on my development system, but I also had the same problem when it was on the RT system (That is how I started, but changed to see if it would make a difference).

I have tried it with and without RT FIFO and with/without buffering. Currently I am using RT-FIFO for the network shared variable (5 waveforms with 66560 point per waveform).

 
I have a timed loop on the RT system that wants to write a 66560 element (doubles) waveform every 213ms, but the ADC board can't rearm fast enough to keep up so it drops frames and maintains phase  so it is writing every 426ms.

On the host side, I read the variable every 250ms. I monitor for late finishes and full buffers and get none of those. I get read time outs, because I am reading faster then the real time system is writing.

When I do get a read timeout, I don't try to do anything with the variable and just wait for the next timed loop execution.

Crashes occur anywhere from 1/2hr to several hours. Absolutly no errors are shown anywhere (Labview error handlers, XP Event Logs, Windbg). The system just freezes.

Rick

[Jon_S.]

I'll set up another test to simulate your setup.  Also, is there another PC that you can look at this variable from the Distributed System Manager?  This way we could narrow it down to the PC or program.  Also, it would probably help if you could post your project with the subVIs if that is possible.

[Rick_F_Foster]

I tried running just the DSM on the development machine instead of my analysis vi.

I added the waveform variable to the watch list.

The development system did a hard crash after about 1 hour only running the DSM.

Monday I will run the DSM on another computer and see what happens.

[Jon_S.]

Have you tried monitoring this variable with the DSM on another computer today?  I have been updating a variable with the same properties as yours and monitoring it with the DSM today and have not experienced a crash.  If you post your project I can try to run your project here and replicate the issue.

[Jon_S.]

I ran an overnight test on my machine and I didn't receive any hard crashes.  Have you had any luck on your end?

[Rick_F_Foster]

Regarding the test on another machine: Working on that. My other possible machine is a x64 dual quad core running Vista. I am not sure if that is a supported confgiuration. I just need the IT folks to give me a patch between my lab and office.

I've attached a small version of the project that is crashing.

Workstation VI launched 1st and RT target second.  Probably need to use browse in Workstation to select file "recent_sp.hws" that is in the project directory.

[Rick_F_Foster]

The attempt to run the analysis VI on my x64 Dual Quadcore Vista box did not work so well. Things sort of ran, but the parts trying the host/access the SVE did not seem to work at all.

[Jon_S.]

Hello Rick,

I looked at your project.  I ran it on my system overnight and added the waveform variable to the watch list on the DSM but didn't produce a hard crash.  Can you try to place all of the variables in one library that is hosted on the RT system.  Let me know if you receive the same errors.