LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Choppy behavior using Shared Variables?

Hello - I have several Shared Variables hosted on a PC (Windows XP). An application here shows info from multiple cRIO systems. I have several RT cRIO systems which write to these variables (specific variables for each cRIO). I notice "choppy" front panel behavior on both the RT Host Application and my PC Application that is showing information obtained from the RT-Host apps. This is discouraging in that I am testing this on only one cRIO system for now and already notice this beahvior. For example, graph indicators suddenly freeze on one or the other or both, then kind of catch up to where they should be. (Maybe loops are running too fast?) Using Datasocket writes on the RT side seem to help a little, but I still notice this behavior. Any suggestions? Thanks
0 Kudos
Message 1 of 4
(3,134 Views)
Hello,

Have you tried running the applications with the front panels of the CompactRIO systems closed?  Having these front panels open with a lot of graphs and indicators that update at a high rate cause a very large amount of network traffic overhead between the CompactRIO system and the host PC.  This also increases the load on the CompactRIO's CPU.  You can close the front Panels of a LabVIEW RT quickly by right-clicking the target in the project and selecting "Disconnect". This will leave the VI running without the front panel, and this will be closer to the final embedded solution as well. 

Let me know if that helps.
--Paul Mandeltort
Automotive and Industrial Communications Product Marketing
0 Kudos
Message 2 of 4
(3,117 Views)

Thank you for your response Paul and yuor suggestion - I will try that.

On that subject - does anyone know what overhead is created by showing RT system front panels like that? In other words... do we know how (or can one determine) how much network traffic is created by showing the front panel?

- Con

0 Kudos
Message 3 of 4
(3,112 Views)
It's difficult to provide a concrete number because many factors play into it.  Keep in mind "overhead" includes both network and more important CPU resources.  There is a fixed amount of overhead plus an incremental amount of overhead for the number of indicators and the rate they are updating.  A good trick is to create a seperate "display" loop that updates at slow 15-30 Hz and simply place the indicators in that loop. You can move the data into that loop using single-element RT FIFO enabled Shared Variables.


--Paul Mandeltort
Automotive and Industrial Communications Product Marketing
0 Kudos
Message 4 of 4
(3,109 Views)