Labview 2014 Waveform Min/Max

[orelb]

Hi.

My VI started crashing repeatedly after I upgraded from Labview 2013 to 2014.  After some detective work, I found which part of my code was causing the trouble (see attached  file "prob.png").

After some work, I was able to find a solution (see attached file "sol.png") which involves more VIs and is a bit less elegant.  However, the final result is the same.

I wanted to share this with others who might be experiencing the same thing, and ask whether or not this is a known issue...?  Maybe there is a better way to fix this...

Thanks,

Orel.

Problematic code:

Solution:

[GerdW]

Hi Orel,

- What means "crashing" to you?

- What's the data in "waveform out 2"?

- What happens when you use a wire instead of a local variable?

[orelb]

Hi GerdW,

1.  "Crashing" is getting the attached "LabVIEW Crash Reporter" window, after Labview closes.

2.  The data is just a regular waveform, for example:

                                                                                                     

3.  Using a wire instead of a local variable seems to work.  Any explanation for this?

Thanks,

Orel.

[GerdW]

Hi Orel,

3. THINK DATAFLOW! 😄

It could be related to some weird kind of compiler optimization in combination with your local variable. With DATAFLOW the compiler is much more sure about when to touch data…

[James.M]

Are you looping as fast as possible? A waveform is a weird thing to have as a local variable because usually you want data points one by one without repeated points.

Like GerdW said, dataflow. If you can, share your code so we can take a look. This could be something weird with that local variable and a race condition.

[orelb]

Yes, but it used to work in Labview 2013 and it works if I change around some VIs...  For some strange reason, only this VI combination crashes it.

[James.M]

There are compiler differences between 2013 and 2014, so if something wasn't coded in the best way it could cause a problem with one compiler and not another. Dataflow is a big deal.

[orelb]

Thanks, James.

It's no problem using a local variable.  I have one infinite loop which acquires the waveform using an FPGA and a second loop which analyzes the data.  There shouldn't be any repeated points because the acquisition is done so that the variable is updated only when a full waveform is output.  I would rather have both loops running independently.  If I use a wire, then that's not the case.

[orelb]

OK thanks.

I'll keep thinking about it.

[GerdW]

Hi Orel,

use a queue, when you want to have a nice producer-consumer scheme!