Flush memory/cache labview

[Harlequinade]

Sometimes it happens that LabVIEW crashes or that it shows unpredicted behavior. In order to solve this, a fresh restart does the trick.

The new starting of LabVIEW than takes longer then if you close it and start it again. It seems to be that LV creates some kind of cache in the memory, that when you start it again, it doesn't have to initialize things again.

Would it be possible, without restarting the computer, to clear or flush the memory/cache of LV and thus start a "fresh" version of LV?

[Ray.R]

Ideally, LabVIEW shouldn't crash.  Can you describe the situation(s) that causes it to crash?

[Harlequinade]

The reason it crashes is not really the topic of the question, it can have diverse or even unknown reasons. 

[Ray.R]

I appreciate that.  However, I've seldom seen LabVIEW crash.  And when it did, it was related to something in the code or with the OS (windows).  If there is a scenario that we can pinpoint to why LabVIEW crashes, then everyone could benefit from that.

Can you tell us which version of LabVIEW and which OS?

[altenbach]

---

Harlequinade wrote:

The new starting of LabVIEW than takes longer then if you close it and start it again. It seems to be that LV creates some kind of cache in the memory, that when you start it again, it doesn't have to initialize things again.

---

You seem to be talking about memory. Instead of fuzzy words such as "longer" vs. "shorter" times, a better description might be the memory footprint as reported by the task manager. It would also be of general interest to look at the memory use over time while the program is running. Do you have any relevant data along these lines?

A restart after a crash also takes longer because of autorecovery attempts.

What does your program do? Typical causes of memory problems are growing arrays without bounds, constantly opening references whithout ever closing them, a queue with a fast producer loop and a slow consumer loop, etc.

There is also the "request deallocation" function.

[Ray.R]

Thanks Altenbach.

My point exactly.. There are so many possibilities of the code causing the crash that, without seeing the code, it is not possible to propose anything.  Especially if it consistently crashes.

I'm not saying that it is the code, but I would not rule it out.