LabVIEW
Long run front panel bugs
If you look in task manager, how does the memory use behave as a function of time?
Is there an update for the graphics driver available, if so, try to upgrade. What else is running on that computer?
Do you have the original LabVIEW code?
Ouch. I think I need a bigger monitor! 😄
(No way to look at that code on my laptop... :()
I was more interested about memory use over time, not the final value. Does memory start low but is creeping up over time at a constant rate?
Who wrote this?
From a quick glance, it is just peppered with illogical code and race conditions. For example the big outer loop can only go to the next iteration after all the inner loops have finished, but at that time the stop button has already been read, causing an additional interation once all inner loops that read the same stop button via locals have finished. If the operator would change the button back to FALSE quicky, the program would be in an unknown state, with some loops running and some not. Looking at the upper left loop inside the big loop, your shift registers never contain more than one element, so why use arrays at all? (You start with an empty array from the uninitialized shift register and append a scalar, creating an array with one element. Then, using an inplace structure for no good reason at all you reverse that array with one element and try to take a much bigger subset (depending on the value of the local), but since the array only contains one element, that's all you get out. Then you reverse the array again.
What's up with all the local variables of hidden controls and all the value property nodes?
At this point I gave up. This is not salvageable or debuggable in its current form. Sorry. 😞
I've wrote this. I started from scratch and a few tutorials, so it's history of my learning and mistakes. It was constant prototyping. But now apparatus, running this VI, works and experiments can be performed (and a user cannot see, how obviously bad the code is). So I have time to refactor the code and apply what i have learned. I will try to remove all "duct tape" solutions, like mentioned big outer loop, that makes additional run.
Hidden controls are used to pass data. I'm going to replace them with Glob.Vars or SubVIs.
Value Nodes used to either allow VI to control itself according experimental data, or to allow user to schedule changes in controls.
Race conditions, due to some specific of experiments, allow to get more statistically reliable data. 
I'm still thinking, should I replace this with synchronized algorithm.
Concerning left up loop. It accumulates data from queue to shift registers. Then the inplace structure looks after array to not exceed certain max size (local variable).
Thanks for your participating! I'l try to accummulate more data about the graphic bug.
