01-17-2013 09:25 AM
Hi folks
I'd like to hear your opinions about using feedback nodes.
- Do you use them in your code?
- Do you prefer them to shift registers? Why?
I don't really like them because in my view they make a program harder to understand and they contradict the LabVIEW "left-to-right"-principle.
I haven't found such a discussion here yet...
01-17-2013 09:35 AM
This has been discussed somewhere on this forum. Not sure where though.
Anyways, I use both. Feedback nodes are great for when you don't have to pass data out of your loop and you only need the value in 1 spot. This is usually a delay function (like comparing current value to previous value).
Now, as using them for FGV: I don't really care. I tend to prefer the FB Node simply because I don't like dealing with a loop that only iterates once.
01-17-2013 09:42 AM
Depending on the target, implementers are used to use different setups.
E.g. pipelining on FPGA is more often implemented with FB nodes while passing data from one iteration to the next one on windows and RT is more often implemented with SR's.
As a matter of fact:
For code performance, it does not make a difference. The LV compiler abstracts FB nodes and SR's to the same transform. So essentially, this discussion does only affect graphical representation of the code.
Norbert
01-17-2013 10:54 AM
I usually use feedback nodes if I don't need a loop or if the code is in a loop, but is localized (for instance, using the output of the tick count function to measure how much time each iteration took) or nested inside case structures.
You can see more here - http://labviewartisan.blogspot.com/2009/03/why-are-you-still-using-loops-in-your.html