LabVIEW
Performance of In Place Element Structure with nested array-cluster structures
Hello,
I have a big cluster containing an array containing a cluster etc., and i want to change one specific boolean in the total bottom of this hierarchy. The hierarchy represents visualisation, their elements are files, their elements are measurements, their elements are channels.
There are two ways in my example, using just indexing for loops and unbundle/bundle by name, or using in place element structures and shift registers. While the second one is more space consuming, something tells me that this is the correct way. Is the performance better with the in place elements, or not? Or should I totally avoid having one cluster with all the info and maybe rather set up kind of relational database with different tables (clusters) with indices pointing to each others elements? There are two views to the question - performance-wise (that I am going to evaluate for this particular usage) and correctness-wise.
Thanks for programmer advices!
The only way to tell would be to benchmark in detail (I have not), but I would go with the upper code, just because it is more compact and cleaner.
Be aware that the compiler is very smart and there is a good chance that both version will compile to the basically the same binary code after all internal optimizations. How big are these arrays?
