05-29-2015 08:02 AM
As the topic states, I'm pondering if there is a performance hit (or gain?) by making class accessor available through property node?
Specifically, I wonder if the 'class accessor property node' is somehow different than 'normal' property nodes, which I tend to avoid as best I can due to how they often involve the GUI thread and are often detrimental to performance.
And since the follow up question to any answer would be to ask if including error handling terminals changes the answer?
-People who write LVOOP on a semi-regular basis or more, is there a concensus on when/if you use the property node option? Is it preferred as a default almost? Is it affected by whether the accessor is dynamic or static?
Thanks!
QFang
05-31-2015 10:40 AM
05-31-2015 11:29 AM
As far as I know, there is no difference between a subVI call and a property call and they should behave identically behind the scenes. There shouldn't be a forced call in the UI thread, because these don't deal with UI elements. It's simply a different syntax, so it should be the same for DD. I have seen some complaints about performance issues and/or bugs in earlier versions, but I don't remember if there was anything concrete. You can search on LAVA.
Likewise, I don't think the error I/O itself causes delays, but both the serialization of the calls and having case structures acting on those errors can. AQ had some posts on that and he dislikes both. Personally, I tend to like the readability, particularly for serialized calls. Note that some people don't like accessors (at least public ones) as a concept, suggesting that having writers for individual fields leads to all kinds of troubles, and that it's generally better to have specific functions with the relevant inputs.
05-31-2015 12:13 PM
Thanks for your input guys! I appreciate it!