Let's say one is going to be embarking on a fairly large application. Large is subjective but assume something like a very large test system that will be commanding a pool of instruments to perform a variety of tasks in different parallel processes. I suppose the closest thing most here would relate to is something like teststand (it's actually quite different in purpose and teststand would not be suitable). Anyway, you get the gist, a large application that will probably involve 50+ modules minimum for just the base core of functionality.
It's not really a thing where data is acquired at a breathtaking pace like what most seem to do around here, but there will be a lot going on and I've seen some crude attempts at a similar thing by some former colleagues when we were not very experienced with LabVIEW and it will absolutely get to the point where things feel slow/unresponsive if one doesn't specifically architect everything to be non-blocking/asynchronous,, as I've witnessed. It's quite distributed between a main server PC using Windows, a main RT server, and about 20 or so windows clients communicating over the network.
Basically, AF and DQMH are quite well suited to the challenge.
But I'm curious if there are any hidden dragons when things get to this scale. Am I asking for memory problems with DQMH since there will be so many duplicates of functionally the same VIs for each module? Should I definitely go 64bit if I go with DQMH? What about AF? Speaking of 64bit, I still seem to avoid it because everytime I look into it it still seems that there are lots issues and lack of feature support, is that still the case? I definitely will need a DB module or three, is the DB toolkit still unsupported for 64bit? Appreciate any thoughts on these questions and about AF/DQMH considerations when things scale up.
