The "trick" of the Producer/Consumer pattern in LabVIEW is that loops can run "in parallel", and moving data around in memory is quite fast. You have three loops: Data Acquisition loop, which generates data at regular intervals, but spends very few cycles "inside the loop" because most of the work takes place in the hardware, a Plotting loop which gets an Array of data from the Acquisition loop (and needs to plot it at least as fast as the data are being generated), and a data-saving Loop, which writes data to disk (and is governed by disk speed, generally also pretty fast.
Data Acquisition -> Data Plotting is one Procedure/Consumer loop. Data Plotting -> Data Saving is a second Producer/Consumer Loop. When the Plotting Loop gets the data from its dequeue (acting as a Consumer), it immediately puts it on a Queue to the Data Saving loop, making it a (second) Producer. Note that running the Producer or the Consumer takes very little time -- it's basically "copying data from one block of memory to another". Because it is in a Queue, no data gets "lost" or "ignored" (unlike the case with a Notifier). Note that Queues are "elastic" -- if one loop needs "a little more time" to consume, the data will "build up" in the Queue, and when the loops resume "full speed", they'll "drain" the Queue.
Bob Schor
