Keng,
That is a loaded question. It assumes that there is a good way to use the build array function in terms of memory usage. Rather than build an array, the most optimal way of constructing an array is to create an array once (outside of a loop) of a know size and then replace elements or subsets of that array in a loop with a shift-register and using the "Replace Array Subset" function. This latter approach is often referred to as operating on data "in place". Building an array requires allocating and dealocating memory each time, but replacing an element or subset of an array merely changes the value of already allocated memory. The later is much faster (even more so as the data gets very large).
For further reading on the topic take a look at the
following PDF File:
LabVIEW� Performance andMemory Management
You can probably find other such documents, articles, and presentations by searching NI's website for keywords like "LabVIEW", "Performance", "Array", "Optimize" "Memory Management", etc.
Best of luck,
-Jim
