Well, you can use an array of clusters, where each cluster holds a 1D array. I think this would take less memory but it would definitly be more complex becuase you're adding the cluster operations. To do this, create a 1D array, drag it into a cluster and drag that cluster into another array. It's possible that this isn't any better. You can measure the performance difference by using VI profiling (Tools>>Advanced) and by using timing (the time the VI stopped minus the time it started).
___________________
Try to take over the world!