Proc Corr with large data set

I need to run Proc Corr on a large dataset. Let's say dimensions of 500x5000. Any machine I can run this one runs out of memory. Now, I could manually create a bunch of lists of subsets of the data and use them in a WITH statement (running proc corr several times and then appending all the results together), but that is tiresome and not likely to be feasible in the long run for this project.

Any idea how I can automate this process? An array would be nice, but my reading indicates that they are data set specific and not for use in proc statements.

Thanks in advance. Any help is appreciated.