Clean up redundant memory space usage in python
Created by: dyang37
Recently I've been running large scale CT problem, and constantly run into out of memory issue on Brown cluster (max allowed memory space 91GB). As a result, we need to clean up unnecessary memory usage in both python and C code. The memory usage issue in C code is partially fixed with PR #50 (see issue #45 (closed) for more details). Issue #4 (closed) also seems to be relevant. For python code, we need to run a large-scale recon problem, calculate roughly how much memory the problem should be taking given the image, sino and matrix sizes. If we see the memory usage grow substantially beyond that, then we need to pinpoint where that's occurring and see if it can be worked around.