Víctor Alejandro Gil Sepúlveda

This does not work: N = len(elements) new_matrix = CondensedMatrix([1.]_((N_(N-1)/2))) for i in range(len(elements)-1): e_i = elements[i] for j in range(i+1,len(elements)): e_j = elements[j] new_matrix[i, j] = old_matrix[e_i, e_j] return...

matrixOneVsFollowing\* were added but only implemented by CUDA calculators. The need to be reimplemented to improve performance with those calculators that allow the calculation of RMSD without calculating the actual...

- Improving Jacobian performance (or solve analitically) and improve overall memory consumption by using short ints.

Currently pickling can only be used by registering the pickling function in client code with copy_reg (http://docs.python.org/2/library/copy_reg.html) See http://stackoverflow.com/questions/12530333/python-pickling-c-objects (defining the **reduce** or pickling needed methods)