According to the rankUpdate docstring:

the update is C <- α*A'A + β*C but while debugging with the penicllin data, I had the following: C is of size (6,6) and A is of size (6,24) which means the update is C <- α*AA' + β*C (note the reverse order for A and its transpose).

Am I understanding something incorrectly or should the docstring be updated?