I have made a simple model that constructs a symmetric Jacobian for a linear ODE system, and applies
eigenvectors_sym, for a 164 dimensional system. The expressions are very simple, yet memory usage saturates immediately: top shows 44 GB allocated on a 16 GB system… the input Rdump file is 400KB.
Are there known memory issues with these functions? For reference, I’ve posted the code on this gist.
edit I checked the scaling of memory usage with problem size, and I’m sure I can’t use this as is. Since I only need leading eigenvectors and values, I will find another way to compute them.