Great job with the facilities for Monte Carlo Simulations.
I have a couple of question about EnsembleProblem
. From the docs, it follows that the data from all the trajectories is saved. Is it possible to use Welford’s online algorithm to accumulate the data on the go (each batch would have its own accumulator)?
It is also convenient to be able to save to file accumulated results after each n
trajectories. This way, the data can be analyzed in the background as the simulation still goes on. So, is it possible to do it?