Get a dummy ash object and bad allocation

[william-denault]

Hello,

I am using ash objects as the basis of one of my algorithms. However, I use a different procedure to estimate the mixture wieght. Is there a simple way to provide to obtain an ash object (using betahat and sdhat) without using the optimization steps (e.g. initializing it with full weight on zero).

Furthermore, we are facing some memory allocation problems when inputting a vector that contains more than 13 millions entries, mixsqp is trying to allocate very large vector (100Gb)

[pcarbo]

@william-denault In the ash function there is an option fixg = TRUE. Would that work?

For your data set with n = 13e6, what is the size of the grid?