JuMP 2-d array Decision Variable computationally expensive

I am creating a decision variable that is a 2-d array of size 25122 x 25122. (length of arcs and od are both 25122)

Is there a more efficient way to initialize a multi-dimensional array as a decision variable? I have checked out this link but didn’t seem to find anything help for multi-dimensional JuMP arrays: Performance Tips · The Julia Language

@variable(model,f[arcs,od] >=0)

I am working on a 8 core, 64 gb. 1 tb machine and attempted to run this line of code for 8+ hours with no completion. All other decision variables are created under 3 seconds.
@variable(model,x[arcs], Bin) where arcs is of size 25122

25122x25122 is 630,000,000 variables. Do you expect to be able to solve a problem of that size?

It probably shouldn’t take 8 hours without doing anything, but you’re probably running out of RAM and swapping.