Something wrong with my memory calculation

I’m coding a NeuralODE in Julia and tried to get a memory use data from the program. On python, when I change my adjoint method I get different but consistent (by method) results for my max RSS but on Julia I always get the same value, here is how I did it :

function meminfo_julia()
    @printf "GC live:   %9.3f MiB\n" Base.gc_live_bytes()/2^20
    @printf "JIT:       %9.3f MiB\n" Base.jit_total_bytes()/2^20
    @printf "Max. RSS:  %9.3f MiB\n" Sys.maxrss()/2^20

my main function : 
for i in 1:N
    global p = modele(strategie[i],strategie_length[i],strategie_iter[i],strategie_rate[i],p,st)


The main function do respond correctly to the variables as for different solving/adjoint algorithms it gives me different values of loss and precision.
Does someone have any idea of what is wrong ?