Properly Running Chains in Parallel

[Markian Pahuta]

Hi

I just want to make sure that I will be running 3 independent chains with the following code (I'm using Linux). Specifically, am I defining rng_seed OK?

library(parallel)
library(rstan)

setwd('~/ions/')

load('Co_data_oct16.dat')

foo <- stan(file = 'oct16.stan', data = datalist, chains=0)
rng_seed <-12345
sflist <- mclapply(
          1:3,
          mc.cores = 3,
           function(i) stan(fit = foo,
                            data = datalist,
                            iter=5000,
                            thin=5,
                            seed = rng_seed,
                            chains = 1,
                            chain_id = i,
                            refresh = -1))

modfit <- sflist2stanfit(sflist)

thanks

Mark

[Bob Carpenter]

Yes, the key is having different chain_id. This advances
the random number generator a very big step (more than you'd
have time to genrate), thus making the draws in the different chains
come from different slices of the pseudo-random number sequence
determined by the seed.

This works the same way in the other interfaces, too.

- Bob

> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.