Error with vectorizing truncated likelihood

Modeling
1

I get the following error when I try to compile the vectorized truncated data example from section 4.2 of the manual

data {
  int<lower=0> N;
  real U;
  array[N] real<upper=U> y;
}
parameters {
  real mu;
  real<lower=0> sigma;
}
model {
  y ~ normal(mu, sigma) T[ , U];
}

Error: “Outcomes in truncated distributions must be univariate.”

This is with rstan version 2.26.22.

Does anyone know why this is happening? Is there a more recent version of rstan that I should be using? Thanks!

2

Vectorized truncations did not become available until (I believe) 2.31

3

Thank you, yeah I was able to switch to cmdstanr and version 2.32 and then it worked.

For a lognormal model with around 20k y values the vectorized 2.3 code ran in about 30 minutes on my laptop vs ~45 min for the non-vectorized version. Still significantly slower than the non-truncated version which ran in a few minutes.