Ito process as numerical solution of stochastic differential equation

I’m about to submit an issue but since I’m getting zero feedback from discourse Performance of stochastic volatility model I’m hesitating to submit a PR. Currently my main concern is

  • This implementation could be completely written in Stan lang, is it still worth to add it as a function.
  • The Wiener process generation asks for a potentially large matrix/vector of iid normals, how’s this going to affect the performance in MCMC?