They are the same in terms of algorithm and the way you sample from prior distribution and pass through deterministic function. The differences is that forward sampling returns samples of the latent variables and the (would be) observed y, and prior predictive only returns the observed y.
2 Likes