Update sim_fixed_n() and relevant vignette

[cmansch]

Update and test the error handling of sim_fixed_n() from "pass" to "stop" to achieve parity with the behavior of sim_gs_n().

          Might be worth a brief mention in the [vignette on parallelization](https://github.com/Merck/simtrial/blob/main/vignettes/parallel.Rmd) that now `sim_gs_n()` supports the same parallelization behavior as `sim_fixed_n()`

Originally posted by @jdblischak in #249 (comment)