wds15

And one more thing: Current partial log link functions as used with the cmdstanr backend do prevent the `SoA` optimisation due to `for`looping. A - hopefully - simple change for...

Thanks @paul-buerkner for trying out the array index thing. I am surprised to slows things down... will look into it once more myself. I know that Steve worked on it...

HI! I can confirm that constrained types are a problem as it seems. Also arrays of matrices cause problems as it looks. Here is an example: ```R library(OncoBayes2) example_model("combo2") ##...

Not sure if the Student-t is still being considered to be included as a fully supported random effects distribution, but what may help is the recent availability of a multi_student_t...

just one thought... as a - very dirty - hack... could one possibly write a function which temporarily redefines the `readLines` function? Maybe the packages doing mocking could help here?...

I just tried this approach and I got amazingly far. The culprit is that it's not just about swapping out `readLines` (which really works line-by-line yack): ```r library(rstan) rf

Ok, I finally solved the mystery here and the good news is that everything is exactly reproducibile, but the behavior is inconsistent overall. So what happened is that the extract...

We should at least add to the documentation of `read_stan_csv` that the random number gen state matters when calling the read_stan_csv function. The much better solution would be to extract...

Thanks for reporting. Will need to reproduce this with Docker to be able to debug it.

Hmmm... right now I cannot reproduce the issue. It actually seems to be a problem with the ubuntu RcppParallel package. What I did is to install ubuntu 20.10 in a...