Hello to everyone, I am looking to parallelise within chain some code in STAN to boost computation time. Unfortunately my target function cannot be aggregated trough a simple sum, i.e `reduce_sum`

in this case does not work. I need some parallel code that returns a vector of known length `b`

and then I have to post process that to get my desired likelihood.

The main issue I am having is that to compute this likelihood I need to pass 4 auxiliary matrices to the likelihood function, say A,B,C, Omega. From these I can the compute the likelihood. It seems to me that currently `map_rect`

only accepts a pre-specified number of inputs;

```
vector map_rect((vector, vector, array[] real, array[] int):vector f,
vector phi, array[] vector thetas,
data array[,] real x_rs, data array[,] int x_is);
```

From this source it seems I can only pass one auxiliary vector `phi`

to the`map_rect`

function, is it true? Ideally I would like to have something like:

```
vector lp(matrix A, matrix B, matrix C, matrix Omega, array[] matrix Y){.....
.................
return lp'
}
```

Where my data is an array of matrices. Can I pass this type of data to `map_rect`

, or does it only work for the data types presented in the Rstan documentation?

Best,

Luca