plan(multisession, workers= a lot!) seed, simChef

[fdavallejos]

Hello everyone, I need help! I have to perform a simulation of an experimental study with a factorial design in which I will have 9 simulation scenarios. My computer has 16 cores, but I'm not sure if it's possible to use the multisesion option so that the specified random seed allows, for example, the simulation on core 1 to end at 5000 and the simulation on core 2 to start at 5001. I'm not sure if this is clear... Basically, I would like to understand the relationship between simulations on different cores. As stated in Tim Morris's article "Using simulation studies to evaluate statistical methods" (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6492164/pdf/SIM-38-2074.pdf), the implication for parallelization is that, while runs for different data-generating mechanisms may be parallelized, it is inadvisable to parallelize repetitions within a specific data-generating mechanism. I am conducting the simulation study with the help of "simChef". Could you tell me how to approach this condition so that my simulation study is replicable? Thanks!!! I'm really stuck on this!!
SIM-38-2074.pdf