I'd like to know why it is considered as so difficult to generate random numbers when using parallel computing like multi threading or the GPU? Wouldn't it be for instance possible to just start a bunch of threads, give everyone a seed and call it a day?
I know that the topic of RNG itself is considered a big challenge in computer science but why does it get even more difficult when working in a parallel computing environment? Could anyone summarize some reasons?
Sources I consumed which hint that it's not trivial:
https://www.youtube.com/watch?v=qO8FAQvlgX0&t=1498s https://www.iro.umontreal.ca/~lecuyer/myftp/papers/parallel-rng-imacs.pdf
Aucun commentaire:
Enregistrer un commentaire