Hi there!
I need to run several hundred million calculations through a grasshopper recurrent script to train some AI algorithms.
I was planning on feeding the input data from a spreadsheet and put to work several instances of rhino + GH in parallel, to take advantage of all the CPU cores.
My ballpark figure would be 4 threads for windows + 4 threads for each instance in parallel: 7 instances in a 16-core ryzen or 31 in a 64-core threadripper.
Despite loosing about 30% boost y should be able to get through the calculations much faster (30sec each, means either 24 or 4.5 hours per million calculations)
Any thoughts? thanks in advance.