start julia with --threads=auto, check with Threads.nthreads()
import Pkg
Pkg.add(["Turing", "StatsPlots", "CSV", "JSON", "DataFrames", "PDMats", "LogExpFunctions"])advanced read:
- http://hakank.org/julia/turing/ or https://github.com/hakank/hakank/tree/master/julia/turing
- https://storopoli.io/Bayesian-Julia/
docs:
distributes sapling maybe faster than threaded sampling: https://discourse.julialang.org/t/trying-to-understand-parallel-performance-in-turing/74963/6