Generating random functions (as opposed to random numbers)

General Usage
,
7

You may have more luck with some of the rngs from packages. RandomNumbers has the following:

julia> let rng = RandomNumbers.Xorshifts.Xoroshiro128()
           global g(r::Number) = (Random.seed!(rng, hash(r)); randn(rng))
       end
g (generic function with 1 method)

julia> g(1)
1.1158401688820527

julia> g(1)
1.1158401688820527

julia> @benchmark f(1)
BenchmarkTools.Trial:
  memory estimate:  152 bytes
  allocs estimate:  4
  --------------
  minimum time:     9.860 μs (0.00% GC)
  median time:      9.950 μs (0.00% GC)
  mean time:        10.096 μs (0.00% GC)
  maximum time:     26.180 μs (0.00% GC)
  --------------
  samples:          10000
  evals/sample:     1

julia> @benchmark g(1)
BenchmarkTools.Trial:
  memory estimate:  0 bytes
  allocs estimate:  0
  --------------
  minimum time:     7.708 ns (0.00% GC)
  median time:      7.768 ns (0.00% GC)
  mean time:        7.889 ns (0.00% GC)
  maximum time:     24.985 ns (0.00% GC)
  --------------
  samples:          10000
  evals/sample:     999

Though I’m not too familiar with the particulars of that generator…

1 Like