In both Julia 1.4 and 1.6, you could try:
using Pkg
Pkg.add(["RecursiveFactorization", "VegaLite", "DataFrames", "BenchmarkTools"])
using RecursiveFactorization
include(joinpath(pkgdir(RecursiveFactorization), "perf", "lu.jl"))
and share your results here.
You could also try benchmarking the particular problem you were solving between versions. Solving that problem uses RecursiveFactorization.jl.
Something else simple you can try to compare versions:
using BenchmarkTools
function mysum(x)
s = zero(eltype(x))
@simd for xᵢ ∈ x
s += xᵢ
end
s
end
x = rand(128);
@btime mysum($x)