Commit(s): pkofod/julia@a4daf3b21f1a98dd208f50bc1b785d4d92f24d3b vs JuliaLang/julia@83007fb8c4c748fb23d38759daafd77fa1d56c2b
Triggered By: link
Tag Predicate: ALL
Commit(s): pkofod/julia@a4daf3b21f1a98dd208f50bc1b785d4d92f24d3b vs JuliaLang/julia@83007fb8c4c748fb23d38759daafd77fa1d56c2b
Triggered By: link
Tag Predicate: ALL
pkm@pkm:~/.julia/v0.6/RemPiO2$ julia6 -O3 test/runtests.jl | |
Testing speed and accuracy of rempio2 in [-pi*9/4, pi*9/4] | |
---------------------------------------------------------- | |
Numbers below are elapsed time returned from @belapsed | |
Every number is checked between the two implementations | |
using a @test such that any difference results in termi- | |
nation of the program. |
julia> Profile.print() | |
11 ./event.jl:68; (::Base.REPL.##3#4{Base.REPL.REPLBackend})() | |
11 ./REPL.jl:95; macro expansion | |
11 ./REPL.jl:64; eval_user_input(::Any, ::Base.REPL.REPLBackend) | |
11 ./boot.jl:234; eval(::Module, ::Any) | |
11 ./<missing>:?; anonymous | |
11 ./profile.jl:16; macro expansion; | |
11 /home/pkm/.julia/v0.5/MDPTools/src/solution/solve.jl:39; solve!; | |
11 /home/pkm/.julia/v0.5/MDPTools/src/solution/solve.jl:40; #solve!#58; | |
11 ./<missing>:0; (::MDPTools.#kw##solve!)(::Array{Any,1}, ::MDPTools.#solve!, ::MDPTools.LinearUtility{Float64}, ::MDP... |
intersect at /home/pkm/julia/julia/src/subtype.c:1628 | |
intersect_ufirst at /home/pkm/julia/julia/src/subtype.c:1032 [inlined] | |
intersect_var at /home/pkm/julia/julia/src/subtype.c:1100 | |
intersect at /home/pkm/julia/julia/src/subtype.c:1628 | |
intersect_ufirst at /home/pkm/julia/julia/src/subtype.c:1032 [inlined] | |
intersect_var at /home/pkm/julia/julia/src/subtype.c:1100 | |
intersect at /home/pkm/julia/julia/src/subtype.c:1628 | |
intersect_ufirst at /home/pkm/julia/julia/src/subtype.c:1032 [inlined] | |
intersect_var at /home/pkm/julia/julia/src/subtype.c:1100 | |
intersect at /home/pkm/julia/julia/src/subtype.c:1628 |
using Optim, Calculus | |
# Let's try with a polynomial. It has very simple Hessian, as there are no cross | |
# products, only quadratic terms. Hence the Hessian is a diagonal matrix where | |
# diag(H) = [2, 2, ..., 2]. Let's try to see if that's what we get! | |
function large_polynomial(x::Vector) | |
res = zero(x[1]) | |
for i in 1:250 | |
res += (i - x[i])^2 | |
end | |
return res |
using BenchmarkTools | |
K = 5 | |
N = 3 | |
z = [rand(20000, K) for i =1:N] | |
P = [rand(20000) for i = 1:N] | |
cache = zeros(20000, K) | |
function test1(P, z) | |
for i in eachindex(P) | |
cache .= P[i].*z[i] |
K = 5 | |
N = 3 | |
z = [rand(20000, K) for i =1:N] | |
P = [rand(20000) for i = 1:N] | |
Pm = rand(20000, N) | |
Pv = [@view Pm[:, j] for j = 1:N] | |
cache = zeros(20000, K) | |
function test1(P, z) | |
for i in eachindex(P) |
function LearnBase.learn!(solver::CrossEntropyMethod, env::AbstractEnvironment, doanim = false) | |
# !!! INIT: | |
# this is a mappable function of θ to reward | |
cem_episode = θ -> begin | |
π = cem_policy(env, θ) | |
R, T = episode!(env, π; maxiter = solver.options[:maxiter]) | |
R | |
end |
julia> using Optim | |
julia> function rosenbrock(x) | |
N = length(x) | |
fval = 0.0 | |
for i in 1:div(N, 2) | |
fval += 100(x[2i-1]^2 - x[2i])^2 + (x[2i-1] - 1.0)^2 | |
end | |
fval | |
end |
using BenchmarkTools | |
A = rand(10000) | |
B = Vector[rand(10) for i = 1:1000] | |
C = rand(10,1000) | |
D = rand(10000) | |
E = Vector[rand(10) for i = 1:1000] | |
F = rand(10, 1000) | |
function fast!(a, c) |