erratic perf using multithreading #18
Description
This is a duplicate from counter perf in ImageFiltering.
using ImageFiltering, BenchmarkTools, ComputationalResources
Test program to reproduce
using ImageFiltering, BenchmarkTools, ComputationalResources
function test_imfilter(; opt=1)
gk = Kernel.gaussian(2.0f0)
I = rand(Float32,1000 ,1000)
println("julia_num_threads=$(Threads.nthreads())")
for i = 1:9
start = time()
if opt == 1
J = imfilter(CPU1(Algorithm.FIR()), I, gk)
elseif opt == 2
J = imfilter(I, gk)
else
error("Wrong opt value $opt, either 1 (for CPU1), 2 CPUThreads")
end
t = round(time() - start, digits=6)
println("$i imfilter $t sec")
end
end
Issue occurs when multithreading using ImageView or Gtk
julia> include("test_imfilter.jl")
test_imfilter (generic function with 1 method)
julia> test_imfilter(opt=1)
julia_num_threads=16
1 imfilter 0.020194 sec
2 imfilter 0.028331 sec
3 imfilter 0.020721 sec
4 imfilter 0.020388 sec
5 imfilter 0.02025 sec
6 imfilter 0.022854 sec
7 imfilter 0.016506 sec
8 imfilter 0.017528 sec
9 imfilter 0.017142 sec
julia> test_imfilter(opt=2)
julia_num_threads=16
1 imfilter 0.279932 sec
2 imfilter 0.006712 sec
3 imfilter 0.006302 sec
4 imfilter 0.011404 sec
5 imfilter 0.003632 sec
6 imfilter 0.004423 sec
7 imfilter 0.002673 sec
8 imfilter 0.002693 sec
9 imfilter 0.002655 sec
julia> test_imfilter(opt=2)
julia_num_threads=16
1 imfilter 0.018668 sec
2 imfilter 0.007212 sec
3 imfilter 0.007689 sec
4 imfilter 0.005306 sec
5 imfilter 0.004047 sec
6 imfilter 0.003786 sec
7 imfilter 0.006157 sec
8 imfilter 0.003978 sec
9 imfilter 0.003345 sec
julia> using ImageView
Gtk-Message: 09:00:34.274: Failed to load module "canberra-gtk-module"
Gtk-Message: 09:00:34.274: Failed to load module "canberra-gtk-module"
julia> test_imfilter(opt=2)
julia_num_threads=16
1 imfilter 5.009421 sec
2 imfilter 5.017527 sec
3 imfilter 4.156348 sec
4 imfilter 0.012405 sec
5 imfilter 5.022431 sec
6 imfilter 0.008102 sec
7 imfilter 0.00735 sec
8 imfilter 5.013069 sec
9 imfilter 5.016844 sec
julia> test_imfilter(opt=1)
julia_num_threads=16
1 imfilter 0.03658 sec
2 imfilter 0.01693 sec
3 imfilter 0.01708 sec
4 imfilter 0.017126 sec
5 imfilter 0.019604 sec
6 imfilter 0.019881 sec
7 imfilter 0.016709 sec
8 imfilter 0.017145 sec
9 imfilter 0.016654 sec
In this part, see interrupt trace below
Julia> test_imfilter(opt=2)
julia_num_threads=16
1 imfilter 5.01519 sec
2 imfilter 1.482307 sec
^CERROR: InterruptException:
Stacktrace:
[1] try_yieldto(::typeof(Base.ensure_rescheduled), ::Base.RefValue{Task}) at ./task.jl:645
[2] wait() at ./task.jl:710
[3] wait(::Base.GenericCondition{Base.Threads.SpinLock}) at ./condition.jl:106
[4] _wait(::Task) at ./task.jl:238
[5] wait(::Task) at ./task.jl:265
[6] macro expansion at ./threadingconstructs.jl:69 [inlined]
[7] _imfilter_tiled_threads!(::CPU1{ImageFiltering.Algorithm.FIRTiled{2}}, ::Array{Float32,2}, ::OffsetArrays.OffsetArray{Float32,2,Array{Float32,2}}, ::OffsetArrays.OffsetArray{Float32,2,Array{Float32,2}}, ::OffsetArrays.OffsetArray{Float32,2,Array{Float32,2}}, ::NoPad{Pad{2}}, ::Array{Tuple{UnitRange{Int64},UnitRange{Int64}},1}, ::Array{Array{Float32,2},1}) at /home/manu/.julia/packages/ImageFiltering/jhBno/src/imfilter.jl:858
...