Home > Net >  Trying to parallelize a matrix rotation
Trying to parallelize a matrix rotation

Time:08-19

I am trying to parallelize a matrix rotation in Julia. I'm using the Distributed package, but I'm still getting an error. My original (serial) code is as follows:

using Rotations
using ImageTransformations
using CoordinateTransformations

data_arr = rand(25,25)
angle_arr = -80:0.01:80

for a in angle_arr
    trfm = recenter(Rotations.RotMatrix(deg2rad(a)),center(data_arr))
    rot_arr = warp(data_arr, trfm);
    rot_arr[rot_arr.<0].=0
end

After attempting to apply the Distributed package my code is now this:

using Distributed
@everywhere using Rotations
@everywhere using ImageTransformations
@everywhere using CoordinateTransformations

data_arr = rand(25,25)
angle_arr = -80:0.01:80

addprocs(8)

nheads = @distributed ( ) for a in angle_arr
    trfm = recenter(RotMatrix(deg2rad(a)),center(data_arr))
    rot_arr = warp(data_arr, trfm);
    rot_arr[rot_arr.<0].=0
end

which still gives me the following error. I'm a bit confused, since I thought I was defining the Rotations package containing RotMatrix everywhere. What am I missing?

ERROR: LoadError: TaskFailedException
Stacktrace:
 [1] wait
   @ .\task.jl:334 [inlined]
 [2] fetch
   @ .\task.jl:349 [inlined]
 [3] preduce(reducer::Function, f::Function, R::UnitRange{Int64})
   @ Distributed C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\macros.jl:274
 [4] top-level scope
   @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\macros.jl:286
 [5] eval
   @ .\boot.jl:373 [inlined]
     [6] include_string(mapexpr::typeof(identity), mod::Module, code::String, filename::String)
   @ Base .\loading.jl:1196

    nested task error: On worker 2:
    UndefVarError: RotMatrix not defined
    Stacktrace:
     [1] top-level scope
       @ none:1
     [2] eval
       @ .\boot.jl:373
     [3] #103
       @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\process_messages.jl:274
     [4] run_work_thunk
       @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\process_messages.jl:63
     [5] run_work_thunk
       @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\process_messages.jl:72
     [6] #96
       @ .\task.jl:423

    ...and 8 more exceptions.

    Stacktrace:
      [1] sync_end
        @ .\task.jl:381
      [2] macro expansion
        @ .\task.jl:400 [inlined]
      [3] remotecall_eval
        @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\macros.jl:219
      [4] macro expansion
        @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\macros.jl:203 [inlined]
      [5] macro expansion
        @ C:\Git\Repositories\water\src\Swamp.jl:146 [inlined]
      [6] #15
        @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\macros.jl:288
      [7] #106
        @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\process_messages.jl:278
      [8] run_work_thunk
        @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\process_messages.jl:63
      [9] macro expansion
        @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\process_messages.jl:278 [inlined]
     [10] #105
        @ .\task.jl:423
    Stacktrace:
     [1] remotecall_fetch(::Function, ::Distributed.Worker, ::Function, ::Vararg{Any}; kwargs::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
       @ Distributed C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\remotecall.jl:469
     [2] remotecall_fetch
       @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\remotecall.jl:461 [inlined]
     [3] #remotecall_fetch#158
       @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\remotecall.jl:496 [inlined]
     [4] remotecall_fetch
       @ C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\remotecall.jl:496 [inlined]
     [5] (::Distributed.var"#169#170"{typeof( ), var"#15#16", UnitRange{Int64}, Vector{UnitRange{Int64}}, Int64, Int64})()
       @ Distributed C:\Users\apung\AppData\Local\Programs\Julia-1.7.2\share\julia\stdlib\v1.7\Distributed\src\macros.jl:270
in expression starting at C:\Git\Repositories\water\src\Swamp.jl:137

CodePudding user response:

@everywhere should be called after addprocs - otherwise you have no workers to load your libraries. The workers added later have no libraries loaded.

using Distributed
addprocs(8)
@everywhere using Rotations
@everywhere using ImageTransformations
@everywhere using CoordinateTransformations

Your loop is not collecting anything - perhaps you want to put rot_arr in the last line just before end. In that case the values of rot_arr would get summed.

  • Related