Problem with nested loops

[ctkelley]

I am trying to get @tasks to work with a nested loop and am able to
find what I need in the docs. Herewith a MWE:

I've tried using @set but with no luck.

function looper(n = 10)
    A = rand(n, n)
    for k = 1:n
        A[:, k] .*= 2.0
        @tasks for j = k+1:n
            for i = k+1:n
                A[i, j] -= A[i, k] * A[k, j]
            end
        end
    end
end

I run Julia from a directory with only this file in it.

% julia --project=.
               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.10.2 (2024-03-01)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

julia> using OhMyThreads

julia> include("looper.jl")
looper (generic function with 2 methods)

julia> looper()
ERROR: ArgumentError: You must either indicate the desired number of chunks (n) or the target size of a chunk (size).
Stacktrace:
  [1] missing_input_err()
    @ ChunkSplitters ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:118
  [2] getchunk(itr::UnitRange{Int64}, ichunk::Int64; n::Int64, size::Int64, split::Symbol)
    @ ChunkSplitters ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:274
  [3] getchunk
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:273 [inlined]
  [4] getchunk
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:293 [inlined]
  [5] iterate
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:142 [inlined]
  [6] iterate
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:141 [inlined]
  [7] iterate
    @ ./generator.jl:44 [inlined]
  [8] collect(itr::Base.Generator{ChunkSplitters.Chunk{…}, OhMyThreads.Implementation.var"#13#21"{…}})
    @ Base ./array.jl:834
  [9] map(f::Function, A::ChunkSplitters.Chunk{UnitRange{Int64}, ChunkSplitters.FixedCount})
    @ Base ./abstractarray.jl:3313
 [10] _tmapreduce(f::Function, op::Function, Arrs::Tuple{…}, ::Type{…}, scheduler::DynamicScheduler{…}, mapreduce_kwargs::@Kwargs{…})
    @ OhMyThreads.Implementation ~/.julia/packages/OhMyThreads/N9muZ/src/implementation.jl:52
 [11] #tmapreduce#10
    @ ~/.julia/packages/OhMyThreads/N9muZ/src/implementation.jl:33 [inlined]
 [12] tmapreduce
    @ ~/.julia/packages/OhMyThreads/N9muZ/src/implementation.jl:25 [inlined]
 [13] #tforeach#70
    @ ~/.julia/packages/OhMyThreads/N9muZ/src/implementation.jl:180 [inlined]
 [14] macro expansion
    @ ~/.julia/packages/OhMyThreads/N9muZ/src/macro_impl.jl:45 [inlined]
 [15] looper(n::Int64)
    @ Main ~/OMTest/looper.jl:5
 [16] looper()
    @ Main ~/OMTest/looper.jl:2
 [17] top-level scope
    @ REPL[3]:1
Some type information was truncated. Use `show(err)` to see complete types.

julia> 

[carstenbauer]

The issue is

julia> using ChunkSplitters

julia> collect(chunks(11:10; n=6))
ERROR: ArgumentError: You must either indicate the desired number of chunks (n) or the target size of a chunk (size).
Stacktrace:
  [1] missing_input_err()
    @ ChunkSplitters ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:118
  [2] getchunk(itr::UnitRange{Int64}, ichunk::Int64; n::Int64, size::Int64, split::Symbol)
    @ ChunkSplitters ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:274
  [3] getchunk
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:273 [inlined]
  [4] getchunk
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:293 [inlined]
  [5] iterate
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:142 [inlined]
  [6] iterate
    @ ~/.julia/packages/ChunkSplitters/R5lnx/src/ChunkSplitters.jl:141 [inlined]
  [7] copyto!(dest::Vector{StepRange{Int64, Int64}}, src::ChunkSplitters.Chunk{UnitRange{Int64}, ChunkSplitters.FixedCount})
    @ Base ./abstractarray.jl:943
  [8] _collect
    @ ./array.jl:765 [inlined]
  [9] collect(itr::ChunkSplitters.Chunk{UnitRange{Int64}, ChunkSplitters.FixedCount})
    @ Base ./array.jl:759
 [10] top-level scope
    @ REPL[32]:1

It occurs because for k=n you get k+1:n == n+1:n which isempty.

Regular for loops support this case (i.e. they don't to anything) so we should as well. However, at least partially, this is an upstream issue because the error message above is clearly wrong and bad. On our side, we might consider checking !isempty(coll).

[carstenbauer]

Will be fixed by #88 (in the next release).