Running into a mysterious zero div error

[sharabhshukla]

I am getting started with the MadNLPGPU and ExaModels, I am already able to run the native ExaModels and solve nlps on a GPU. However, while using the JuMP Interface, I am running into a mysterious error,

ERROR: LoadError: DivideError: integer division error
Stacktrace:
  [1] div
    @ ./int.jl:295 [inlined]
  [2] div
    @ ./div.jl:308 [inlined]
  [3] div
    @ ./div.jl:353 [inlined]
  [4] fld
    @ ./div.jl:319 [inlined]
  [5] mod
    @ ./int.jl:287 [inlined]
  [6] #4
    @ ~/.julia/packages/KernelAbstractions/zPAn3/src/nditeration.jl:121 [inlined]
  [7] ntuple
    @ ./ntuple.jl:48 [inlined]
  [8] partition
    @ ~/.julia/packages/KernelAbstractions/zPAn3/src/nditeration.jl:119 [inlined]
  [9] partition
    @ ~/.julia/packages/KernelAbstractions/zPAn3/src/KernelAbstractions.jl:628 [inlined]
 [10] launch_config
    @ ~/.julia/packages/CUDA/htRwP/src/CUDAKernels.jl:88 [inlined]
 [11] (::KernelAbstractions.Kernel{CUDABackend, KernelAbstractions.NDIteration.DynamicSize, KernelAbstractions.NDIteration.DynamicSize, typeof(MadNLPGPU.gpu__force_lower_triangular!)})(::CuArray{Int32, 1, CUDA.Mem.DeviceBuffer}, ::Vararg{CuArray{Int32, 1, CUDA.Mem.DeviceBuffer}}; ndrange::Int64, workgroupsize::Nothing)
    @ CUDA.CUDAKernels ~/.julia/packages/CUDA/htRwP/src/CUDAKernels.jl:108
 [12] Kernel
    @ ~/.julia/packages/CUDA/htRwP/src/CUDAKernels.jl:105 [inlined]
 [13] force_lower_triangular!
    @ ~/.julia/packages/MadNLPGPU/DvKUv/src/interface.jl:314 [inlined]
 [14] create_kkt_system(::Type{MadNLP.SparseCondensedKKTSystem}, cb::MadNLP.SparseCallback{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, ExaModel{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, ExaModelsKernelAbstractions.KAExtension{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, Nothing, CuArray{Tuple{Int64, Int64}, 1, CUDA.Mem.DeviceBuffer}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, CUDABackend}, ExaModels.ObjectiveNull, ExaModels.ConstraintAug{ExaModels.ConstraintAug{ExaModels.Constraint{ExaModels.ConstraintNull, ExaModels.SIMDFunction{ExaModels.Null{Nothing}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, Int64}, ExaModels.SIMDFunction{Pair{ExaModels.ParIndexed{ExaModels.ParSource, 3}, ExaModels.Node2{typeof(*), ExaModels.Var{ExaModels.ParIndexed{ExaModels.ParSource, 1}}, ExaModels.ParIndexed{ExaModels.ParSource, 2}}}, ExaModels.Compressor{Tuple{Int64}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64, Float64, Int64}, 1, CUDA.Mem.DeviceBuffer}}, ExaModels.SIMDFunction{ExaModels.Null{Float64}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64}, 1, CUDA.Mem.DeviceBuffer}}}, MadNLP.RelaxBound, MadNLP.RelaxEquality}, ind_cons::@NamedTuple{ind_eq::CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, ind_ineq::CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, ind_fixed::CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, ind_lb::CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, ind_ub::CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, ind_llb::CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, ind_uub::CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}}, linear_solver::Type{MadNLPGPU.CUDSSSolver}; opt_linear_solver::MadNLPGPU.CudssSolverOptions, hessian_approximation::Type)
    @ MadNLP ~/.julia/packages/MadNLP/RRGPv/src/KKT/sparse.jl:480
 [15] MadNLPSolver(nlp::ExaModel{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, ExaModelsKernelAbstractions.KAExtension{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, Nothing, CuArray{Tuple{Int64, Int64}, 1, CUDA.Mem.DeviceBuffer}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, CUDABackend}, ExaModels.ObjectiveNull, ExaModels.ConstraintAug{ExaModels.ConstraintAug{ExaModels.Constraint{ExaModels.ConstraintNull, ExaModels.SIMDFunction{ExaModels.Null{Nothing}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, Int64}, ExaModels.SIMDFunction{Pair{ExaModels.ParIndexed{ExaModels.ParSource, 3}, ExaModels.Node2{typeof(*), ExaModels.Var{ExaModels.ParIndexed{ExaModels.ParSource, 1}}, ExaModels.ParIndexed{ExaModels.ParSource, 2}}}, ExaModels.Compressor{Tuple{Int64}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64, Float64, Int64}, 1, CUDA.Mem.DeviceBuffer}}, ExaModels.SIMDFunction{ExaModels.Null{Float64}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64}, 1, CUDA.Mem.DeviceBuffer}}}; kwargs::@Kwargs{})
    @ MadNLP ~/.julia/packages/MadNLP/RRGPv/src/IPM/IPM.jl:155
 [16] MadNLPSolver
    @ ~/.julia/packages/MadNLP/RRGPv/src/IPM/IPM.jl:115 [inlined]
 [17] madnlp(model::ExaModel{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, ExaModelsKernelAbstractions.KAExtension{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, Nothing, CuArray{Tuple{Int64, Int64}, 1, CUDA.Mem.DeviceBuffer}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, CUDABackend}, ExaModels.ObjectiveNull, ExaModels.ConstraintAug{ExaModels.ConstraintAug{ExaModels.Constraint{ExaModels.ConstraintNull, ExaModels.SIMDFunction{ExaModels.Null{Nothing}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, Int64}, ExaModels.SIMDFunction{Pair{ExaModels.ParIndexed{ExaModels.ParSource, 3}, ExaModels.Node2{typeof(*), ExaModels.Var{ExaModels.ParIndexed{ExaModels.ParSource, 1}}, ExaModels.ParIndexed{ExaModels.ParSource, 2}}}, ExaModels.Compressor{Tuple{Int64}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64, Float64, Int64}, 1, CUDA.Mem.DeviceBuffer}}, ExaModels.SIMDFunction{ExaModels.Null{Float64}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64}, 1, CUDA.Mem.DeviceBuffer}}}; kwargs::@Kwargs{})
    @ MadNLP ~/.julia/packages/MadNLP/RRGPv/src/IPM/solver.jl:10
 [18] madnlp(model::ExaModel{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, ExaModelsKernelAbstractions.KAExtension{Float64, CuArray{Float64, 1, CUDA.Mem.DeviceBuffer}, Nothing, CuArray{Tuple{Int64, Int64}, 1, CUDA.Mem.DeviceBuffer}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, CUDABackend}, ExaModels.ObjectiveNull, ExaModels.ConstraintAug{ExaModels.ConstraintAug{ExaModels.Constraint{ExaModels.ConstraintNull, ExaModels.SIMDFunction{ExaModels.Null{Nothing}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Int64, 1, CUDA.Mem.DeviceBuffer}, Int64}, ExaModels.SIMDFunction{Pair{ExaModels.ParIndexed{ExaModels.ParSource, 3}, ExaModels.Node2{typeof(*), ExaModels.Var{ExaModels.ParIndexed{ExaModels.ParSource, 1}}, ExaModels.ParIndexed{ExaModels.ParSource, 2}}}, ExaModels.Compressor{Tuple{Int64}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64, Float64, Int64}, 1, CUDA.Mem.DeviceBuffer}}, ExaModels.SIMDFunction{ExaModels.Null{Float64}, ExaModels.Compressor{Tuple{}}, ExaModels.Compressor{Tuple{}}}, CuArray{Tuple{Int64}, 1, CUDA.Mem.DeviceBuffer}}})
    @ MadNLP ~/.julia/packages/MadNLP/RRGPv/src/IPM/solver.jl:9
 [19] top-level scope
    @ ~/julianlpcuda/test_simple_nlp.jl:11
in expression starting at /home/ubuntu/julianlpcuda/test_simple_nlp.jl:11

Here is the script I am running,

using MadNLP, JuMP, MadNLPGPU, CUDA, ExaModels


model = Model()
@variable(model, x >= 0)
@variable(model, 0 <= y <= 3)
@NLobjective(model, Min, 12x + 20y)
@constraint(model, c1, 6x + 8y >= 100)
@constraint(model, c2, 7x + 12y >= 120)
em = ExaModel(model; backend = CUDABackend())
result = madnlp(em)

[sshin23]

Thanks for reporting this; this seems to be a bug. We'll work on the fix

[sshin23]

We have fixed this in #77 and MadNLP/MadNLP.jl#326. We'll do a patch release in the next few days.

A side note: we currently do not support @NL.... Please try @objective

[sharabhshukla]

@sshin23 , Thank you for the prompt response. Looking forward to the new release

[sharabhshukla]

Ok, I tried objective instead of NL, still the same error

[sshin23]

@sharabhshukla It will work once MadNLP/MadNLP.jl#326 is merged and a new version is released