malloc() error while running EBSDMeshGenerator

[ashishdhole]

Hello,
I am trying to run a grain growth simulation using EBSD file that I generated from DREAM3D. I am getting the following error

mpirun -np 4 pfubc-opt -i ebsd.i 


*** Warning, This code is deprecated and will be removed in future versions:
Please update your main.C to adapt new main function in MOOSE framework, see'test/src/main.C in MOOSE as an example of moose::main()'. 

malloc(): invalid size (unsorted)
malloc(): invalid size (unsorted)

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 182467 RUNNING AT ashish-Inspiron-14-5430
=   EXIT CODE: 134
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
YOUR APPLICATION TERMINATED WITH THE EXIT STRING: Aborted (signal 6)
This typically refers to a problem with your application.
Please see the FAQ page for debugging suggestions

Following are the files for the reference.
output_data.txt
valgrind.txt
my input file is

[Mesh]
  [file]
    type = EBSDMeshGenerator
    filename = output_data.txt
    #pre_refine = 2
  []
[]
[GlobalParams]
  op_num = 15
  var_name_base = gr
[]

[UserObjects]
  [ebsd_reader]
    type = EBSDReader
  []
  [ebsd]
    type = PolycrystalEBSD
    coloring_algorithm = jp
    ebsd_reader = ebsd_reader
    enable_var_coloring = true
  []
  [grain_tracker]
    type = GrainTracker
    flood_entity_type = ELEMENTAL
    compute_halo_maps = true # For displaying HALO fields
    polycrystal_ic_uo = ebsd
    compute_var_to_feature_map = true
  []
[]

[ICs]
  [PolycrystalICs]
    [PolycrystalColoringIC]
      polycrystal_ic_uo = ebsd
    []
  []
[]

[Variables]
  [PolycrystalVariables]
  []
[]

[AuxVariables]
  [bnds]
  []
  [unique_grains_ic]
    order = CONSTANT
    family = MONOMIAL
  []
  [unique_grains]
    order = CONSTANT
    family = MONOMIAL
  []
  [ghost_elements]
    order = CONSTANT
    family = MONOMIAL
  []
  [halos]
    order = CONSTANT
    family = MONOMIAL
  []
  [var_indices_ic]
    order = CONSTANT
    family = MONOMIAL
  []
  [var_indices]
    order = CONSTANT
    family = MONOMIAL
  []
  [ebsd_grains]
    family = MONOMIAL
    order = CONSTANT
  []
[]

[Kernels]
  [PolycrystalKernel]
  []
[]

[AuxKernels]
  [BndsCalc]
    type = BndsCalcAux
    variable = bnds
    execute_on = 'initial timestep_end'
  []
  [ghost_elements]
    type = FeatureFloodCountAux
    variable = ghost_elements
    field_display = GHOSTED_ENTITIES
    execute_on = 'initial timestep_end'
    flood_counter = grain_tracker
  []
  [halos]
    type = FeatureFloodCountAux
    variable = halos
    field_display = HALOS
    execute_on = 'initial timestep_end'
    flood_counter = grain_tracker
  []
  [var_indices_ic]
    type = FeatureFloodCountAux
    variable = var_indices_ic
    execute_on = 'initial'
    flood_counter = ebsd
    field_display = VARIABLE_COLORING
  []
  [unique_grains_ic]
    type = FeatureFloodCountAux
    variable = unique_grains_ic
    execute_on = 'initial'
    flood_counter = ebsd
    field_display = UNIQUE_REGION
  []
  [var_indices]
    type = FeatureFloodCountAux
    variable = var_indices
    execute_on = 'initial timestep_end'
    flood_counter = grain_tracker
    field_display = VARIABLE_COLORING
  []
  [unique_grains]
    type = FeatureFloodCountAux
    variable = unique_grains
    execute_on = 'initial timestep_end'
    flood_counter = grain_tracker
    field_display = UNIQUE_REGION
  []
  [grain_aux]
    type = EBSDReaderPointDataAux
    variable = ebsd_grains
    ebsd_reader = ebsd_reader
    data_name = 'feature_id'
    execute_on = 'initial timestep_end'
  []
[]

[Modules]
  [PhaseField]
    [EulerAngles2RGB]
      crystal_structure = cubic
      euler_angle_provider = ebsd_reader
      grain_tracker = grain_tracker
    []
  []
[]

[Materials]
  [Copper]
    # T = 500 # K
    type = GBEvolution
    T = 500
    wGB = 0.6 # um
    GBmob0 = 2.5e-6 # m^4/(Js) from Schoenfelder 1997
    Q = 0.23 # Migration energy in eV
    GBenergy = 0.708 # GB energy in J/m^2
    molar_volume = 7.11e-6 # Molar volume in m^3/mol
    length_scale = 1.0e-6
    time_scale = 1.0e-6
    #ebsd_reader = ebsd_reader
    #grain_tracker = grain_tracker
  []
  [GB_type]
    # The new developed Miso Bnds Aux Kernel
    type = ComputeGBMisorientationAngle
    ebsd_reader = ebsd_reader
    grain_tracker = grain_tracker
    output_properties = 'gb_type'
    outputs = exodus
  []
[]

[Postprocessors]
  [dt]
    type = TimestepSize
  []
  [n_elements]
    type = NumElements
    execute_on = 'initial timestep_end'
  []
  [n_nodes]
    type = NumNodes
    execute_on = 'initial timestep_end'
  []
  [DOFs]
    type = NumDOFs
  []
[]

[Executioner]
  type = Transient
  scheme = bdf2
  solve_type = PJFNK #Preconditioned JFNK (default)
  petsc_options_iname = '-pc_type'
  petsc_options_value = 'lu'
  #petsc_options_iname = '-pc_type -pc_hypre_type -snes_type'
  #petsc_options_value = 'hypre    boomeramg      vinewtonrsls'
  steady_state_detection = true
  steady_state_start_time = 0
  steady_state_tolerance = 1e-8
  l_abs_tol = 1.0e-3
  nl_abs_tol = 1.0e-3
  l_tol = 1.0e-3
  l_max_its = 5
  nl_max_its = 5
  nl_rel_tol = 1.0e-3
  start_time = 0.0
  #num_steps = 200
  end_time = 149.973
  dtmin = 1e-20
  #dt = 1

  [TimeStepper]
    type = SolutionTimeAdaptiveDT
    dt = 1
  []

  #[Adaptivity]
  #  initial_adaptivity = 2
  #  refine_fraction = 0.7
  #  coarsen_fraction = 0.1
  #  max_h_level = 2
  #[]
[]

[Outputs]
  file_base = 'poly_out/fp_0p0_no_ani'
  exodus = true
  checkpoint = truee website
  perf_graph = true
[]

can you tell me what is going on and why it is failing. when I use the example ebsd file (IN100_120x120.txt)on the moose website the simulation works perfectly fine.
Thank you

[GiudGiud]

Hello

You are on the right track with using valgrind to find the bug, but you will want to use a debug version of the code.
So instead of make -j6, use METHOD=dbg make -j6 then use the newly created dbg executable

[ashishdhole]

This is what I got with the debug executable.
valgrind _dbg.txt

[GiudGiud]

hello

is this file in Moose? I can't find ComputeGBMisorientationAngle

 ==227349==    at 0x90F153C: ComputeGBMisorientationAngle::getMisorientationAngles() (ComputeGBMisorientationAngle.C:234)

[ashishdhole]

No. I made that material to calculate the misorientation angle derived from ComputeGBMisorientationType. But the error is still there with this material. But input file runs fine with the example ebsd file.
Following is the valgrind output with ComputeGBMisorientationType.
valgrind_dbg_1.txt

[GiudGiud]

looking at the code, it fails in this last line

  const auto grain_num = _ebsd_reader.getGrainNum();
  _euler_angle.resize(grain_num);
  _quat_angle.resize(grain_num);

  // Get Euler Angle of Orientation
  for (const auto i : make_range(grain_num))
  {
    auto grain_id = _ebsd_reader.getFeatureID(i);
    _euler_angle[grain_id] = _ebsd_reader.getEulerAngles(i);

you can debug further by printing the size of the arrays.
My guess is that there are more grains specified in the reader than there are Euler angles specified. So we are reading out of bounds

[GiudGiud]

I cannot reproduce this on my machine.
I am using the output_data.txt file and the ComputeGBMisorientationType (since ComputeGBMisorientationAngle is custom) and the error does not occur, even in debug mode.
I printed and we are not accessing out of bounds.

I will add an assert to that code and you can see if debug mode hits the assertion on your machine. This could be a clue
otherwise please print graid_id and grain_num in the loop where it crashes and see if it goes wrong

[ashishdhole]

Hello Guillaume,
Sorry for not updating you on this. I think I figured it out what was the issue here. In the EBSD file (output_data.txt) generated by Dream3d, the grainID was indexed starting from 1 instead of 0. When I changed that, the simulations ran smoothly.
It was a very small change.
Thank you again for help.
Ashish

[GiudGiud]

thanks for reporting back. I made a note of it and we'll catch it properly in the future