ExaCA

An exascale-capable cellular automaton for nucleation and grain growth

ExaCA is a cellular automata (CA) code for grain growth under additive manufacturing conditions by ExaAM within the Exascale Computing Project.

Build

ExaCA-Kokkos uses Kokkos and MPI for parallelism.

Dependencies

Dependency	Version	Required	Details
CMake	3.11+	Yes	Build system
Kokkos	3.2+	Yes	Provides portable on-node parallelism.
MPI	GPU Aware if CUDA/HIP Enabled	Yes	Message Passing Interface
nlohmann_json	3.10+	Yes	Input parsing
GoogleTest	1.10+	No	Unit test framework
CUDA	9+	No	Programming model for NVIDIA GPUs
HIP	3.5+	No	Programming model for AMD GPUs

Kokkos and MPI are available on many systems; if not, obtain the desired versions:

git clone https://github.com/kokkos/kokkos.git --branch 3.4.00

Backend options

Note that ExaCA runs with the default enabled Kokkos backend (see https://github.com/kokkos/kokkos/wiki/Initialization).

ExaCA has been tested with Serial, OpenMP, Pthreads, CUDA, and HIP backends.

Build CPU

First, if Kokkos is not already built on your system, build Kokkos:

# Change this path to Kokkos source
cd ./kokkos
mkdir build
cd build
# Check the CPU architecture flag matches the hardware
cmake \
  -D CMAKE_BUILD_TYPE="Release" \
  -D CMAKE_INSTALL_PREFIX=install \
  -D Kokkos_ENABLE_OPENMP=ON \
  -D Kokkos_ARCH_POWER9=ON \
  .. ;
make install
cd ../..

Note that there are other host backends available. The Kokkos architecture flag must match the hardware you run on and will improve performance, if used.

Then build ExaCA, including the path to the Kokkos build:

# Change this path to Kokkos installation
export KOKKOS_INSTALL_DIR=./kokkos/build/install

# Change this path to ExaCA source
cd ./ExaCA
mkdir build
cd build
cmake \
  -D CMAKE_BUILD_TYPE="Release" \
  -D CMAKE_PREFIX_PATH=$KOKKOS_INSTALL_DIR \
  -D CMAKE_INSTALL_PREFIX=install \
  ..;
make install
cd ../..

Build CUDA

If running on NVIDIA GPUs, build Kokkos with additional inputs:

# Change this path to Kokkos source
cd ./kokkos
mkdir build
cd build
# Check the GPU architecture flag matches the hardware
cmake \
  -D CMAKE_BUILD_TYPE="Release" \
  -D CMAKE_INSTALL_PREFIX=install \
  -D Kokkos_ENABLE_CUDA=ON \
  -D Kokkos_ENABLE_CUDA_LAMBDA=ON \
  -D Kokkos_ARCH_VOLTA70=ON \
  .. ;
make install
cd ../..

Note the two flags needed for the Kokkos::Cuda backend. The Kokkos architecture flag must match the hardware you run on and will improve performance. By default, the host will use Kokkos::Serial; other parallel host backends can also be used, e.g. by adding -D Kokkos_ENABLE_OPENMP.

Building ExaCA with Kokkos CUDA is identical to the OpenMP example above (Kokkos automatically uses nvcc internally as needed).

Build HIP

Again, first build Kokkos, this time with the hipcc compiler:

cd ./kokkos
mkdir build
cd build
cmake \
    -D CMAKE_BUILD_TYPE="Release" \
    -D CMAKE_CXX_COMPILER=hipcc \
    -D CMAKE_INSTALL_PREFIX=install \
    -D Kokkos_ENABLE_HIP=ON \
    -D Kokkos_ARCH_VEGA908=ON \
    .. ;
make install

And build ExaCA, where the only difference from above is the hipcc compiler:

# Change this path to Kokkos installation
export KOKKOS_INSTALL_DIR=./kokkos/build/install

cd ./ExaCA
mkdir build
cd build
cmake \
    -D CMAKE_BUILD_TYPE="Release" \
    -D CMAKE_CXX_COMPILER=hipcc \
    -D CMAKE_PREFIX_PATH="$KOKKOS_INSTALL_DIR" \
    -D CMAKE_INSTALL_PREFIX=install \
    .. ;
make install

Test

Unit tests can be run if the ExaCA_ENABLE_TESTING CMake option is enabled by running ctest in the build directory (this requires the GoogleTest framework). Tests are automatically generated for the enabled Kokkos backend.

Run

ExaCA-Kokkos runs using an input file, passed on the command line. Example problems are provided in the examples/ directory - A separate README file located in the examples/ directory goes into more detail on the problem types, the optional and required arguments needed for each problem type, and additional files used by ExaCA. The example input files present in this repository are:

Inp_DirSolidification.json: simulates grain growth from a surface with a fixed thermal gradient and cooling rate
Inp_SmallDirSolidification.json: a smaller and simpler version of the previous
Inp_SpotMelt.json: simulates overlapping spot melts with fixed a fixed thermal gradient and cooling rate
Inp_SmallSpotMelt.json: a smaller and simpler version of the previous

Example problems only possible with external data (available via https://github.com/LLNL/ExaCA-Data):

Inp_SingleLine.json: simulates melting and solidification of a single line of melt pool data
Inp_TwoLineTwoLayer.json: simulates two layers consisting of segments of two overlapping melt pools

Run by calling the created executable with an ExaCA input file:

mpiexec -n 1 ./build/install/bin/ExaCA-Kokkos examples/Inp_DirSolidification.json

Automated input file generation using Tasmanian (https://tasmanian.ornl.gov/)

Within the utilities directory, an example python script for the generation of an ensemble of input files is available. By running the example script TasmanianTest.py, 69 ExaCA input files are generated with a range of heterogenous nucleation density, mean nucleation undercooling, and mean substrate grain size values, based on the ranges in python code (N0Min-N0Max, dTNMin-dTNMax, and S0Min-S0Max), respectively. Running the python script from the ExaCA source directory, via the command

python utilities/TasmanianTest.py PathToTemperatureFile1 PathToTemperatureFile2 ...

the script will generate an ensemble of input files in the examples directory, for a series of simulations that will use the thermal history or histories described in PathToTemperatureFile1(s) being repeated for a certain number of layers (56 in this example). If a simulation repeating multiple thermal histories is desired (for example, and even layer and an odd layer scan pattern), both paths to/file names of the thermal history data should be given on the command line. Running this code will generate N = 1 to 69 input files named examples/Inp_TasmanianTest_[N].json. Other CA inputs, such as the time step or cell size, must be adjusted manually inside of the python script. Separate instances of ExaCA can be run with each ensemble member to probe microstructure dependency on nucleation and substrate.

Output and post-processing analysis

If the "Print file of grain misorientations" option is turned on within an input file, ExaCA will output a scalar field "Angle_z" as a vtk file ending with "Misorientations.vtk". Angle_z corresponds to the orientation (in degrees) of a given grain relative to the positive Z direction in a simulation (the thermal gradient direction for directional solidification problems, the build/layer offset direction for other problems). Epitaxial grains (from the initial grain structure or powder layer) are assigned values between 0 and 62.7, while nucleated grains (not present in the initial grain structure) are assigned values between 100 and 162.7 (the offset of 100 is simply used to ensure the two types of grains are differentiated, but a nucleated grain with Angle_z = 135 actually has a misorientation of 35 degrees).

If the "Print Paraview vtk file" option is turned on within an input file, post-processing can be performed on the output data set. This functionality is a separate executable from ExaCA, located in the analysis/ directory and is linked to the ExaCA library for input utilities.

Specifying debug check options can be done to print various ExaCA data fields to files following simulation initialization. The "reduced" debug check will print "CritTimeStep" (the time step at which each cell goes below the liquidus for the final time), "LayerID" (the layer associated with each cell going below the liquidus for the final time, with layer 0 being the first layer, and -1 for all cells that did not undergo solidification) and "CellType" (integers corrsponding to cell types specified in src/CAtypes.hpp). The "extensive" debug check will, in addition to the "reduced" data fields, also print "UndercoolingChange" (the rate at which a cell cools per time step after reaching its "CritTimeStep" value), "UndercoolingCurrent" (the initial undercooling of each cell), and "Melted" (1 for cells that are part of the melt pool, 0 for cells that are not).

ExaCA can optionally print the system state at intermediate time values as part of a series of vtk files that can be read by Paraview to make animations, if the "Print intermediate output frames" option is turned on. "Increment to separate frames" is the separation between intermediate output files in microseconds - if there is a long time period between solidification events (such as two overlapping melt pools formed via line scan with a long dwell time between them), setting "Intermediate output even if system is unchanged from previous state" to off will skip printing of those files.

Running ExaCA for the test problem Inp_DirSolidification.txt yields the output files TestProblemDirS.vtk (containing LayerID, GrainID, and Melted data) and TestProblemDirS.json (containing information regarding the simulation parameters used, simulation dimensions, and some timing data). To analyze this data, run grain_analysis (installed in the same location as ExaCA-Kokkos), with two command line arguments: the first being the path to/name of the analysis input file, and the second being the path to and filename (excluding extensions) of the .vtk and .json files associated with the data set of interest.

./build/install/bin/grain_analysis analysis/examples/AnalyzeDirS.json TestProblemDirS

Within the analysis/examples directory, there are example analysis input files. Note that the microstructure data files TestProblemDirS.vtk and TestProblemDirS.json must both be in the location given on the command line.

The analysis executable, in addition to outputting grain statistics, can also output files that can be further post-processing in Matlab using the MTEX toolbox to generate pole figures, inverse pole figures, and inverse pole figure-colored cross-sections. More details on this are provided in analysis/README.md

Citing ExaCA

If you use ExaCA in your work, please cite the following paper:

@article{rolchigo2022,
  title = {ExaCA: A performance portable exascale cellular automata application for alloy solidification modeling},
  journal = {Computational Materials Science},
  volume = {214},
  pages = {111692},
  year = {2022},
  issn = {0927-0256},
  doi = {https://doi.org/10.1016/j.commatsci.2022.111692},
  url = {https://www.sciencedirect.com/science/article/pii/S0927025622004189},
  author = {Matt Rolchigo and Samuel Temple Reeve and Benjamin Stump and Gerald L. Knapp and John Coleman and Alex Plotkowski and James Belak},
}

If you would like to cite the software itself, cite the current release or version used from Zenodo.

Contributing

We encourage you to contribute to ExaCA. Please check the contribution guidelines.

License

ExaCA is distributed under an MIT license.

Release

LLNL-CODE-821827

nncarlson / ExaCA