OLCF Crusher TDS

If this is your first time building MFIX-Exa on Joule2, please review the general notes below and Basics section first.

To access Summit, you will need an OLCF account and an RSA token.
The name of the MFIX-Exa account is cfd122
These instructions build MFIX-Exa on the login nodes using -j8 CPUs. You may have to decrease this value if there is high traffic or you may want to increase this value if you are on a compute node interactively.
The cmake instructions compile to a build directory. The gmake instructions compile to a exec directory.
For the dependencies, it is assumed that you have set the following environment variables:
```
export HYPRE_INSTALL_DIR=$HOME/<path/to/my/hypre-install-dir>
export CSG_INSTALL_DIR=$HOME/<path/to/my/csg-dep-install-dir>
export ASCENT_INSTALL_DIR=$HOME/<path/to/my/ascent-install-dir>
export BOOST_INSTALL_DIR=$HOME/<path/to/my/boost-install-dir>
```
to a path that you have read/write access to. You will need to recall these paths later if you want to build MFIX-Exa with the optional dependencies. Boost is only needed for PrgEnv-cray; there is a module available for PrgEnv-gnu.
After building the mfix executable (with cmake), you can build the PIC-to-DEM restarter app by executing the following command in the build directory
```
cmake --build . --target pic2dem
```

Warning

MFIX-Exa currently experiences a floating point exception when running with ascent support. Make sure all of the fpe traps in your inputs are off.

Basics

Source code

Before building, first obtain a copy of the source code following the instructions on the MFIX-Exa website.

Modules

OLCF provides three programming environments on Crusher: PrgEnv-amd, PrgEnv-cray, and PrgEnv-gnu. We’ve had the most success with the latter two and only provide build instructions for them below. Note that in the PrgEnv-gnu env you will set different compilers if you are compiling for CPU only (e.g., most of the dependencies) or if you are compiling for the GPU with HIP-support.

module purge
module reset
module load cmake/3.23.2
module load cpe/22.08
module load craype-accel-amd-gfx90a
module load rocm/5.2.0
module load cray-mpich/8.1.21
module load cce/15.0.0  # must reload after rocm

export MPICH_GPU_SUPPORT_ENABLED=1

export CFLAGS="-I${ROCM_PATH}/include"
export CXXFLAGS="-I${ROCM_PATH}/include"
export LDFLAGS="-L${ROCM_PATH}/lib -lamdhip64"

Modules

module purge
module reset
module swap PrgEnv-cray PrgEnv-gnu/8.3.3
module load cmake/3.23.2
module load craype-accel-amd-gfx90a
module load cray-mpich/8.1.21
module load rocm/5.2.0
module load cce/14.0.2
module load boost/1.79.0

CPU compilers

export CC=$(which cc)
export CXX=$(which CC)
export FC=$(which ftn)

GPU compilers

export CC=$(which hipcc)
export CXX=$(which hipcc)

export MPICH_GPU_SUPPORT_ENABLED=1    # GPU-aware MPI

export CFLAGS="-I${MPICH_DIR}/include -I${ROCM_PATH}/include"
export CXXFLAGS="-I$OLCF_BOOST_ROOT/include -I${MPICH_DIR}/include -I${ROCM_PATH}/include"
export LDFLAGS="-L${MPICH_DIR}/lib -L${CRAY_MPICH_ROOTDIR}/gtl/lib -I${ROCM_PATH}/lib -lmpi -lmpi_gtl_hsa"
export HIPFLAGS="--amdgpu-target=gfx90a"

Building MFIX-Exa

The commands below are the superbuild instructions, i.e., AMReX is built as part of the MFIX-Exa build process. To build MFIX-Exa with hypre, csg and/or ascent dependencies, you first need to build and install these libraries and their dependencies. Instructions on building the necessary dependencies are below and should be successfully installed first. These instructions are currently only provided for cmake.

cmake -DMFIX_MPI=yes \
      -DMFIX_OMP=no \
      -DMFIX_GPU_BACKEND=NONE \
      -DAMReX_TINY_PROFILE=no \
      -DMFIX_CSG=no \
      -DMFIX_HYPRE=no \
      -DCMAKE_BUILD_TYPE=Release \
      ../
make -j8

export AMREX_AMD_ARCH=gfx90a

cmake -DMFIX_MPI=yes \
      -DMFIX_OMP=no \
      -DMFIX_CSG=no \
      -DMFIX_HYPRE=no \
      -DMFIX_GPU_BACKEND=HIP \
      -DAMReX_TINY_PROFILE=no \
      -DCMAKE_BUILD_TYPE=Release \
      ../
make -j8

export HYPRE_DIR=$HYPRE_INSTALL_DIR
export HYPRE_ROOT=$HYPRE_DIR
export HYPRE_LIBRARIES=$HYPRE_DIR/lib
export HYPRE_INCLUDE_DIRS=$HYPRE_DIR/include

export ASCENT_DIR=$ASCENT_INSTALL_DIR
export CONDUIT_DIR=$ASCENT_DIR
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:$ASCENT_DIR/lib/cmake/ascent
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:$ASCENT_DIR/lib/cmake/conduit

export CSG_DIR=$CSG_INSTALL_DIR
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:$CSG_DIR

export BOOST_HOME=$MY_INSTALL_DIR/boost            #PrgEnv-cray ONLY
export Boost_INCLUDE_DIR="-I$BOOST_HOME/include"   #PrgEnv-cray ONLY

cmake -DMFIX_MPI=yes \
      -DMFIX_OMP=no \
      -DMFIX_CSG=yes \
      -DMFIX_HYPRE=yes \
      -DAMReX_ASCENT=yes \
      -DAMReX_CONDUIT=yes \
      -DMFIX_GPU_BACKEND=NONE \
      -DAMReX_TINY_PROFILE=no \
      -DCMAKE_BUILD_TYPE=Release \
      ../mfix
make -j8

export HYPRE_DIR=$HYPRE_INSTALL_DIR
export HYPRE_ROOT=$HYPRE_DIR
export HYPRE_LIBRARIES=$HYPRE_DIR/lib
export HYPRE_INCLUDE_DIRS=$HYPRE_DIR/include

export ASCENT_DIR=$ASCENT_INSTALL_DIR
export CONDUIT_DIR=$ASCENT_DIR
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:$ASCENT_DIR/lib/cmake/ascent
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:$ASCENT_DIR/lib/cmake/conduit

export CSG_DIR=$CSG_INSTALL_DIR
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:$CSG_DIR

export BOOST_HOME=$MY_INSTALL_DIR/boost            #PrgEnv-cray ONLY
export Boost_INCLUDE_DIR="-I$BOOST_HOME/include"   #PrgEnv-cray ONLY

export AMREX_AMD_ARCH=gfx90a

cmake -DMFIX_MPI=yes \
      -DMFIX_OMP=no \
      -DMFIX_CSG=yes \
      -DMFIX_HYPRE=yes \
      -DAMReX_ASCENT=yes \
      -DAMReX_CONDUIT=yes \
      -DMFIX_GPU_BACKEND=HIP \
      -DGPUS_PER_NODE=8 \
      -DAMReX_TINY_PROFILE=no \
      -DCMAKE_BUILD_TYPE=Release \
      ../mfix
make -j8

Optional build dependencies

The following dependencies need to be built and installed prior to following any of the full build instructions above.

HYPRE

git clone https://github.com/hypre-space/hypre.git
pushd hypre/src/
git checkout v2.26.0
./configure --prefix=$HYPRE_INSTALL_DIR --with-MPI
make -j8 install
popd

git clone https://github.com/hypre-space/hypre.git
pushd hypre/src/
git checkout v2.26.0
./configure --prefix=$HYPRE_INSTALL_DIR \
            --without-superlu \
            --disable-bigint \
            --without-openmp \
            --enable-shared \
            --with-hip \
            --with-gpu-arch=gfx90a \
            --enable-rocsparse \
            --enable-rocrand \
            --enable-unified-memory \
            --enable-device-memory-pool \
            --with-MPI-lib-dirs="${MPICH_DIR}/lib ${CRAY_MPICH_ROOTDIR}/gtl/lib ${ROCM_PATH}/lib" \
            --with-MPI-libs="mpi mpi_gtl_hsa amdhip64" \
            --with-MPI-include="${MPICH_DIR}/include {ROCM_PATH}/include"
make -j8 install
popd

Catch2

git clone --depth 1 --branch v2.13.7 https://github.com/catchorg/Catch2
pushd Catch2/
cmake -S . -B build -DCMAKE_INSTALL_PREFIX=$CSG_INSTALL_DIR
cd build/
make -j8 install
popd

GMP

wget --no-check-certificate https://ftp.gnu.org/gnu/gmp/gmp-6.2.1.tar.xz
tar -xf gmp-6.2.1.tar.xz
pushd gmp-6.2.1
./configure --prefix=$CSG_INSTALL_DIR
make -j8 install
popd

MPFR

wget https://www.mpfr.org/mpfr-current/mpfr-4.2.0.tar.gz
tar -zxvf mpfr-4.2.0.tar.gz
pushd mpfr-4.2.0/
./configure --prefix=$CSG_INSTALL_DIR \
            --with-gmp-lib=${CSG_INSTALL_DIR}/lib \
            --with-gmp-include=${CSG_INSTALL_DIR}/include
make -j8 install
popd

CGAL

git clone --depth 1 --branch v5.3 https://github.com/CGAL/cgal
pushd cgal/
cmake -S . -B build -DCMAKE_INSTALL_PREFIX=$CSG_INSTALL_DIR
cd build/
make -j8 install
popd

PEGTL
```
git clone --branch 3.2.2 https://github.com/taocpp/PEGTL
pushd PEGTL/
cmake -S . -B build -DCMAKE_INSTALL_PREFIX=$CSG_INSTALL_DIR
cd build/
make -j8 install
popd
```
Warning

For some reason, something in PEGTL does not set the experimental filesystem when installing in PrgEnv-gnu (but it does in PrgEnv-cray). So you you have to force that manually. One hack to do so is shown below.

Make sure that “experimental” precedes filesystem on or near lines 47 and 51 in CSG_INSTALL_DIR/include/tao/pegtl/internal/filesystem.hpp
```
...
#include <experimental/filesystem>

namespace TAO_PEGTL_NAMESPACE::internal
{
namespace filesystem = ::std::experimental::filesystem;
...
```

Boost

wget https://boostorg.jfrog.io/artifactory/main/release/1.81.0/source/boost_1_81_0.tar.gz
tar -zxvf boost_1_81_0.tar.gz
pushd boost_1_81_0/
./bootstrap.sh
./b2 install --prefix=$BOOST_INSTALL_DIR
popd

Warning

Only needed for PrgEnv-cray. This appears to build with the native gcc/g++ compilers, not the cray wrapped cc/CC compilers, but “it works.”

Conduit

git clone --recursive https://github.com/LLNL/conduit.git
pushd conduit/
git checkout v0.8.6
mkdir build && cd build
cmake -S ../src -DCMAKE_INSTALL_PREFIX=$ASCENT_INSTALL_DIR \
      -DCMAKE_BUILD_TYPE=Release \
      -DENABLE_OPENMP=OFF \
      -DENABLE_MPI=ON \
      -DENABLE_CUDA=OFF \
      -DENABLE_PYTHON=OFF \
      -DENABLE_GTEST=OFF \
      -DENABLE_TESTS=OFF
make -j8 install
popd

Vtk-m

git clone --branch master https://gitlab.kitware.com/vtk/vtk-m.git
pushd vtk-m/
git checkout v1.9.0
mkdir build && cd build/
cmake -S ../ -DCMAKE_INSTALL_PREFIX=$ASCENT_INSTALL_DIR \
      -DCMAKE_BUILD_TYPE=Release \
      -DVTKm_ENABLE_OPENMP=OFF \
      -DVTKm_ENABLE_MPI=ON \
      -DVTKm_ENABLE_CUDA=OFF \
      -DVTKm_USE_64BIT_IDS=OFF \
      -DVTKm_USE_DOUBLE_PRECISION=ON \
      -DVTKm_USE_DEFAULT_TYPES_FOR_ASCENT=ON \
      -DVTKm_NO_DEPRECATED_VIRTUAL=ON \
      -DVTKm_ENABLE_TESTING=OFF
make -j8 install
popd

Ascent

git clone --recursive https://github.com/Alpine-DAV/ascent.git
pushd ascent
git checkout v0.9.0
mkdir build && cd build/
cmake -S ../src -DCMAKE_INSTALL_PREFIX=$ASCENT_INSTALL_DIR \
      -DCMAKE_BUILD_TYPE=Release \
      -DCONDUIT_DIR=$ASCENT_INSTALL_DIR \
      -DVTKM_DIR=$ASCENT_INSTALL_DIR \
      -DENABLE_VTKH=ON \
      -DENABLE_FORTRAN=OFF \
      -DENABLE_PYTHON=OFF \
      -DENABLE_DOCS=OFF \
      -DBUILD_SHARED_LIBS=ON \
      -DENABLE_GTEST=OFF \
      -DENABLE_TESTS=OFF
make -j8 install
popd

Running Jobs

Common Slurm commands:

sinfo see available/allocated resources
sbatch runit_cpu.sh submit a cpu job to the queue
squeue -u USER check job status of user USER
squeue -p PARTITION check job status of partition PARTITION
scancel JOBID kill a job with id JOBID
salloc -N 1 -A CFD122_crusher -J build -t 01:00:00 grab an interactive node for an hour

Example run scripts:

#!/bin/bash -l

#SBATCH -A CFD122_crusher
#SBATCH -J mfix-timing
#SBATCH -o job_%x-%j.out
#SBATCH -e job_%x-%j.err
#SBATCH --threads-per-core=1
#SBATCH --exclude=crusher[026,027,028,081,126,114,115]
#SBATCH -t 00:05:00
#SBATCH -N 2

nodes=2
nrs=12
omp=1

module purge
module reset
module load cpe/22.08
module load craype-accel-amd-gfx90a
module load rocm/5.2.0
module load cray-mpich/8.1.21
module load cce/15.0.0  # must be loaded after rocm

export OMP_NUM_THREADS=$omp
export MPICH_GPU_SUPPORT_ENABLED=1    # remove for cpu only
#export FI_MR_CACHE_MAX_COUNT=0       # libfabric disable caching
export FI_MR_CACHE_MONITOR=memhooks   # alt cache monitor
export FI_CXI_RX_MATCH_MODE=software
export FI_CXI_REQ_BUF_SIZE=12582912
export FI_CXI_REQ_BUF_MIN_POSTED=6
export FI_CXI_DEFAULT_CQ_SIZE=131072

srun -N $nodes -n $nrs -c1 --ntasks-per-gpu=1 --gpu-bind=closest ./mfix inputs > screen.txt

#cpu: srun -N $nodes -n $nrs -c1 ./mfix inputs > screen.txt

#!/bin/bash -l

#SBATCH -A CFD122_crusher
#SBATCH -J mfix-timing
#SBATCH -o job_%x-%j.out
#SBATCH -e job_%x-%j.err
#SBATCH --threads-per-core=1
#SBATCH --exclude=crusher[026,027,028,081,126,114,115]
#SBATCH -t 00:05:00
#SBATCH -N 2

nodes=2
nrs=12
omp=1

module purge
module reset
module load cpe/22.08
module load craype-accel-amd-gfx90a
module load rocm/5.2.0
module load cray-mpich/8.1.21
module load cce/15.0.0  # must be loaded after rocm

export OMP_NUM_THREADS=$omp
export MPICH_GPU_SUPPORT_ENABLED=1    # remove for cpu only
#export FI_MR_CACHE_MAX_COUNT=0       # libfabric disable caching
export FI_MR_CACHE_MONITOR=memhooks   # alt cache monitor
export FI_CXI_RX_MATCH_MODE=software
export FI_CXI_REQ_BUF_SIZE=12582912
export FI_CXI_REQ_BUF_MIN_POSTED=6
export FI_CXI_DEFAULT_CQ_SIZE=131072

srun -N $nodes -n $nrs -c1 --ntasks-per-gpu=1 --gpu-bind=closest ./mfix inputs > screen.txt

#cpu: srun -N $nodes -n $nrs -c1 ./mfix inputs > screen.txt