Crusher (OLCF)

The Crusher cluster is located at OLCF.

On Crusher, each compute node provides four AMD MI250X GPUs, each with two Graphics Compute Dies (GCDs) for a total of 8 GCDs per node. You can think of the 8 GCDs as 8 separate GPUs, each having 64 GB of high-bandwidth memory (HBM2E).

Introduction

If you are new to this system, please see the following resources:

Crusher user guide
Batch system: Slurm
Production directories:
- $HOME: per-user directory, use only for inputs, source and scripts; backed up; mounted as read-only on compute nodes, that means you cannot run in it (50 GB quota)
- $PROJWORK/$proj/: shared with all members of a project, purged every 90 days, Lustre (recommended)
- $MEMBERWORK/$proj/: single user, purged every 90 days, Lustre (usually smaller quota, 50TB default quota)
- $WORLDWORK/$proj/: shared with all users, purged every 90 days, Lustre (50TB default quota)

Note: the Orion Lustre filesystem on Frontier and Crusher, and the older Alpine GPFS filesystem on Summit are not mounted on each others machines. Use Globus to transfer data between them if needed.

Preparation

Use the following commands to download the WarpX source code:

git clone https://github.com/BLAST-WarpX/warpx.git $HOME/src/warpx

We use system software modules, add environment hints and further dependencies via the file $HOME/crusher_warpx.profile. Create it now:

cp $HOME/src/warpx/Tools/machines/crusher-olcf/crusher_warpx.profile.example $HOME/crusher_warpx.profile

Edit the 2nd line of this script, which sets the export proj="" variable. For example, if you are member of the project aph114, then run vi $HOME/crusher_warpx.profile. Enter the edit mode by typing i and edit line 2 to read:

export proj="aph114"

Exit the vi editor with Esc and then type :wq (write & quit).

Important

Now, and as the first step on future logins to Crusher, activate these environment settings:

source $HOME/crusher_warpx.profile

Finally, since Crusher does not yet provide software modules for some of our dependencies, install them once:

bash $HOME/src/warpx/Tools/machines/crusher-olcf/install_dependencies.sh
source $HOME/sw/crusher/gpu/venvs/warpx-crusher/bin/activate

Script Details

#!/bin/bash
#
# Copyright 2023 The WarpX Community
#
# This file is part of WarpX.
#
# Author: Axel Huebl
# License: BSD-3-Clause-LBNL

# Exit on first error encountered #############################################
#
set -eu -o pipefail


# Check: ######################################################################
#
#   Was crusher_gpu_warpx.profile sourced and configured correctly?
if [ -z ${proj-} ]; then echo "WARNING: The 'proj' variable is not yet set in your crusher_warpx.profile file! Please edit its line 2 to continue!"; exit 1; fi


# Check $proj variable is correct and has a corresponding CFS directory #######
#
if [ ! -d "${PROJWORK}/${proj}/" ]
then
    echo "WARNING: The directory $PROJWORK/$proj/ does not exist!"
    echo "Is the \$proj environment variable of value \"$proj\" correctly set? "
    echo "Please edit line 2 of your crusher_warpx.profile file to continue!"
    exit
fi


# Remove old dependencies #####################################################
#
SW_DIR="${HOME}/sw/crusher/gpu"
rm -rf ${SW_DIR}
mkdir -p ${SW_DIR}

# remove common user mistakes in python, located in .local instead of a venv
python3 -m pip uninstall -qq -y pywarpx
python3 -m pip uninstall -qq -y warpx
python3 -m pip uninstall -qqq -y mpi4py 2>/dev/null || true


# General extra dependencies ##################################################
#

# BLAS++ (for PSATD+RZ)
if [ -d $HOME/src/blaspp ]
then
  cd $HOME/src/blaspp
  git fetch
  git checkout v2024.05.31
  cd -
else
  git clone -b v2024.05.31 https://github.com/icl-utk-edu/blaspp.git $HOME/src/blaspp
fi
rm -rf $HOME/src/blaspp-crusher-gpu-build
CXX=$(which CC) cmake -S $HOME/src/blaspp -B $HOME/src/blaspp-crusher-gpu-build -Duse_openmp=OFF -Dgpu_backend=hip -DCMAKE_CXX_STANDARD=17 -DCMAKE_INSTALL_PREFIX=${SW_DIR}/blaspp-2024.05.31
cmake --build $HOME/src/blaspp-crusher-gpu-build --target install --parallel 16
rm -rf $HOME/src/blaspp-crusher-gpu-build

# LAPACK++ (for PSATD+RZ)
if [ -d $HOME/src/lapackpp ]
then
  cd $HOME/src/lapackpp
  git fetch
  git checkout v2024.05.31
  cd -
else
  git clone -b v2024.05.31 https://github.com/icl-utk-edu/lapackpp.git $HOME/src/lapackpp
fi
rm -rf $HOME/src/lapackpp-crusher-gpu-build
CXX=$(which CC) CXXFLAGS="-DLAPACK_FORTRAN_ADD_" cmake -S $HOME/src/lapackpp -B $HOME/src/lapackpp-crusher-gpu-build -DCMAKE_CXX_STANDARD=17 -Dbuild_tests=OFF -DCMAKE_INSTALL_RPATH_USE_LINK_PATH=ON -DCMAKE_INSTALL_PREFIX=${SW_DIR}/lapackpp-2024.05.31
cmake --build $HOME/src/lapackpp-crusher-gpu-build --target install --parallel 16
rm -rf $HOME/src/lapackpp-crusher-gpu-build


# Python ######################################################################
#
python3 -m pip install --upgrade pip
python3 -m pip install --upgrade virtualenv
python3 -m pip cache purge
rm -rf ${SW_DIR}/venvs/warpx-crusher
python3 -m venv ${SW_DIR}/venvs/warpx-crusher
source ${SW_DIR}/venvs/warpx-crusher/bin/activate
python3 -m pip install --upgrade pip
python3 -m pip install --upgrade build
python3 -m pip install --upgrade packaging
python3 -m pip install --upgrade wheel
python3 -m pip install --upgrade setuptools[core]
python3 -m pip install --upgrade cython
python3 -m pip install --upgrade numpy
python3 -m pip install --upgrade pandas
python3 -m pip install --upgrade scipy
MPICC="cc -shared" python3 -m pip install --upgrade mpi4py --no-cache-dir --no-build-isolation --no-binary mpi4py
python3 -m pip install --upgrade openpmd-api
python3 -m pip install --upgrade matplotlib
python3 -m pip install --upgrade yt
# install or update WarpX dependencies such as picmistandard
python3 -m pip install --upgrade -r $HOME/src/warpx/requirements.txt
# optional: for libEnsemble
python3 -m pip install -r $HOME/src/warpx/Tools/LibEnsemble/requirements.txt
# optional: for optimas (based on libEnsemble & ax->botorch->gpytorch->pytorch)
#python3 -m pip install --upgrade torch --index-url https://download.pytorch.org/whl/rocm5.4.2
#python3 -m pip install -r $HOME/src/warpx/Tools/optimas/requirements.txt

Compilation

Use the following cmake commands to compile the application executable:

cd $HOME/src/warpx
rm -rf build_crusher

cmake -S . -B build_crusher -DWarpX_COMPUTE=HIP -DWarpX_FFT=ON -DWarpX_QED_TABLE_GEN=ON -DWarpX_DIMS="1;2;RZ;3"
cmake --build build_crusher -j 16

The WarpX application executables are now in $HOME/src/warpx/build_crusher/bin/. Additionally, the following commands will install WarpX as a Python module:

rm -rf build_crusher_py

cmake -S . -B build_crusher_py -DWarpX_COMPUTE=HIP -DWarpX_FFT=ON -DWarpX_QED_TABLE_GEN=ON -DWarpX_APP=OFF -DWarpX_PYTHON=ON -DWarpX_DIMS="1;2;RZ;3"
cmake --build build_crusher_py -j 16 --target pip_install

Now, you can submit Crusher compute jobs for WarpX Python (PICMI) scripts (example scripts). Or, you can use the WarpX executables to submit Crusher jobs (example inputs). For executables, you can reference their location in your job script or copy them to a location in $PROJWORK/$proj/.

Update WarpX & Dependencies

If you already installed WarpX in the past and want to update it, start by getting the latest source code:

cd $HOME/src/warpx

# read the output of this command - does it look ok?
git status

# get the latest WarpX source code
git fetch
git pull

# read the output of these commands - do they look ok?
git status
git log     # press q to exit

And, if needed,

update the crusher_warpx.profile file,
log out and into the system, activate the now updated environment profile as usual,
execute the dependency install scripts.

As a last step, clean the build directory rm -rf $HOME/src/warpx/build_crusher and rebuild WarpX.

Running

MI250X GPUs (2x64 GB)

After requesting an interactive node with the getNode alias above, run a simulation like this, here using 8 MPI ranks and a single node:

runNode ./warpx inputs

Or in non-interactive runs:

Listing 2 You can copy this file from $HOME/src/warpx/Tools/machines/crusher-olcf/submit.sh.

#!/usr/bin/env bash

#SBATCH -A <project id>
#    note: WarpX ECP members use aph114
#SBATCH -J warpx
#SBATCH -o %x-%j.out
#SBATCH -t 00:10:00
#SBATCH -p batch
#SBATCH --ntasks-per-node=8
# Since 2022-12-29 Crusher is using a low-noise mode layout,
# making only 7 instead of 8 cores available per process
# https://docs.olcf.ornl.gov/systems/crusher_quick_start_guide.html#id6
#SBATCH --cpus-per-task=7
#SBATCH --gpus-per-task=1
#SBATCH --gpu-bind=closest
#SBATCH -N 1

# From the documentation:
# Each Crusher compute node consists of [1x] 64-core AMD EPYC 7A53
# "Optimized 3rd Gen EPYC" CPU (with 2 hardware threads per physical core) with
# access to 512 GB of DDR4 memory.
# Each node also contains [4x] AMD MI250X, each with 2 Graphics Compute Dies
# (GCDs) for a total of 8 GCDs per node. The programmer can think of the 8 GCDs
# as 8 separate GPUs, each having 64 GB of high-bandwidth memory (HBM2E).

# note (5-16-22, OLCFHELP-6888)
# this environment setting is currently needed on Crusher to work-around a
# known issue with Libfabric
#export FI_MR_CACHE_MAX_COUNT=0  # libfabric disable caching
# or, less invasive:
export FI_MR_CACHE_MONITOR=memhooks  # alternative cache monitor

# Seen since August 2023 on Frontier, adapting the same for Crusher
# OLCFDEV-1597: OFI Poll Failed UNDELIVERABLE Errors
# https://docs.olcf.ornl.gov/systems/frontier_user_guide.html#olcfdev-1597-ofi-poll-failed-undeliverable-errors
export MPICH_SMP_SINGLE_COPY_MODE=NONE
export FI_CXI_RX_MATCH_MODE=software

# note (9-2-22, OLCFDEV-1079)
# this environment setting is needed to avoid that rocFFT writes a cache in
# the home directory, which does not scale.
export ROCFFT_RTC_CACHE_PATH=/dev/null

export OMP_NUM_THREADS=1
export WARPX_NMPI_PER_NODE=8
export TOTAL_NMPI=$(( ${SLURM_JOB_NUM_NODES} * ${WARPX_NMPI_PER_NODE} ))
srun -N${SLURM_JOB_NUM_NODES} -n${TOTAL_NMPI} --ntasks-per-node=${WARPX_NMPI_PER_NODE} \
    ./warpx.3d inputs > output_${SLURM_JOBID}.txt

Post-Processing

For post-processing, most users use Python via OLCFs’s Jupyter service (Docs).

Please follow the same guidance as for OLCF Summit post-processing.

Known System Issues

Note

Please see the Frontier Known System Issues due to the similarity of the two systems.