Specfem 3D globe -- Bench readme

Note that this guide is still a work in progress and is currently being reviewed.

Get the source

Clone the repository in a location of your choice, let's say $HOME.

cd $HOME
git clone https://github.com/geodynamics/specfem3d_globe.git

Then use a fixed and stable version of specfem3D_globe (the one of October 31, 2017 for example, see https://github.com/geodynamics/specfem3d_globe/commits/master)

cd $HOME/specfem3d_globe
git checkout b1d6ba966496f269611eff8c2cf1f22bcdac2bd9

If this is not done, clone the ueabs repository.

cd $HOME
git clone https://repository.prace-ri.eu/git/UEABS/ueabs.git

In the specfem3D folder of this repo, you will find test cases in the test_cases folder, you will also find environment and submission scripts templates for several machines

Load the environment

You will need a fortran and a C compiler and a MPI library. The following variables are relevent to compile the code:

• LANG=C
• FC
• MPIFC
• CC
• MPICC

Compiling with CUDA to run on GPUs, you will also need to load the cuda environment and set the two following variables

• CUDA_LIB
• CUDA_INC

An exemple (compiling for GPUs) on the ouessant cluster at IDRIS - France:

LANG=C

module purge
module load pgi cuda ompi

export FC=`which pgfortran`
export MPIFC=`which mpif90`
export CC=`which pgcc`
export MPICC=`which mpicc`
export CUDA_LIB="$CUDAROOT/lib64"
export CUDA_INC="$CUDAROOT/include"

Once again, you will find in the specfem3D folder of this repo a folder named env, with files named env_x which gives examples of the environment used on several supercomputers during the last benchmark campaign

Compile specfem

As arrays are staticaly declared, you will need to compile specfem once for each test case with the right Par_file On some environement, depending on MPI configuration you will need to replace use mpi statement with include mpif.h, use the script and prodedure commented below.

First you will have to configure.

On GPU platform you will have to add the following arguments to the configure:--build=ppc64 --with-cuda=cuda5.

cp -r $HOME/specfem3d_globe specfem_compil_${test_case_id}
cp $HOME/bench_spec/test_case_${test_case_id}/DATA/Par_file specfem_compil_${test_case_id}/DATA/

cd specfem_compil_${test_case_id}

### replace `use mpi` if needed ###
# cd utils
# perl replace_use_mpi_with_include_mpif_dot_h.pl
# cd ..
####################################

./configure --prefix=$PWD

On Xeon Phi, since support is recent you should replace the following variables values in the generated Makefile:

FCFLAGS = -g -O3 -qopenmp -xMIC-AVX512 -DUSE_FP32 -DOPT_STREAMS -align array64byte  -fp-model fast=2 -traceback -mcmodel=large
FCFLAGS_f90 = -mod ./obj -I./obj -I.  -I. -I${SETUP} -xMIC-AVX512
CPPFLAGS = -I${SETUP}  -DFORCE_VECTORIZATION  -xMIC-AVX512

Note: Be careful, in most machines login node does not have the same instruction set so, in order to compile with the right instruction set, you'll have to compile on a compute node (salloc + ssh)

Finally compile with make:

make clean
make all

-> You will find in the specfem folder of ueabs repository the file "compile.sh" which is an compilation script template for several machines (different architectures : KNL, SKL, Haswell and GPU)

Launch specfem

You can use or be inspired by the submission script template in the job_script folder using the appropriate job submission command :

• qsub for pbs job,
• sbatch for slurm job,
• ccc_msub for irene job (wrapper),
• llsubmit for LoadLeveler job.

Gather results

The relevant metric for this benchmark is time for the solver. Using slurm, it is easy to gather as each mpirun or srun is interpreted as a step wich is already timed. So the command line sacct -j <job_id> allows you to catch the metric.

Or you can find more precise timing info at the end of this output file : specfem3d_globe/OUTPUT_FILES/output_solver.txt