602f257a4c | ||
---|---|---|
.. | ||
install_fluent.sh | ||
readme.md | ||
run_fluent_hpcx.sh |
readme.md
Install and run Fluent Benchmarks
Prerequisites
Cluster is built with the desired configuration for networking, storage, compute etc. You can see the tutorial or examples folder in this repo for how to set this up.
Dependencies for binary version:
- None
Installation
NOTE: Update the path to the fluent installer tar file in $azhpc_dir/apps/fluent/install_fluent.sh
First copy the apps directory to the cluster. The azhpc-scp
can be used to do this:
azhpc-scp -r $azhpc_dir/apps hpcuser@headnode:.
Alternatively you can checkout the azurehpc repository but you will need to update the paths according to where you put it.
azhpc-run -u hpcuser apps/fluent/install_fluent.sh
Note: This will install into
/apps
.
Connect to the headnode
azhpc-connect -u hpcuser headnode
Setup the benchmark files
Copy the benchmark tarball to /apps/ansys_inc/v193/fluent then run
tar xvf <tarfilename>.tar
This will place the required files where fluent will find them
Running
NOTE: In the run script (run_fluent_hpcx.sh) you will need to update the license server. Currently it is set to localhost which would require a tunnel to be created (currently the ssh tunnel command commented out in the script).
Now, you can run as follows:
for ppn in 60 45 30; do
for nodes in 2 4 8 16 32 64 128; do
name=racecar_hpcx_${nodes}x${ppn}
mkdir $name
cd $name
cp ../run_fluent_hpcx.sh .
qsub -l select=${nodes}:ncpus=${ppn}:mpiprocs=${ppn},place=scatter:excl -N $name ./run_fluent_hpcx.sh
cd -
done
done