ParRes Kernels Experiment (MPI)

This experiment runs a set of the ParRes Kernels as a microbenchmark for Granny's MPI implementation.

Start AKS cluster

In the experiment-base terminal, run:

(faasm-exp-base) inv cluster.provision --vm Standard_D8_v5 --nodes 3 cluster.credentials

Deploy the cluster:

(faasm-exp-faabric) faasmctl deploy.k8s --workers=2

Upload the WASM file:

(faasm-exp-faabric) inv kernels-mpi.wasm.upload

and run the experiment with:

(faasm-exp-faabric) inv kernels-mpi.run.wasm

finally, delete the Granny cluster:

faasmctl delete

Deploy the OpenMPI cluster:

inv kernels-mpi.native.deploy

inv kernels-mpi.run.native

finally, delete the OpenMPI cluster

inv kernels-mpi.native.delete

To plot the results, just run:

inv kernels-mpi.plot

the plot will be available in ./plots/kernels-mpi/mpi_kernels_slowdown.pdf, we also include it below:

Finally, delete the AKS cluster:

(faasm-exp-base) inv cluster.delete