To run OPM Flow software on Azure, you just need the right HPC hardware and a shared storage location for the input and output files as our new example workload for https://docs.microsoft.com/azure/architecture/example-scenario/infrastructure/reservoir-simulation shows.
These compute-hungry models are used to predict complex flows of fluids such as oil, water, and gas beneath the earth’s surface. OPM Flow software comes from the popular open-source Open Porous Media Initiative.
This example deploys a small HPC cluster with a Linux head node running OPM Flow, along with the PBS Pro 19.1 job-scheduling software. The head node mounts a 4-terabyte (TB) NFS space as the HPC file share, but depending on your model and IO requirements, you can use other storage options.
Azure HB-series high-performance computing (HPC) virtual machines provide the horsepower, although you can also use general-purpose virtual machines for the head node.
For details, see the https://docs.microsoft.com/azure/architecture/example-scenario/infrastructure/reservoir-simulation example workload in the Azure Architecture Center. This example also includes https://github.com/Azure/azurehpc/tree/master/tutorials/oil_and_gas_opm on GitHub.
Related resources
Reservoir Simulation on Azure: HPC for Oil & Gas blog
HPC: Oil and gas in Azure blog
https://docs.microsoft.com/azure/architecture/example-scenario/infrastructure/hpc-cfd
