The Intel® MPI Library is one of the most widely used MPI implementations for distributed computing, with continual optimizations that help HPC developers and data scientists deliver performance and flexibility.
In this session, Technical Consulting Engineer James Tullos gives a tour of the tool’s latest incarnation, including how it:
- Increases performance on OpenFabrics Interfaces (OFI), a framework specifically designed to meet the performance and scalability requirements of HPC applications running in a tightly coupled network environment.
- Provides tight integration for users of Singularity* containers which are used to package entire scientific workflows, software and libraries, and data.
- Optimizes performance on AWS* Elastic Fabric Adapter, a network interface that enables you to run HPC apps at scale, including computational fluid dynamics, weather modeling, and reservoir simulation.
Get the software
- Intel® MPI Library—This library is one of five free Intel® Performance Libraries. Download one or all of them today, free.