High performance computing (HPC) users often want to run workloads in the cloud for exactly the same reasons as more mainstream cloud users:

• Availability of resources whenever they are needed
• Access to the latest CPUs, GPUs, and other hardware
• Ability to use services like statistics, analytics, and machine learning
• Budget limitations that put owning the equipment out of reach

For the purposes of this eBook, we’re not concerned with the ability to run grand challenges in the cloud. Rather, we’ll focus on the wide variety of “high-performance workloads” that require more compute resources and greater shared I/O bandwidth than the typical enterprise cloud application. Examples include genomics, electronic design automation (EDA), seismic data processing, media and entertainment, high-performance databases, and even large software builds.