Autonomous Driving, 5g, Precision Medicine And More: The Impact of High-Performance Computing and its Future in Cloud

A public cloud platform is, by definition, the only solution that enables companies to take global advantage of elasticity over massive specialized compute scale. An organization can leverage both on-premises infrastructure environment while bursting to a cloud environment to better accommodate the variable demand for compute, limit wait times for engineers, upgrade performance throughout the year without any lock-in as new Cloud infrastructure becomes available, map directly their usage against their current place in the life cycle while having the ability to share the data globally with peers.

We see this clearly emerging with Microsoft’s financial services customers, like UBS who need to use large amounts of compute during month-end and year-end cycles to comply to local regulation while ensuring they don’t have unused infrastructure investments during non-peak period.

Paramount for HPC is performance and access to a productive infrastructure and Microsoft has powerful, scalable, and cost-performant HPC instances. Indeed, Exabyte.io’s independent research published in February 2017 found “Microsoft Azure to deliver the best results, and demonstrated that the performance per single computing core on public cloud to be comparable to modern traditional supercomputing systems.” Exabyte suggests that “the concept of high-performance computing in the cloud is ready for widespread adoption.”

Microsoft is a leader in the HPC cloud space by consistently making significant investments in the HPC area. Microsoft has addressed customer concerns and ensured companies get the best value for performance, scalability, and cost-efficiency by designing compute infrastructure with the applications in mind.

For example, large scale parallel simulations like oil and gas seismic processing, weather simulations, automotive structural analysis or material development molecular dynamics are now possible through the use of the new 100Gb Infiniband-backed performance-tuned CPU. The recently launched Azure HC – “C stands for “Compute” - virtual machine series enables structural mechanics such as crash test simulation, airplane engine design or Azure HB – B stands for “memory bandwidth” - series designed for fluid dynamics workloads which can help scale distributed simulations (“message passing interface” simulations) up to 18,000 cores, linearly.

Notably, Microsoft also delivers fast orchestration and reduced boot times for customers through technology it acquired with CycleCloud, an orchestrator that can enable thousands of cores to be spun up in the span of a few minutes, and Avere, a hot caching mechanism which can reduce boot times, latency to data source and read throughput.

Furthermore, Microsoft has ensured we can provide custom solutions for our customer’s hardest to solve problems through partnerships like Cray Supercomputers and ensured support in this new Cloud world with technology partnership like Ansys, Altair, Schlumberger, Halliburton, Siemens Star CCM+, Dassault Exa-Powerflow, LS-Dyna, PBS Pro, etc.

In today’s fast-moving business environment, companies need to respond faster to internal and customer needs to be competitive. Having the most performant and yet flexible HPC infrastructure enables companies to be more productive and get the best value for performance, scalability, and cost-efficiency. We believe Microsoft has made the right technology investments and partnerships to exceed our customer’s expectations in this space.