High Performance Computing is changing and is no longer limited to running high volumes of batch jobs on an HPC only compute cluster. We are in the era of multi-purpose and embedded High Performance Computing.
Converged, Multi-Purpose
HPC and AI are pieces of a single puzzle.
Most business processes start with seemingly minor questions like will our new wing design fly, will this new heart pump overheat, will our insurance policy expose us to a major loss? The next step is usually the job of HPC. Take this question, apply math to it and generate potential outcomes. The third step is taking the large data set generated by HPC and crunch through it with high performance analysis techniques and with AI.
The hardware running AI and HPC systems are now indistinguishable: fast many core CPU’s, accelerators, low latency network interconnects, high performance storage arrays (https://www.hpe.com/us/en/solutions/hpc-high-performance-computing.html).
We are seeing orchestration techniques converge as well, with Containers.
As AI and HPC technologies get closer together, they will continue to create the business insights companies need to stay ahead.
Embedded
Until recently, the only place to run an HPC workload was an HPC cluster.
Now, HPC is closer to the end users then ever before. Cloud computing brings HPC capabilities to the fingertips of app developers and their users become HPC consumers without even hearing the term.
We’ve seen three great examples with the Living Heart, Living Brain and Building Law projects.
As AI and HPC find their ways to new use-cases and get embedded in other applications, masses will gain access to these technologies. To learn more about how you can benefit from HPC in the cloud, contact us for a free consultation.
