High Performance Computing (HPC): Hard to Define…

Our customers continually demand greater computational power, because they want to reduce the execution time of their applications and overcome the limitations of desktop computing architectures. But what are we able to offer them? You may have heard of supercomputing and monster machines, that work on some of mankind’s biggest problems in science and engineering. But these supercomputers are only good for specialized problems and need a lot of budget and expertise. What our customers need are robust, cost-effective alternatives – without compromising efficiency – that stand the test of time. The keyword is: HPC. An HPC machine is more complex than a simple desktop computer but the basics aren’t that much more difficult to understand.


What is high performance computing (HPC) anyway?

Nowadays there is still no kind of universal definition of “HPC”. And for me, it turns out that defining HPC is a bit like defining the word “Smartphone” — although you probably know what a smartphone is, it’s still kind of hard to write a concise definition that includes every important aspect. Even Wikipedia is no help, because the term “High-performance computing” redirects to the “Supercomputer” definition and for narrower definitions of HPC the site points you to the definitions of “high-throughput computing” and “many-task computing”. As you can see, this is all a bit confusing. So let’s try to avoid grey theory and take a more realistic approach.


Let’s put an end to the theory and let HPC come to life…

According to CIO Review, Kontron’s dedication to R&D and technology innovation is helping to revamp the notion of HPC through its High Performance Embedded Computing (HPEC) solutions.Kontron has also been selected as one of the top 20 most promising high performance computing (HPC) technology providers – but that’s just an aside . Actually we are providing application-ready solutions that help our customers better analyze the growing volume of data from the IoT.  Kontron’s StarVX – for example – is a strong HPEC that is designed to help developers dramatically streamline the process from design to field deployment of next-generation HPEC radar, sonar and imagine processing.  This powerful platform is able to host multiple VPX (ANSI Standard) cards which enable customers to quickly build custom solutions. Coupled with our VPX processing boards, StarVX can be utilized to perform real time HPEC processing at rates demanded by the explosion of sensor data.


ra2 studio/@fotolia.com


One real-life example for HPC is farm vehicles, which are now being equipped with more sensor and automatic controls. This data can be coupled with satellite images of a farm (such as Google Maps) resulting in a HPEC big data application that helps farmers quickly visualize where to fertilize, seed and harvest their farm in the best way. By leveraging our platforms, correlations analysis of these images in real time is possible as the vehicles are traversing the farms – and you need no complicate definition to know, that this is the power of HPC at the edge.

So what about you? Do you have your own definition of HPC?

Thank you!

Your comment was submitted.

An error occured on subscribing!:

{{comment.date.format('MMMM DD, YYYY')}}


There are no comments yet.

Stay connected