Yesterday, Today and The Future – Disruptive Technologies for Supercomputing Applications

Author: Nick Ihli, Director of Field Services, Adaptive Computing

Last week I attended IDC’s HPC User Forum held in Broomfield, Colorado. There were many interesting topics including an informative panel on President Obama’s National Strategic Computing Initiative to develop exascale computing. I also had the privilege of being on a panel for Disruptive Technologies. Multiple companies presented some amazing new technologies that really will help take HPC to the next level. Paul Muzio from CUNY opened the discussion with talking about disruptive technologies throughout time and he showed a picture of a wheel and then a sports car. This really got me thinking about how disruptive technologies open up new capabilities but also new problems that need solving.

Look at what happened with the invention of the car. Previously, everyone rode around on horse or horse drawn carriages. There was one main problem with this method, the waste. A typical horse would produce 15-35 pounds of manure per day. This brought flies and disease. It was even predicted that by 1945 “London would be buried by nine feet of manure.” Sounds like great living conditions!

Well, the disruptive technology of the automobile solved this problem, but brought along with it new problems. How would we handle having multiple vehicles driving around the city at fast speeds without total chaos? A traffic system and new laws were created. What about the mud, rough dirt and rocky roads causing havoc on wheels and tires? Roads were paved and a greater infrastructure was built across the United States enabling a better way to ship commerce. A new kind of waste was now an issue, this time being released into the atmosphere. How might that be solved? Many agree electric vehicles may solve the smog issues, but then we have a greater dependency on electricity, and with that a need to produce more electricity.

Cars are also getting smarter, solving another problem the automobile created, high-speed wrecks. According to CNN Money, there were more than 18,600 motor-vehicle deaths from January through June of 2015. How do we solve the problem of the fallible human? Why not try an “infallible” computer. As cars begin to talk with each other, constantly monitoring the road, conditions, pedestrians, and other cars within a certain radius, we will sit back and relax and know we’ll get to our destination on time and safe. No more accidents, no more stress, no more traffic congestion….right?

That might be the case, but once again new problems will need solving, even some we haven’t even thought of. Who’s at fault if one of the cars “loses connection” or “a little bug” brings the system down causing a wreck. The car manufacturer? The software company? What about security? These cars would need to be highly secure, so any nefarious hacking is thwarted. Maybe these cars will have mini-supercomputers, running live simulations based on millions of constantly changing data points about the environment to be as safe as possible.

Let’s now return to the exascale initiative. Not only will exascale systems be a disruptive technology, but it will also enable organizations to create their own disruptive technology through new discoveries on these massive supercomputers. The development of the self-driving car will definitely have exascale systems to thank. But as we make our way to “Exascale”, just like other disruptive technologies, new problems will need to be solved. For instance, how to handle the energy efficiency and cooling of these systems that would need the power of small a city? Interconnects, memory, storage, parallel algorithms all will need massive improvements. From a scheduling perspective, schedulers like Moab will need tighter integration with power, cooling, interconnects, memory, storage, etc. There’s many incremental steps left to take, but the journey is exciting.

With the raw power of the Cray-1 in the palm of our hands with today’s smartphones, what will supercomputing systems and resulting technology look like in another 40 years? Where in the stars will it take us? What diseases will be eradicated? When will I get my hoverboard?

I can’t wait to find out. See you in the future!