Supercomputing: Coming to a data center—or even a desktop—near you
The term supercomputing, also known as high-performance computing, might conjure up the image of a bunch of energy-guzzling, LED-flashing black cabinets with a footprint the size of two tennis courts. And with names like Aurora, Deep Blue, Pangea, and Frontera, supercomputers seem mysterious to most, performing scientific calculations beyond comprehension, for only specific large-scale projects.
But all that is changing as supercomputing enters an era in which upcoming exascale machines combined with AI promise to not only help solve some of the world's biggest problems, such as climate change and drug discovery, but to drive everyday business applications like automotive design and financial modeling.
Greatly accelerating the speed at which calculations can be performed, this new generation of supercomputing is becoming increasingly attractive—and accessible—to a growing number of industries looking to capitalize on their expanding volumes of data, says Bill Mannel, vice president and general manager of HPC at Hewlett Packard Enterprise. With more affordable options, including as a service, more and more businesses are able to tap the power of supercomputers "to make better decisions, build better products, and bring products to market faster," he says.
In this episode of Technology Untangled, Mannel joins host Michael Bird and guests Jacob Balma, HPC and AI engineering research scientist at HPE, and Andrew Emerson, a researcher at the Cineca supercomputing center in Italy, for a look at what's coming up in supercomputing—from crash testing automobiles to simulating the human brain. They explain how the technology works, what exascale and artificial intelligence will bring to the table, and current and future applications, including real-world use cases that will benefit businesses across industries.
So, why do we need a computer that does a billion, billion floating-point calculations per second? "Well, because problems that we want to solve keep getting bigger and bigger," says Mannel, who points to applications like combustible engine modeling to increase fuel efficiency—something scientists are not able to do today.
Please listen to: How HPE Spaceborne Computer-2 will help earthlings
In the financial world, where speed is a critical element in buying and selling stocks and other activities, big banks and securities firms are investing "heavily" in supercomputing technology, Mannel says.
And of course, speed is also a key factor in drug research, as highlighted in the race to develop a vaccination for the coronavirus, notes Emerson, who worked on that as part of Cineca's team.
How will AI play into all this?
Supercomputing boosted by machine learning is a game changer, says Balma, particularly when it comes to drug discovery and personalized healthcare. He explains his work on the PharML project, which trained a neural-net machine learning model to predict which drugs will bind to a protein and which drugs will not.
Bringing exascale and AI together will lead to other advances as well, like the ability to simulate portions of the human brain and "solve problems like cancer and all these [wild] permutations on coronavirus," he says.
Supercomputing in the real world
Look for supercomputers to become viable tools across a range of industries, from agriculture to manufacturing, Mannel says. With advances driving performance up and costs down, "a lot of organizations can actually afford a supercomputer now, or they can at least use one as a service," he says.
Listen to other episodes:
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.