That Was Fast: Summit Already Speeding Research into Addiction, Superconductors

by Paresh Kharya

Just weeks after its debut, Summit, the world’s fastest supercomputer, is already blasting through scientific applications crucial to breakthroughs in everything from superconductors to understanding addiction.

Summit, based at the Oak Ridge National Laboratory, in Tennessee, already runs CoMet — which helps identify genetic patterns linked to diseases — 150x faster than its predecessor, Titan. It’s running another application, QMCPACK — which handles quantum Monte Carlo simulations for discovering new materials such as next-generation superconductors — 50x faster than Titan.

The ability to quickly accelerate widely-used scientific applications such as these comes thanks to our more than a decade of investment across what technologists call “the stack.” That is, everything from architecture improvements in our GPU parallel processors to system design, software, algorithms, and optimized applications. While innovating across the entire stack hard, it’s also essential, because, with the end of Moore’s law, there are no automatic performance gains.

Summit, powered by 27,648 NVIDIA GPUs, is the latest GPU-powered supercomputer built to accelerate scientific discovery of all kinds. Built for the U.S. Department of Energy, Summit is the world’s first supercomputer to achieve over a 100 petaflops, accelerating the work of the world’s best scientists in high-energy physics, materials discovery, healthcare and more.

But Summit delivers more than just speed. Instead of one GPU per node with Titan, Summit has six Tensor Core GPUs per node. That gives Summit the flexibility to do traditional simulations along with the GPU-driven deep learning techniques that have upended the computing world since Titan was completed six years ago.

How Volta Stacks the Deck

With Volta, we reinvented the GPU. Its revolutionary Tensor Core architecture enables multi-precision computing. So it can crank through deep learning at 125 teraflops at FP16 precision. Or when greater range or precision are needed, such as for scientific simulations, it can compute at FP64 and FP32.

This fusing of highly efficient processing required for deep learning and highly precise calculations required for scientific simulations enables Volta to be a computational powerhouse for both AI and HPC.

Our-full stack optimization approach means researchers can put the raw speed of systems based on Volta, such as Summit, to work faster. That, in turn, accelerates solutions to our most important challenges. Researchers with early access to Summit have already demonstrated the benefits of this approach in genomic and materials science research that promise real-world benefits.

Solving Genetic Mysteries

Genes hold blueprints for diseases and conditions we’re predisposed to. Diseases like Alzheimer’s or conditions like chronic pain can lead to opioid prescriptions, and 10 percent of those prescribed opioids will become addicted.

The human genome is composed of 3 billion nucleotides and massive computing is required to understand combinations of genes and leading to chronic pain or opioid addiction — possibilities larger than the number of atoms in universe.

Powered by six V100 Tensor Core GPUs, each node of Summit is providing 150x higher performance than Titan and is enabling discovery of previously impossible genetic patterns. This will help pave the way for medical breakthroughs in a range of conditions such as heart diseases, cancer, Alzheimer’s and opioid addiction.

Unlocking Power of Superconductivity

Superconductors are among the most exciting materials yet discovered. They can transmit electricity without any loss and can be used to store energy indefinitely. They can be used to develop powerful scientific magnets for MRI equipment, for levitating trains, and magnetic fusion devices among other uses.

A key challenge: superconductors operate only at extremely low temperatures (-243° C) that require expensive setup such as the use of liquid helium.

High-temperature superconductors (HTS) can be operated at -70° C, but are brittle, hard to manufacture. Quantum Monte Carlo (QMC) electronic structure calculations can help identify new HTS materials with metal like properties.

QMCPACK is optimized for Summit’s Volta GPUs and runs 50x faster on Summit node than Titan node. This is enabling researchers to greatly increase the complexity of materials that they can simulate and dramatically accelerate ORNL’s research for new, cost-effective superconductors.

AI and Simulation: A Powerful Combination

Summit is a shining example of the big shift in the current breed of supercomputers to machines that are both fast enough to accelerate scientific simulations and smart enough to gather insights from massive volumes of data.

It’s a powerful combination that promises to help this generation of scientists accomplish wonders. With Summit and systems like it, we’re doing everything we can to ensure they do just that.