Where Curiosity Leads: Fulbright College Researchers Seek New Knowledge with High-Performance Computing

by | Oct 18, 2019 | Faculty, Featured Posts, Field Notes, Guest Writers, Research, Uncategorized

Curiosity may have killed the cat, but it also gave birth to modern humans. From turning sticks and stones into spears and arrows all the way to space exploration, curiosity has been the driving force behind our species’ success.

That desire to know more tomorrow than you did today is also the driving force behind the J. William Fulbright College of Arts and Sciences at the University of Arkansas (U of A), which is classified among the top three percent of universities with the highest level of research activity according to the Carnegie Foundation.

The J. William Fulbright College of Arts and Sciences at the University of Arkansas is among the top three percent of universities with the highest level of research activity.

U.S. News & World Report also ranks the U of A among its top American public research universities, with Fulbright College being a leader on campus receiving 134 research grants totaling more than $19.87 million last year alone.

“Our scientists are making great strides in critical research on nanoforges, early cancer detection and treatment methods, neurological disorder treatment, and in climate change and clean energy – to name but a small sample of their numerous endeavors,” says Jeannine Durdik, associate dean of Fulbright College. “We also partnered with the U of A’s College of Engineering and the Sam M. Walton College of Business to create a new undergraduate data sciences degree launching next fall, which is one of the first of its kind in the nation.”

Durdik says that Fulbright College is home to a wide array of researchers whose fields of study range from the natural to social sciences. While their focuses differ, one factor unites many of the college’s top researchers—reliance on high-performance computing (HPC). 

“It is very rewarding to complement and support the amazing scientists and researchers in the college who continue to advance discovery and insight through HPC,” says Jeff Pummill, co-director of the Arkansas High Performance Computing Center (AHPCC). “We provide the physical resources and expertise to support computational science at Fulbright College and throughout the University of Arkansas.”

To fully understand how Fulbright College embraces the power of HPC, let’s take a look at some of the coolest recent projects coming out of this college.

Feng Wang, chemistry

Drug discovery is a research area where the benefits are easy to understand: new methods mean new drugs, which translates to improved quality of life for the sick and injured.

<strong>Paracetamol molecules</strong>—which form the popular analgesic, Tylenol—are known to crystallize in at least nine different forms. Wang is developing a computational model to accurately predict the stability and interconversions between these forms. Such models facilitate cheaper, safer and more efficient medicines. Courtesy Feng Wang and Ryan Rogers.

Paracetamol molecules—which form the popular analgesic, Tylenol—are known to crystallize in at least nine different forms. Wang is developing a computational model to accurately predict the stability and interconversions between these forms. Such models facilitate cheaper, safer and more efficient medicines. Courtesy Feng Wang and Ryan Rogers.

But many molecules known to have pharmaceutical efficacy aren’t currently being used in drug formulations because of unsuitable characteristics such as size or solubility. That’s why Feng Wang’s work is so exciting.

A professor of physical chemistry who holds the Charles Scharlau Professorship in Chemistry, Wang uses computation to predict the behavior of molecules. By improving the accuracy and efficiency of computer simulations, researchers can discover new structures that will remove those obstacles. Once they hit on a promising crystal structure, it can then be tested in a physical laboratory.

Because Wang’s models require a huge number of integers and large matrices, a supercomputing environment is absolutely necessary. “Just storing those matrices can take a terabyte of local storage,” says Wang. “What we are doing isn’t possible without high-performance computing.” 

Computational chemistry is a relatively new field, and the combination of technology and experimentation is opening up new horizons for researchers. “Things have changed a lot since I was a grad student,” says Wang. “Now we can address much bigger problems.”

Andrew Alverson, biological sciences

We often think that trees and other land plants are the sole producers of breathable air. But Andrew Alverson, associate professor of biological sciences and Twenty-First Century Chair in Bioinformatics, is focused on another, overlooked source of the Earth’s oxygen.

<strong>Scanning electron micrographs</strong> show the ornate silica cell walls of six diverse diatom species. These cells are roughly 20–200 micrometers in size. Courtesy Andrew Alverson and Elizabeth Ruck.

Scanning electron micrographs show the ornate silica cell walls of six diverse diatom species. These cells are roughly 20–200 micrometers in size. Courtesy Andrew Alverson and Elizabeth Ruck.

“Half of the world’s photosynthesis takes place in the oceans,” says Alverson. “Diatoms produce about twenty percent of our oxygen.”

Diatoms are some of the most important lifeforms that you’ve probably never heard of. These single-celled algae call Earth’s waterways home, and they generally don’t get bigger than 0.5 millimeters.

Diatoms have been around for a long time, and more than 100,000 species live in our oceans, lakes and rivers. To better understand their evolutionary history, Alverson is sequencing the genomes of different varieties of diatoms to see what distinguishes them.

“We’re working in the science of the tree of life, or phylogenetics, reconstructing how different species are related to one another,” says Alverson. In the past, scientists used to build phylogenetic trees based on the sequences of just a few genes. “Now we can do that with thousands or even tens of thousands of genes, so at the scale of the whole genome,” says Alverson.

But it’s HPC that makes whole genome sequencing possible. “What we need is lots and lots of memory to assemble the genome,” says Alverson. “We’re using nodes that have 500 – 800 GB of memory. What we do just can’t be done on a desktop computer.”

Salvador Barraza-Lopez, physics

One of the amazing properties of HPC is that it allows us to understand systems that are too small to see. Salvador Barraza-Lopez, associate professor of physics, studies anatomically-thin ferroelectric materials which have a thickness of less than half a nanometer and possess a built-in electric field.

<strong>Snapshot</strong> of a SnSe monolayer at finite temperature (Nano Lett. 16, 1704 (2016). Courtesy Edmund Harriss.

Snapshot of a SnSe monolayer at finite temperature (Nano Lett. 16, 1704 (2016). Courtesy Edmund Harriss.

Such materials could one day serve as memory devices, or could be useful for optical communication in ultrathin, flexible devices. To get a better idea of how these materials behave, Barraza-Lopez relies on HPC computer simulations.

“The research demands simulating these two-dimensional materials with molecular dynamics,” says Barraza-Lopez. “In the simulations, we update interatomic forces constantly using the rules of quantum mechanics: this is a costly and time-consuming endeavor.”

Thanks to the high-performance computing capabilities at the University of Arkansas, Barraza-Lopez’s team was able to publish work on the simulations shortly before the first experimental confirmation became available. 

Says Barraza-Lopez, “I get a great sense of accomplishment from contributing research that, with continued support, may become part of everyday technologies.”

Tulin Kaman, mathematical sciences

Fusion power promises a nearly inexhaustible source of clean, sustainable energy. Problem is, no one has yet figured out how to build a reactor that generates more energy than it takes to run. 

But assistant professor and Lawrence Jesser Toll Jr. Chair of Mathematical Sciences Tulin Kaman believes in the promise of fusion and is working to make it a reality. Her research concentrates on better understanding the fluid instabilities that can prevent fusion from being achieved.

Kaman turns to HPC for realistic numerical simulations of 3D turbulent mixing. The multi-fluid Navier-Stokes equations she relies on are computationally very expensive, and obtaining accurate and precise solutions requires a lot of computing power. Even small jobs can run over 24 hours on thousands of cores.

“Without HPC,” Kaman says, “it would not be possible to develop a production-quality multiphysics simulation package that supports a range of physics as well as perform validation and verification studies.”

Kaman and her colleagues have been able to answer some long-standing problems in turbulent mixing by using computational simulations. They have focused on modeling the radiation energy and its evolution by studying the growth of hydrodynamic instabilities and the resulting mixture during the inertial confinement of the fusion implosion process. At the same time, they have developed new parallelization strategies to better take advantage of HPC systems—ultimately simulating fusion energy and bringing the concept one step closer to reality.

“From biology to chemistry to physics and beyond, our Fulbright College researchers are taking full advantage of computing resources to ask bigger questions and solve bigger problems,” says Todd Shields, dean of Fulbright College. “With so many talented researchers at the college tapping into the transformative power of high-performance computing, just imagine where their curiosity may lead them and what they might discover next.”

Kevin Jackson

ScienceNode.org