News

Student threads computers together for science


JOSEPH CANNING | canni001@rangers.uwp.edu

At UW-Parkside in a small, unassuming room tucked away in the second floor of Greenquist Hall, dozens of computers of all sorts are buzzing incessantly. These computers have a common purpose—they are a unit, joined by a jungle of wires to run billions of calculations every minute for the benefit of science.

This unit is the pet project of a single Parkside student: Evan Macintosh.

Macintosh_Computer_B&W

Evan Macintosh poses in front of his computers as they run calculations.

The beginnings

“It started half as a joke,” said Macintosh, smiling. In 2015, a friend of his had facetiously suggested they set up a Beowulf cluster, a method of parallel computing. The method involves linking together many consumer-grade PCs. To his friend’s surprise, Evan responded seriously: “Let’s do it!”

Soon after, the two collected old PCs and made their first cluster in one of their basements. It was a modest affair and has since been disassembled, but it gave Macintosh the necessary experience to expand the project.

A new chapter

On campus, he found new computers and a new space with which to set up a better cluster. With the guidance of physics professor William Parker, Macintosh set up a new project in a tiny closet.

The research work the cluster was performing eventually demanded more space, and the physics department allowed the computers to be moved to their current location. Now the cluster has grown to include powerful computers with Xeon processors and nVidia GPUs as well as more mundane hardware.

Math machines

The research that’s being done is on the behavior of electrons within atomic structures. The cluster runs complex physics simulations based on density functional theory of these structures using a software amusingly branded Quantum ESPRESSO.

Because the math involved in the simulations is so intense, tremendous processing power is required. Macintosh elegantly explained the advantages of parallel computing with an analogy: “Let’s say you have a test with 144 questions, and you can do one question a minute; the test takes you 144 minutes. Now let’s say you have a friend… now that takes you 72 minutes.”

Currently the simulations being run are what he referred to as “baby steps.” They only reproduce one monolayer (eight atoms) thick material, but he hopes to eventually simulate materials with thicknesses of up to 20 or 50 monolayers. The calculations involved in representing even a single monolayer of material takes days to complete. Future endeavors will not finish for weeks.

An ever-evolving project

Professor Parker and Macintosh have been awarded $9000 in grant money to continue their research, and they hope to spend it on better hardware. About the project, Macintosh said, “It started out as credit, now it’s just turning into a personal project… I see myself doing a lot of this later in life, and I think… this will increase my skills in this area.”

If you are interested in parallel computing or with getting involved in physics research, Macintosh said he is open to more students joining in. You can contact him at his email—macin001@rangers.uwp.edu.

Categories: News

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s