High school student learns HPC with Anvil
Last Spring, under the guidance of PhD student Anastasia Neuman, a high schooler completed her senior capstone project by conducting research utilizing Purdue University’s Anvil supercomputer. The pair used Anvil to run simulations elucidating the thermodynamics of nanocomposites.
Sarah Will recently 
graduated from the Science and Mathematics Academy at Aberdeen High School (SMA) in Aberdeen, Maryland. SMA is a school-within-a-school magnet program that provides academically talented students with educational experiences that integrate science, technology, engineering, and mathematics beyond the traditional advanced program. As part of her senior year graduation requirements, Will had to complete a year-long capstone project wherein she worked alongside a scientist in any field she wished. Will chose Neuman as her mentor, who taught her the ins-and-outs of high-performance computing (HPC) through a hands-on, real-world research project.
Anastasia Neuman is a PhD student in the Chemical and Biomolecular Engineering Department at the University of Pennsylvania (UPenn). She focuses on computational research that utilizes HPC resources for simulations involving polymer physics. In her role at UPenn, Neuman often works with experimental polymer physicists, taking their experimental data and figuring out how to create a model that explains the underlying physics of the project. Neuman also happened to graduate from SMA and is very familiar with the capstone process, as she had to undergo it herself. As such, Neuman was thrilled to become a mentor for Will. The pair decided to expand on research previously conducted by Neuman, which looked into how confinement within nanoparticle packings affected the miscibility (the ability to be mixed at a molecular level to produce one homogeneous phase) of a bulk polymer blend. For this project, the two used Anvil to simulate the effects nanoparticle packings have on block copolymers (BCPs). The results of the new project were unexpected, but will help experimental researchers produce materials with specific BCP phase structures. This could lead to novel polymer properties (e.g., improved malleability or conductivity), which could lead to solutions for problems such as CO2 separation, rechargeable batteries, food packaging, and tissue engineering.
Polymers are materials and substances made up of long, repeating chains of very large molecules, called macromolecules. These chains are called monomers. Different monomer types will exhibit repulsion when combined due to chemical dissimilarity. Using HPC simulations, both projects aimed to determine the order to disorder transition (ODT)—the temperature at which the polymer achieves phase separation, i.e., goes from being mixed randomly to separated.
In Neuman’s original project, she studied a homopolymer bulk blend, in which there was only one monomer type per chain. Half of the mixture was made up of one type of monomer chain, the other half consisted of a secondary type. The research showed that nanoconfinement increased the ODT, and therefore, the miscibility in the blend. But Will and Neuman found that the opposite was true for the BCPs.
BCPs are made up 
of blocks of two different monomer types, all connected linearly on the same chain. Using Anvil’s GPU nodes, Will and Neuman employed a simulation technique known as self-consistent field theory, allowing them to observe the thermodynamics and morphology of BCPs confined within nanoparticle packing. The pair found that, with BCPs, the strength of the repulsion required to induce phase separation decreased as confinement increased, a surprising result given what Neuman found for the bulk blend. But regardless of the unexpected difference between the bulk blend and the BCP, Will and Neuman were thrilled with both the results and Anvil’s performance.
“Anvil was really great,” says Neuman. “So we are running 3D simulations involving multiple inputs. We make these random nanoparticle packings with a dynamic simulation, and then we feed those into our field theory simulations. It is very difficult to run these simulations without good GPU nodes. I basically use the [other] GPU nodes I have access to for testing changes in the CUDA code, but I can’t get results without Anvil.”
Neuman was happy not only with Anvil’s GPU nodes, but also with the supercomputer’s accessibility and ease of use. Will was a first-time HPC user. She had no experience working within a terminal and was unfamiliar with HPC server environments. On top of that, she would also need to do her computational work from her high school’s computers, which presented its own unique set of issues.
“One challenge that Anvil helped overcome was the issue of software. On the high school computers, users aren’t allowed to download any software. So how are you supposed to do any computational work? But thanks to Open OnDemand, Sarah was able to log into the cluster via a web browser. The browser wasn’t blocked by the school, so she could work on the project in the classroom without having to download anything. It was really nice.”
Neuman continues, “I also think it’s easier for students because they are used to working with a web browser more than they are a terminal. So being able to access Anvil with Open OnDemand made it a lot more user-friendly and a great introduction to computational work. I’ve even recommended Anvil to my supervisor at UPenn, who teaches many classes that introduce computational work to undergraduates.”
To wrap up her capstone project, Will successfully presented her research via a poster presentation at her school. She has since graduated and is now enrolled in the chemical engineering program at the University of Maryland, Baltimore County (UMBC). Neuman noted that she will definitely be keeping in touch with Will and is encouraging her to continue pursuing computational research.
To learn more about HPC and how it can help you, please visit our “Why HPC?” page.
Anvil is Purdue University’s most powerful supercomputer, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States.
Researchers may request access to Anvil via the ACCESS allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.
Written by: Jonathan Poole, poole43@purdue.edu