The second publication in the ACS Technology White Paper series is on the topic of high-performance computing, or “HPC.”Written by Academic Computing Consultant for the Sciences Chris Gahn, it explains how researchers— often in the sciences, but not always— are turning from expensive lab computers to powerful remote computers for their number crunching.
What is HPC?
Have you ever tried to run some code or perform some data analysis on your personal computer, and it ended up taking several hours or even days to finish? Frequently, we interact with data sets or projects that are simply too taxing for a single computer (even a very powerful one) to complete. This happens very frequently in a variety of fields across disciplines. High-Performance Computing (or HPC as it is known) is a field in technology concerned with providing advanced computing resources to researchers in order to speed up their data processing or modeling projects. Typically, these computing resources take the form of what is known as a computing “cluster”, which is really just a fancy name for a large number of computers that are all connected together and process data in unison.
Who can benefit from HPC?
HPC at Vassar (on-campus)
Here at Vassar, we have a computing cluster named “Junior” that was built in 2010. Junior has been used by many faculty and students over the years to run countless analyses and simulations for coursework and research in the Sciences and Humanities. The big advantage of using a system like Junior is that it has what’s called a job scheduler program installed. In Junior’s case, the scheduler program is called SLURM. Slurm enables users to submit the code or analysis that they want to run, and then the system will automatically load the required packages and software to complete the job, and output it into a user-specified format. This means that a user can submit a job that might take the system several days to finish, and then go and work on something else while awaiting the results, confident that work is being performed by the automated processes on the computer the whole time.
HPC at Vassar (off-campus/remote)
Here at Vassar, we have access to off-campus HPC resources as well:
- Through an agreement with the NSF-supported XSEDE system (The Extreme Science and Engineering Discovery Environment), Vassar researchers are able to apply for computing allocations on a variety of cluster environments that provide abundant libraries of packages, software, compilers, and user interfaces. The best part – It’s completely 100% free for Vassar! Your tax dollars are hard at work creating and maintaining this extensive network of HPC resources for use by all researchers.
- Vassar has an agreement with Amazon Web Services (or AWS) to provide for Virtual computing environments hosted in Amazon’s many data centers around the country. While we do pay by the hour for resources through AWS, the scalability and versatility afforded to us through this system are incredibly useful. Computing environments can be built and made accessible to the end-user very quickly and easily by administrators on campus.
- We are exploring additional resources such as Google Cloud, Microsoft Azure, products from IBM, and partnerships with other colleges and universities with more robust computing infrastructure. We are also looking into avenues for the upgrade and/or replacement of Junior.
HPC Projects & Initiatives at Vassar
Many faculty at Vassar have been involved in using HPC in the course of their research and teaching. Courses and projects in Biology, Chemistry, Cognitive Science, Computer Science, Mathematics, and Physics & Astronomy all make use of Junior and other HPC resources for course work and projects.
Just a few of the specific HPC projects underway or already completed include:
- Chemistry Molecular research by Franco Caruso and Miriam Rossi which utilized Materials Studio in a cluster environment and has resulted in the publication of two successful journal submissions with a third in progress.
- Biology research on viruses and bacteria using QIIME and other genetic analysis tools on both the local cluster and in AWS by David Esteban.
- Deep learning research and course work using GPU-enhanced computational systems in the cloud by Joshua de Leeuw.
- Computational Quantum Chemistry research by Leah I. Bendavid on XSEDE.
Find out More!
If you’re interested in learning more about HPC, or getting in touch with other people at Vassar who are using HPC resources, please email Chris Gahn, the ACS Consultant for the Sciences.
The VSA organization Vassar Urban Enrichment asked us to create an aerial image of the entire main campus (i.e. not including the Vassar Farm, Townhouses, golf course, etc.) After a few tries with the ACS drone, we determined that the wifi signal it uses to communicate with its controller was insufficient for handling the distance we needed and the occasional obstacles in between. Drone pilot Chad Fust then used his own drone, which uses an RF signal, rather than wifi, and we were able to complete the project in two sections. The result is a merge (“orthomosaic”) of about 900 individual photos.
Prof. Lynn Christenson of the Environmental Studies program and Keri VanCamp of the Collins Field Station are interested in using the drone to acquire various types of imaging of the Vassar Farm and Ecological Preserve. One area of focus is the beaver pond, which they’d like to view from above at different times of the year and over the years. After several unsuccessful attempts, we were able to collect a series of 200 images and stitch them together into the following visualization.
In December 2016, Environmental Studies major Rachel Marklyn asked us to produce an aerial view of the succession plots, on the environmental preserve. The succession plots are designated areas of a field that receive different treatments— mowing, tilling, and goat grazing— at different intervals, in order to study the long-range effect of those treatments on forest growth.
For Rachel’s senior project, she was creating a series of information signs for the preserve, one of which was about the succession plots. The wintertime photo was not very colorful, but the borders of the different plots were clear.
ART 386, Embodying Compassion in Buddhist Art: A Curatorial Training Seminar was taught by Karen Lucic during Fall semester, 2014. The purpose of the class was to give students the opportunity to research and curate an exhibition at the Frances Lehman Loeb Art Center. During the summer of 2014, Professor Lucic contacted me to discuss creating a website for her students to use as a repository for their research and eventually this site would become a companion site to the final exhibition in April, 2015. Over the course of the Fall semester, students wrote and compiled content for the site and they worked closely with ACS to design and populate the site.
From ART 386 Syllabus:
Each student will be responsible for the digital content and interpretation of 2-3 works in the exhibition. The instructor will assign the objects to each student, based on her/his experience and preparation. These student contributions will be posted on the exhibition application and/or website. (60% of grade.)
Students will work in teams to produce additional resources for the exhibition: gallery guide, interactive maps, guide to web resources, etc. Students in the team will also give feedback on other team members’ work before submission (20% of grade.)
From ART 386 Assignment Sheet:
The purpose of this assignment is to create digital educational content for the exhibition. Always remember who your audience is: visitors to the exhibition, or online users, who might not know much—or anything—about the topic. What you write, and your choice of materials should be based on your assessment of what will enhance their experience and understanding of the exhibition. Texts should be concise and to the point. Other materials should be short but engaging.
For each work you have been assigned:
1) Write an interpretative text (no more than 100 words) for app/website
2) Select one comparative image (must be open access and high resolution); include full caption of comparative image
3) Write a text (no more than 100 words) explaining the comparison
4) Select an audio file, if possible, that enhances the work (no more than two minutes). Examples: chanting, singing, mantra recitation, etc.
5) Select a video file, if possible, that enhances understanding of the work (no more than 2 minutes). Examples: practitioners circumambulating, prostrating, spinning prayer wheels, making sand mandalas, offering incense, etc.
6) If there are no appropriate audio or video files, choose another comparative image.
7) Compile a list of unfamiliar terms from your texts, with definitions
8) Map your work, at least by country. With some works (Putuoshan, Nachi, etc.), it may be possible to be more precise about locations.
9) Record your written contributions.
While students were working to create the content for the site, ACS student employee Bryce Daniel worked on building a wireframe for the WordPress site. Professor Lucic also collaborated with Duke University students to design an App for the exhibition. The App hosted audio files recored and edited by ACS Consultant, Baynard Bailey. These recordings, narrated by both students and Professor Lucic, include short commentaries describing individual pieces in the exhibit, as well as a pronunciation guide for a glossary of terms.
The Embodying Compassion WordPress site is a comprehensive online exhibit reference guide, containing audio, video, images, interactive hotspot maps, and extensive research, curated by ART 386 students. This project proved to be a excellent example of how students, faculty, and ACS consultants collaborate to produce educational materials for the classroom and public audience.
The Frances Lehman Loeb Art Center: Embodying Compassion is a first-of-its-kind exhibition celebrating one of the most important figures in Buddhist art, April 23-June 28, 2015
ACS has a portable scanner that’s available for Vassar faculty members to borrow. It’s ideal for bringing to an archive to scan documents (if allowed) for later study. While it’s true that you can save an image of a document page with your phone/camera, a scan will give you a much more consistent and high-quality result.
Most scanners are too bulky and heavy to consider traveling with, but not this one. The Canon Canoscan LiDE 700F is essentially the size of a laptop computer. It’s less than 2 inches high, with a footprint of 11.5 x 16 inches. It weighs only 4.6 pounds. It can scan a letter-size document (8.5″ x 11″) with a resolution of up to 4800×4800 dpi. It fits snugly in a 12″ x 17″ 2.5″ padded carrying case.
Contact your ACS liaison if you’re interested in borrowing it.