The Scope and Impact of Supercomputing Seemingly Has No Boundaries
One is in the field of digital humanities, an emerging area of study that amplifies and extends humanities research through the use of such advanced information technologies as technology, media and computational methods. An academic who is striving to elevate the use of HPC within digital humanities is Dr. Mariann Hardey, associate professor in the Department of Management and Marketing at Durham University Business School and a member of the Advanced Research Group (ARC) at the University of Durham.
Digital Humanities is an extensive area of study that brings people from multiple disciplines together to better understand and address issues in human society and culture,” says Mariann. “And we are finding ourselves on the cusp of possibilities in digital humanities through the utilization of supercomputing.”
One of her proposed trailblazing projects involves using HPC to process and create a visual data analysis based on fifty years of images from South Africa that focuses on engagement, participation, and protest on the road to democracy in the region. As Mariann explains it, the HPC-based project was designed on “building up a corpus of data for analysis, and then sharing the cultural index (findings) with research partners” in various humanities disciplines. This cultural index would enable comparison cross-nationally to better understand the role that political protest plays in political action.
According to Mariann, who has been working on digital humanities research projects for five years at Durham, the field has exploded recently. Part of that increase in activity has stemmed from efforts to introduce the concept of digital humanities to researchers and explain how HPC can potentially impact their work.
One project, for example, has been an investigating through image analysis how supercomputing could be used to foster better and more meaningful interactions online, particularly in the arts – a field that was severely impacted by the pandemic. Others have concentrated on Research Software Engineering (RSE) and how those in that field can develop computational processes and best practices to scale digital humanities projects from the desktop to the HPC cluster.
There has been a hive of excitement, as the technology and study of digital humanities has really started to match with our aspirations,” notes Mariann.
Supercomputing & the Sounds of Music
Researchers at the University of Edinburgh are using HPC to push the limits of possibility in two areas very relevant in music – synthetic sound and room acoustics.
Funded by the European Research Council, the NESS (Next Generation Sound Synthesis) project – under the direction of Stefan Bilbao – is attempting to use HPC techniques and best practices to explore new ways of creating sound-producing systems. Their algorithms are being used to build simulations of musical instruments and the spaces they play in, as well as create new instruments and sounds based off those that exist in real life.
Two of the project’s application consultants – Larisa Stoltzfus, who has recently completed her PhD in informatics, and software architect James Perry – have a special interest in NESS as they are also musicians. Both are pianists and James plays the viola as well.
It was interesting to work with musicians and get their feedback on the sound simulation,” says James, whose primary responsibility for the project was to port and optimize the code developed for NESS to HPC, a significant challenge. “Taking their knowledge, translating it into improving our work, and then hearing them play the simulated instruments back to us made for an interesting day at the office.”
Larisa joined the project at a later stage, and was charged with ensuring that the work being done was portable so that it could be run on different architectures without retuning or rewriting the codebase. “The method behind this involved taking these algorithms and breaking them down into building blocks, and then finding the best configuration for them across different architectures,” she explains. “Originally the codes only ran on Nvidia GPUs but we wanted to make the code as accessible as possible.
As the project progressed, James and Larisa saw the similarities at the code level that translated into benefits for both arms of the project – synthetic sound creation and room acoustics. “For example, we learned that drum simulations could benefit acoustic simulation design,” notes James. “When you simulate a drum, you look at the inside of the drum and the membrane – and this same method could be translated to looking at acoustic space design and simulating how sound travels.”
Adds Larisa, “Because a lot of simulations share the same origin mathematically, there is a larger benefit to being able to reapply a similar scientific methodology in order to benefit more than one field.”
This November, at SC21 in St. Louis, we’ll explore the ways HPC is fueling breakthroughs across every imaginable industry by harnessing data to unlock insights faster and more accurately than ever before.
Cristin Merritt, SC21 Inclusivity Liaison for Communications