A discussion facilitated by Michela Taufer, PhD, General Chair, SC19
On the surface, two highly-popular computing industry conferences – SC and SIGGRAPH – would appear to have little in common.
The former draws stakeholders throughout the technical computing community who are involved in high-performance computing (HPC), all driven by an urgent need to provide computational solutions for the world’s greatest challenges. And the latter is the world’s largest conference on computer graphics and interactive techniques.
Dig a bit deeper, however, and you’ll find that these two powerhouse conferences are quite complementary, given the increased use of simulation data and visualization to improve the discovery of science; and the pervasive impact HPC is having across countless industries, including the development of computer graphics solutions.
And to further that notion, I’ve asked two of the conferences’ leaders to share their views on how SC and SIGGRAPH are furthering the advancement of innovation across our increasingly ubiquitous and diverse computing fields:
Janine C. Bennett is acting manager of the Extreme-Scale Data Science and Analytics Department at Sandia National Laboratories. She is serving as chair of the Scientific Visualization & Data Analytics Poster Showcase for SC19.
Olga Sorkine-Hornung is professor of computer science at ETH Zurich, where she leads the Interactive Geometry Lab. She is serving as Technical Papers chair for SIGGRAPH 2019.
Michela: For your respective roles at SC19 (Janine) and SIGGRAPH 2019 (Olga), what are you hoping to accomplish?
Olga: SIGGRAPH is kind of like the Oscars of our field. It’s our biggest and most prestigious conference, and definitely the top place for scientists to go to present research and hear about other people’s work.
My goal is to maintain Technical Papers as the top venue for disseminating research in computer graphics technology and interactive techniques. With so many new technologies and trends in computer science, we are opening it up to be more inclusive in such areas as AI, machine learning, VR and visualization.
Janine: The SC19 Scientific Visualization & Data Analytics Showcase provides a forum for this year’s most instrumental videos in HPC. It’s a platform for visually communicating scientific discoveries, as well as showcasing innovations in high performance visualization and data analytics.
To emphasize the importance of both types of video, this year we will have two submission categories: explanatory visualizations, which aim to convey a science story in a manner accessible by a broad audience; and exploratory visualizations, which illuminate research discoveries in a format targeted towards science domain experts.
Michela: Do you see SC and SIGGRAPH as complementary, or do they serve distinctly different audiences?
Janine: Both highlight emerging technologies as well as hardware and software advances in computing. In my opinion, SIGGRAPH’s technical program is centered on computer graphics, animation, virtual reality and gaming; and SC’s technical program is centered on HPC research that addresses the technical needs of scientific and engineering research communities.
But I do think there are a lot of parallels between the two conferences, both of which focus on compute capabilities and heterogeneous platforms – with one more focused on industry (SIGGRAPH) and one more on science (SC). I see a lot of potential for cross-fertilization of ideas between the two communities.
Olga: I think that maybe 10 years ago or so, the topics our conferences covered were more separate. But in recent years, that was changed. Supercomputing permeates more and more of computer graphics, because we have so much new technology to acquire massive amounts of data from the real world – like scanning entire cities in 3D.
This flood of data and research in computer graphics poses a challenge in our computing capabilities that we’ve never seen before.
Michela: Can you provide a brief overview of your own research interests?
Olga: My research is in the geometry part of computer graphics, which is called geometric modeling or geometry processing. Basically, I’m interested in digital representation of 3D objects; how to create them on a computer, or how to process the geometry that we acquire in the real world through 3D scanning.
Janine: My research includes developing scalable data analysis techniques for science and engineering applications. What I’m most passionate about is working with scientists and engineers to help them define features in their data, track them in space and time, and develop algorithms and tools to enable them to do that. Over the years, my work has evolved to include a broader HPC systems-level emphasis.
Michela: What about the future? How do you imagine HPC and visualization will impact science in the next five years?
Janine: Today, we have desktop computers and small clusters that match the compute power of the biggest supercomputers from a decade ago. Over the next five years, HPC innovations will continue to support this increased accessibility to compute resources, empowering researchers from a broader set of technical backgrounds than was ever possible before. Taken together with future innovations in large-scale visualization, I am excited to see how these advances will democratize the dissemination of scientific insights to the general public.
Olga: I feel like we’re at a turning point. This whole explosion of data science, machine learning and neural networks is something few of us could have anticipated even five years ago. In my opinion, this emergence of AI technology and evolvement of neural networks is not something that is temporary. It’s going to stay. And that means it will affect more and more of the research we are doing, and its outcomes.
Michela Taufer, PhD, General Chair, SC19
Michela Taufer is the Dongarra Professor in the Min H. Kao Department of Electrical Engineering & Computer Science, Tickle College of Engineering, University of Tennessee, Knoxville.