The Yin and Yang of Understanding Data

By Nicole Hemsoth

November 3, 2006

When the Department of Energy's Office of Science announced the latest round of awards in the Scientific Discovery through Advanced Computing (SciDAC) program in September, the funded projects included a new Center for Enabling Technologies that will focus on meeting the visualization and analytics needs of scientists. Called the SciDAC Visualization and Analytics Center for Enabling Technologies, or VACET, the project will be co-led by Wes Bethel, head of the Visualization Group at Lawrence Berkeley National Laboratory, and Chris Johnson, director of the Scientific Computing Institute of the University of Utah. The VACET team also includes researchers at Lawrence Livermore and Oak Ridge national laboratories, the University of Utah and the University of California at Davis.

At the upcoming SC06 conference in Tampa, Bethel will give an overview and discuss the objectives of VACET in a talk at the LBNL booth (1812) at 11 a.m. Wednesday, Nov. 15. In the runup to SC06, Bethel took some time to talk about the field of analytics, how it contributes to computational science and where the fields of analytics and visualization are headed.

Question: First of all, how did the team for your SciDAC project come together?

Bethel: We wanted to strike a balance where we have members with research and production backgrounds. The result is a team with members from well-known university and DOE laboratory research programs along with representation from DOE's production visualization programs. This kind of blend will assure that we have top-quality research and development as well as the means to deploy the new technology at the computational facilities where the data will be computed, collected, stored and made available to distributed teams of SciDAC science researchers.

Chris Johnson from the University of Utah and myself share leadership duties of this project. Our team consists of Valerio Pascucci, Hank Childs and Peer-Timo Bremer from LLNL; Ken Joy and Bernd Hamann from UC Davis; Chris Johnson, Chuck Hansen, Claudio Silva, Steve Parker, Allen Sanderson, Xavier Tricoche and Marty Cole from the University of Utah; Sean Ahern, George Ostrouchov, and Jeremy Meredith from Oak Ridge National Laboratory; and of course Cristina Siegerist and myself from Berkeley Lab.

Every person on this team is “a superstar.” In four of the past eight years, members of this team have authored or co-authored Best Paper Award-winning papers at the IEEE Visualization conference. All have an impeccable record of publications and are recognized leaders in the field. But not only that, this team has architected and developed a vast amount of visualization software that has had a broad impact in DOE and NSF science and research programs. Not only are they first-class professionals, they are also fun to hang out with. One thing that is special about this team is that we all get along great. This combination really makes this team special to me. I'm honored to be working with such a great bunch of people.

Q: Last year, SC05 incorporated analytics into the conference program and this year, DOE agreed to fund your proposal to establish a center for analytics and visualization. Clearly, it's an emerging field. What's your definition of analytics and how does it fit into HPC?

Bethel: That's a question that reminds me of freshman philosophy class and the question, “What is love?” You can give a number of examples, but it's hard to define.

We can start by defining scientific visualization, which is the process of creating images from abstract scientific data. Analytics has been described as the science of analytical reasoning. To me, analytical reasoning means being able to draw conclusions based on hypothesis testing and the exploration of large, complex and occasionally incomplete data. Visual analytics is a way of facilitating such research through the use of interactive visual interfaces. Visualization and analytics are complementary technologies with no clearly defined boundary between them. Visualization helps accelerate analytics by relying on humans' vast cognitive processing abilities, the “yin” side of data understanding. Analytics gives hard, quantifiable measures, and is the “yang” side of data understanding.

An example of analytics could be how the California Department of Forestry responds when a forest fire is reported. Their incident response is based on figuring out what's wrong, what they need to do and what resources they have to do that. In simulation science, lives and property may not be at stake, but there can be substantially more data involved. For example, policy makers want to understand how greenhouse gas emissions at various levels will impact global climate. Climate scientists study this question by performing lots of simulation runs where they vary the input parameters to the simulation. Visualization and analytics are tools for discovering relationships between cause and effect in a highly complex system that produces many terabytes worth of data. Neither of these is really a complete “definition of analytics,” but both are examples of how analytics is put into practice to solve complex problems.

As DOE and other agencies progress toward petascale computing, we are increasingly able to produce data more quickly than we can understand it. And visualization by itself is not enough to get the job done. We can make an image of a terabyte of data, but the result is so complex that it can be impossible to gain any insight at all. Scientists are probably not interested in that whole terabyte. More typically, their interest is “Did the science I'm interested in happen somewhere in this terabyte of data?”

Another important dimension to the problem is one of data management. With all the excitement and effort being applied to more highly resolved simulations that run on larger machines, we can sometimes forget that storing, accessing and processing a petabyte worth of data is not a trivial task. The SciDAC program funds efforts in Scientific Data Management to work on these very problems. Our center has a very close relationship with Arie Shoshani's Scientific Data Management Center — large-scale visualization and analytics solutions will, by definition, incorporate technologies for large-scale data management. The two issues are really inseparable.

In our world, analytics is best thought of as a combination of visualization, data analysis, data management and scientific workflow management.

Q: In your proposal for VACET, you mention that visualization and analytics R&D need to be more attuned to the needs of scientists and engineers. How are you planning to address this?

Bethel: At the end of the day, any software tool, be it visualization, database, or even word processing, the scientist or engineer has to be able to say, “This tool or technology helped me do my job.  It helped me discover some new scientific phenomenon or it helped me eliminate some time-consuming, labor-intensive activity so I can focus more time on science.” Ultimately, we measure our success by our positive impact on science and that impact can come in many different forms.

We are engineering our project in such a way that the needs of a specific set of science applications provide the basis for our R&D. For example, our computational astrophysics customers are interested in what triggers supernova explosions. And they know what lines of scientific inquiry they need to follow to answer that question. We attend meetings with them and talk to members of their team to better understand what new visualization and analytic tools they need to do new science. We're doing business like this for half a dozen scientific applications. Our work plans are engineered to deliver the technologies our customers believe they need to do their science.

One theme common across all our customers is that they are overwhelmed with data. We're all aware of how the technological advances in recent years have produced faster computers, bigger disk drives, better networks and so forth. One result from all these technology advances is an “information big bang” — we're literally creating information faster than we can understand it. Overcoming that limitation — being able to understand data at the rate it is generated — is often viewed as being one of the bottlenecks in contemporary science. Stated differently, those who have the ability to quickly draw insight from data have a competitive edge. Our Center takes direct aim at solving that problem for our customers.

Q: Speaking of doing business, you could say that SciDAC centers are in the business of creating production software. What is VACET going to do in this regard?

Bethel: What we're doing is leveraging two established, well-known efforts in visualization as the basis for production deployment of new visualization and analytics technologies capable of tackling current and future scientific datasets: One is the VisIt application from LLNL, which was developed under ASCI for performing parallel and distributed visualization and analysis of ASCI's large datasets, including those from BlueGene/L. VisIt is already in widespread use for visual data analysis, runs on basically all parallel platforms, and won an R&D 100 award in 2005. The other is SCIrun, the University of Utah application that has been successfully used for providing both domain-specific visualization applications as well as the infrastructure for supporting visualization research. Both are open source visualization environments.

Furthermore, we plan to layer new approaches atop this existing infrastructure to deploy at DOE's large computing facilities, as well as on desktops. All of this software infrastructure will be freely available as open source software for use by the community.

Our team feels it is important to minimize R&D risk while maximizing the impact on our science customers while providing a solid substrate on which to conduct future analytics and visualization research for working with petascale data.

Q: VACET was approved as a five-year project. Looking ahead to 2011, what do you think will be the state of visualization and analytics software?

Bethel: DOE plans to have petascale systems in operation by that time. I would expect we would be quite adept at interactive visual analytics of petascale datasets at that point. If what the science projects have planned comes true, we can expect that datasets will also become a lot more complex, thereby exacerbating the need for highly capable analytics and visualization technology. Going back to the climate change example, our climate customers are working hard on coupling many different types of computational models — atmospheric, ocean, geo- and biochemical, among others — into a single coherent whole. These large, multi-grid datasets will span multiple of orders of magnitude in time and space and will have hundreds of fields per zone. There are similar plans in many other areas of science as experts in specialized fields combine forces to challenge what can be thought of as “whole model” simulations. My friends in the experimental sciences tell similarly awe-inspiring stories of the large amounts of data they will collect that they will need to analyze and visualize. As with all good science projects, there exists now a need to compare simulation and experiment. As the data grow larger and more complex, performing such comparisons becomes increasingly challenging.

To some extent, the field of visualization faces many of the same set of challenges faced by the rest of the HPC scientific community: moving targets for standards, changing machine architectures, increasing parallelism, impedance matching between computational power, memory hierarchy, I/O and networking speeds. We can't solve those problems on our own; that's why programs like SciDAC — that integrate math, computer science and domain sciences — are so important.

—–

Wes BethelWes Bethel, a staff scientist at Berkeley Lab, is the author of more than 60 technical publications covering themes such as remote and distributed visualization, visualization architectures, high performance visualization, visualization algorithms, and applications of visualization to understand data in many different scientific domains. He has served as PI or co-PI on 15 different DOE-sponsored graphics and visualization research projects over the past 10 years. He earned his BS in Information Systems in 1984 and MS in Computer Science in 1986 from the University of Tulsa. He is a member of ACM/Siggraph and IEEE, and serves actively in the visualization community as a technical reviewer for many different publications and as a program committee member for IEEE Visualization. He was also co-leader of the teams that won the first three Bandwidth Challenges at the SC conferences, beginning at SC2000 in Dallas.

—–

Source: Jon Bashor, Lawrence Berkeley National Laboratory

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

Nvidia Appoints Andy Grant as EMEA Director of Supercomputing, Higher Education, and AI

March 22, 2024

Nvidia recently appointed Andy Grant as Director, Supercomputing, Higher Education, and AI for Europe, the Middle East, and Africa (EMEA). With over 25 years of high-performance computing (HPC) experience, Grant brings a Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Houston We Have a Solution: Addressing the HPC and Tech Talent Gap

March 15, 2024

Generations of Houstonian teachers, counselors, and parents have either worked in the aerospace industry or know people who do - the prospect of entering the fi Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire