How HPC is Shaking Up Modeling of Mysterious Earthquakes

By Oliver Peckham

July 6, 2021

Earth modeling is extraordinarily complex – and when the earth quakes, it’s no different. The mechanisms of many earthquakes consistently flummox seismologists, making it difficult to understand why those earthquakes occurred, let alone perform the most important and elusive function in seismology: earthquake prediction. In a talk during ISC21, Alice-Agnes Gabriel, a professor of seismology at the Ludwig Maximilian University of Munich, explained how HPC-enabled earthquake codes have worked to close these gaps and take advantage of quickly accelerating supercomputing capabilities.

Dr. Alice-Agnes Gabriel

“Computational seismology has been a pioneering field for high-performance computing,” Gabriel said, “and [it] has itself been pioneered by HPC in order to image Earth’s interior to understand the dynamics of the metal, for example, and also to track down energy resources.” Computational seismology, she explained, just made sense for HPC: seismology is data-rich, and many of its dynamics could often be treated as linear systems.

On the other hand: “If you’re thinking about computational earthquake seismology, we’re facing a very different picture. Computational earthquake seismology is trying to understand earthquakes in a physics-based manner – we’re not imaging earthquakes, but we’re trying to model earthquakes based on physical first-order principles. We’re solving for spontaneous dynamic earthquake rupture as non-linear interaction of frictional failure across prescribed zones of weakness that are geological faults, that can take complicated shapes and interact with each other … and these frictional failure processes are linked to seismic wave propagation in a non-linear manner.”

So, in layperson’s terms: hunting for oil deposits, for instance, is easier than figuring out how many earthquakes occurred. And, indeed, there are many baffling earthquakes: earthquakes that do more damage than seismologists expect, earthquakes that occur along seemingly low-tension faults and even earthquakes that jump.

To model earthquakes, Gabriel explained, seismologists combine knowledge from a wide range of fields: seismology, geodesy, geology, tectonophysics, hydrology, numerical computing, data science, machine learning, applied mathematics, continuum mechanics, tribology, rock mechanics, materials science, and engineering. They amalgamate these processes in bootstrapped models, developing solvers that are able to read data from fault stresses and output rupture dynamics, ground deformation and synthetic observables (like simulated seismograms).

These models are robust, but three major challenges remain: first, earthquake source processes are very ill-constrained and highly non-linear; second, researchers are still working to identify which physical processes are dominant and relevant in real earthquakes and how to computationally budget for those processes’ inclusion in models; and third, teams struggle to assimilate all available knowledge in a manner both hardware- and software-suitable for modern HPC systems.

Progress on these challenges has, thankfully, been rapidly accelerating. Gabriel cited some of her team’s prior research from 2014 – itself a Gordon Bell Prize finalist – that modeled a large 1992 earthquake in California that confused researchers by seemingly jumping from fault to fault without a continuous connection. Ever since, she said, researchers had been modeling earthquakes at a petaflop scale, linking spontaneous dynamic rupture simulations with fault geometry and a range of other physical fields.

In the intervening seven years, of course, the scale is increasing: researchers are working to port those models to ever-larger spatiotemporal ranges in order to capture megathrust earthquakes (which represent all recorded earthquakes over magnitude 9.0) and tsunamis. “The extreme scale of megathrust earthquakes and tsunamis,” Gabriel said, “also requires extreme-scale runs.”

The investments required for such runs are rapidly decreasing. A “hero run” simulation of the 2004 Sumatra earthquake and tsunami, Gabriel said, took 14 hours on LRZ’s SuperMUC Phase 2 (2.8 Linpack petaflops) in 2017; now, typical runs take under six hours on just 16 nodes of LRZ’s SuperMUC-NG (19.5 Linpack petaflops).

Gabriel’s team uses a seismic solver called – appropriately – SeisSol. Using SeisSol, the researchers can solve seismic wave equations on unstructured tetrahedral meshes, combining complex geometries with heterogeneous inputs and high resolutions. Initially, SeisSol was simply a Fortran-based solver with MPI parallelization. Over time, it evolved: in 2014, for simulations of the 1992 California quake, the team implemented hybrid MPI/OpenMP parallelization, parallel I/O and a multiphysics offload function for many-core architectures, allowing the simulations to reach 96 billion degrees of freedom at 200,000 time steps.

For the Sumatra quake, SeisSol was further optimized through a bevy of features, including asynchronous I/O, overlapping computation and communication, and a feature called cluster-based local time-stepping that allowed each element to have its own time step. These optimizations allowed SeisSol to expand to 111 billion degrees of freedom at 3.3 million time steps, offering a 14-fold speedup and allowing the team to complete Sumatra simulations in hours rather than days. Essentially, Gabriel said, these speedups were the difference between a feasible simulation and an impossible simulation.

Image courtesy of Alice-Agnes Gabriel.

“We’ve been using this approach to shed light on puzzling earthquake observations,” Gabriel said. Lately, for instance, the researchers have been studying the 2016 Kaikōura earthquake, a magnitude 7.8 event that struck New Zealand to devastating – and befuddling – effect, perhaps most notably when the earthquake leapt from one fault to another fault 15km away, three times the previously thought maximum distance for a “jumping” earthquake. Other questions stumped researchers, too: who did the earthquake move so slowly? Why did the Hope Fault, which was thought to pose the greatest seismic hazard in the region, barely react?

A representation of the Kaikōura earthquake simulation. Image courtesy of Alice-Agnes Gabriel.

SeisSol, it turns out, was able to answer many of these questions that would have dead-ended researchers a decade ago. Using the solver, the team was able to show, for instance, that another fault – the Point Kean fault – acted as a crucial link between the two endpoints of the jump, and that unfavorable dynamic stresses on the Hope Fault led to its lack of activity.

The researchers are excited to develop even more functionality, exploring the crevices of earthquake modeling. Gabriel elaborated on how her team is modeling manmade earthquakes, such as the 2017 magnitude 5.5 earthquake in South Korea, and modeling the 2018 Sulawesi earthquake that devastated Indonesia – which was complicated by the unavailability of close seismometers recording the event, leading Gabriel and her colleagues to turn to sparse data from satellites and social media.

Improvements in the code, of course, are complemented by improvements in supercomputing. Gabriel’s team has made extensive use of top-tier supercomputers, including modern leaders like SuperMUC-NG, the 5.5-Linpack petaflops Shaheen II at KAUST, the 5.4-Linpack petaflops Mahti at CSC and the 23.5-Linpack petaflops Frontera at TACC.

All of this will be amplified by the forthcoming GPU port of SeisSol – and, of course, the imminent shift to the exascale era will open up a wide range of further possibilities. Gabriel mused on what they could learn from running the models a thousand times faster: higher resolutions, uncertainty quantification, rapid earthquake response simulations, more multi-physics components – a paradigm shift toward more holistic earthquake system modeling. “The interplay of advances in high-performance computing and dense observations,” Gabriel said, “will allow us to go beyond scenario-based analysis.”

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!

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's latest weapon in the AI battle with GPU maker Nvidia and clou Read more…

ISC 2024 Student Cluster Competition

May 16, 2024

The 2024 ISC 2024 competition welcomed 19 virtual (remote) and eight in-person teams. The in-person teams participated in the conference venue and, while the virtual teams competed using the Bridges-2 supercomputers at t Read more…

Grace Hopper Gets Busy with Science 

May 16, 2024

Nvidia’s new Grace Hopper Superchip (GH200) processor has landed in nine new worldwide systems. The GH200 is a recently announced chip from Nvidia that eliminates the PCI bus from the CPU/GPU communications pathway.  Read more…

Europe’s Race towards Quantum-HPC Integration and Quantum Advantage

May 16, 2024

What an interesting panel, Quantum Advantage — Where are We and What is Needed? While the panelists looked slightly weary — their’s was, after all, one of the last panels at ISC 2024 — the discussion was fascinat Read more…

The Future of AI in Science

May 15, 2024

AI is one of the most transformative and valuable scientific tools ever developed. By harnessing vast amounts of data and computational power, AI systems can uncover patterns, generate insights, and make predictions that Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top500 list of the fastest supercomputers in the world. At s Read more…

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

Europe’s Race towards Quantum-HPC Integration and Quantum Advantage

May 16, 2024

What an interesting panel, Quantum Advantage — Where are We and What is Needed? While the panelists looked slightly weary — their’s was, after all, one of Read more…

The Future of AI in Science

May 15, 2024

AI is one of the most transformative and valuable scientific tools ever developed. By harnessing vast amounts of data and computational power, AI systems can un Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

ISC 2024 Keynote: High-precision Computing Will Be a Foundation for AI Models

May 15, 2024

Some scientific computing applications cannot sacrifice accuracy and will always require high-precision computing. Therefore, conventional high-performance c Read more…

Shutterstock 493860193

Linux Foundation Announces the Launch of the High-Performance Software Foundation

May 14, 2024

The Linux Foundation, the nonprofit organization enabling mass innovation through open source, is excited to announce the launch of the High-Performance Softw Read more…

ISC 2024: Hyperion Research Predicts HPC Market Rebound after Flat 2023

May 13, 2024

First, the top line: the overall HPC market was flat in 2023 at roughly $37 billion, bogged down by supply chain issues and slowed acceptance of some larger sys Read more…

Top 500: Aurora Breaks into Exascale, but Can’t Get to the Frontier of HPC

May 13, 2024

The 63rd installment of the TOP500 list is available today in coordination with the kickoff of ISC 2024 in Hamburg, Germany. Once again, the Frontier system at 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…

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…

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…

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…

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…

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…

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…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Leading Solution Providers

Contributors

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…

Eyes on the Quantum Prize – D-Wave Says its Time is Now

January 30, 2024

Early quantum computing pioneer D-Wave again asserted – that at least for D-Wave – the commercial quantum era has begun. Speaking at its first in-person Ana Read more…

The GenAI Datacenter Squeeze Is Here

February 1, 2024

The immediate effect of the GenAI GPU Squeeze was to reduce availability, either direct purchase or cloud access, increase cost, and push demand through the roof. A secondary issue has been developing over the last several years. Even though your organization secured several racks... Read more…

Intel Plans Falcon Shores 2 GPU Supercomputing Chip for 2026  

August 8, 2023

Intel is planning to onboard a new version of the Falcon Shores chip in 2026, which is code-named Falcon Shores 2. The new product was announced by CEO Pat Gel Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

GenAI Having Major Impact on Data Culture, Survey Says

February 21, 2024

While 2023 was the year of GenAI, the adoption rates for GenAI did not match expectations. Most organizations are continuing to invest in GenAI but are yet to Read more…

How the Chip Industry is Helping a Battery Company

May 8, 2024

Chip companies, once seen as engineering pure plays, are now at the center of geopolitical intrigue. Chip manufacturing firms, especially TSMC and Intel, have b 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…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire