Covering Disruptive Technology Powering Business in The Digital Age

(NVIDIA)

Artificial intelligence (AI) and accelerated computing will help climate researchers achieve the miracles they need to achieve breakthroughs in climate research, NVIDIA Founder and CEO Jensen Huang said during a keynote at the Berlin Summit for the Earth Virtualization Engines initiative.

“Richard Feynman once said that ‘what I can’t create, I don’t understand’ and that’s the reason why climate modeling is so important,” Huang told 180 attendees at the Harnack House in Berlin, a storied gathering place for the region’s scientific and research community.

“And so the work that you do is vitally important to policymakers to researchers to the industry,” he added.

To advance this work, the Berlin Summit brings together participants from around the globe to harness AI and high-performance computing for climate prediction.

In his talk, Huang outlined three miracles that will have to happen for climate researchers to achieve their goals, and touched on NVIDIA’s own efforts to collaborate with climate researchers and policymakers with its Earth-2 efforts.

The first miracle required will be to simulate the climate fast enough, and with a high enough resolution—on the order of just a couple of square kilometres.

The second miracle needed will be to emulate the physics of the climate system, using AI, to explore all possible futures at lightning speed, with high fidelity. Generative AI breakthroughs promise new ways of predicting Earth’s climate, at kilometre-scale.

Simulation and AI will generate vast quantities of data

The third miracle is the ability to visualise all this data interactively with NVIDIA Omniverse to “put it in the hands of policymakers, businesses, companies and researchers.”

The Next Wave of Climate and Weather Innovation

The Earth Virtualisation Engines initiative, known as EVE, is an international collaboration that brings together digital infrastructure focused on climate science, HPC and AI aiming to provide, for the first time, easily accessible kilometre-scale climate information to sustainably manage the planet.

“The reason why Earth-2 and EVE found each other at the perfect time is because Earth-2 was based on these three fundamental breakthroughs,” Huang said.

The initiative promises to accelerate the pace of advances, advocating coordinated climate projections at 2.5-kilometre resolution. It’s an enormous challenge, but it’s one that builds on a huge base of advancements over the past 25 years.

A sprawling suite of applications already benefits from accelerated computing, including ICON, IFS, NEMO, MPAS, WRF-G and more—and much more computing power for such applications is coming.

Delivering Higher Performance

The NVIDIA GH200 Grace Hopper Superchip is a breakthrough accelerated CPU designed from the ground up for giant-scale AI and HPC applications. It delivers up to 10x higher performance for applications running terabytes of data.

It is built to scale, and by connecting large numbers of these chips together, NVIDIA can offer systems with the power efficiency to accelerate the work of researchers at the cutting edge of climate research. “To the software, it looks like one giant processor,” Huang said.

To help researchers put vast quantities of data to work, quickly, to unlock understanding, Huang spoke about NVIDIA Modulus, an open-source framework for building, training and fine-tuning physics-based machine learning models, and FourCastNet, a global, data-driven weather forecasting model, and how the latest AI-driven models can learn physics from real-world data.

Boosting Prediction Accuracy

Using raw data alone, FourCastNet is able to learn the principles governing complex weather patterns. Huang showed how FourCastNet was able to accurately predict the path of Hurricane Harvey by modeling the Coriolis force, the effect of the Earth’s rotation, on the storm.

AI models, when tethered to “checkpoints” created by traditional simulation, not only enable the generation and interactive exploration of massive ensembles of possible trajectories at high fidelity but also provide massive data compression. The longer the distance between checkpoints, the larger the compression achieved. FourCastNet today is capable of tethering between checkpoints spaced a month apart achieving 700x data compression. Huang then demonstrated how FourCastNet-generated ensembles, running on NVIDIA GPUs, anticipated an unprecedented North African heatwave.

By running FourCastNet in Modulus, NVIDIA was able to generate 21-day weather trajectories of 1,000 ensemble members in one-tenth the time it previously took to do a single trajectory—and with 1,000x less energy consumption.

Lastly, NVIDIA technologies promise to help all this knowledge become more accessible with digital twins able to create interactive models of increasingly complex systems — from Amazon warehouses to the way 5G signals propagate in dense urban environments.

Huang then showed a stunning, high-resolution interactive visualisation of global-scale climate data in the cloud, zooming in from a view of the globe to a detailed view of Berlin. This approach can work to predict climate and weather in locations as diverse as Berlin, Tokyo and Buenos Aires, Huang said.

Earth: The Final Frontier

To help meet challenges such as these, Huang outlined how NVIDIA is building more powerful systems for training AI models, simulating physical problems and interactive visualization.

“These new types of supercomputers are just coming online,” Huang said. “This is as fresh a computing technology as you can imagine.”

Huang ended his talk by thanking key researchers from across the field and playfully suggesting a mission statement for EVE.

“Earth, the final frontier, these are the voyages of EVE,” Huang said. Its “mission is to push the limits of computing in service of climate modeling, to seek out new methods and technologies to study the global-to-local state of the climate to inform today the impact of mitigation and adaptation to Earth’s tomorrow, to boldly go where no one has gone before.”

Learn more about Earth-2.

(0)(0)