Nvidia Earth-2 (E-2) Supercomputer

Main article: Supercomputers

2021: Plans for the most powerful supercomputer

On November 12, 2021, the company Nvidia announced plans to create the most powerful in the world supercomputer artificial intelligence with to predict climate change. The Earth-2 system, or E-2, will create in. digital double Lands Omniverse

This system will be an analogue of Cambridge-1, a powerful supercomputer with artificial intelligence for medical research. We introduced Cambridge-1 in the UK in the summer of 2021, and it is already used by a number of leading medical companies.

The earth is heating up. The last seven years will be the seven warmest in the history of observations. Greenhouse gas emissions from human activities have caused an average warming of about 1.1 ° C since 1850-1900.

What we are experiencing is very different from the global average. We are experiencing extreme weather conditions - historical droughts, unprecedented heat waves, severe hurricanes, severe storms and catastrophic floods. Climate disasters are the norm.

We need to fight climate change. Nevertheless, we will not feel the benefits of our efforts for decades. It is difficult to mobilize action for something in the future. But we must know our future, see it and feel it - to act without delay.

To make our future a reality, simulation is the answer.

To develop better mitigation and adaptation strategies, we need climate models that can predict climate in different regions of the globe for decades.

Unlike weather forecasting, which primarily models atmospheric physics, climate models are multi-decade simulations that model physics, chemistry and biology of the atmosphere, water, ice, earth and human activities.

Climate simulations are configured with a resolution of 10 to 100 kilometers.

But more resolution is needed to simulate changes in the global water cycle - the movement of water from the ocean, sea ice, land surface and groundwater through the atmosphere and clouds. Changes in this system lead to increased storms and droughts.

To simulate clouds reflecting sunlight back into space, a resolution on a meter scale is required. Scientists estimate that these resolutions will require millions to billions of times more processing power than what is currently available. It will take decades to achieve this through the usual course of computer achievements, which is accelerated 10 times every five years.

For the first time, we have technology for ultra-high-resolution climate modeling, for transitioning to light speed and predicting changes in extreme weather conditions in the region for decades to come.

We can achieve a million-fold increase in speed by combining three technologies: computing with acceleration on the GPU; deep learning and breakthroughs in neural networks based on physics; and supercomputerivAI, as well as a huge amount of observable and model data that you can learn from.

And with ultra-high resolution techniques, we might be able to make the one billion-fold leap needed to model the ultra-high resolution climate. Countries, cities and towns can receive early warnings to adapt and enhance the sustainability of infrastructure. And with more accurate forecasts, people and countries will act more quickly.

So we will dedicate ourselves and our significant resources to directing the scale and expertise of NVIDIA in computing science to join the global community of climate scientists.