Atos and NVIDIA are partners in different fields such as supercomputers, edge solutions and health with the Clara platform. On November 15, they announced the creation of a new artificial intelligence laboratory, the Excellence AI Lab (EXAIL), where scientists and researchers will aim to accelerate the progress of computing technologies, education and research in Europe using high performance computing and AI. The initial work of the lab will focus on five research areas: climate research, health and genome analysis, hybridization with quantum computing, edge AI/computer vision and cybersecurity.
For this work, Atos will develop a BullSequana X supercomputer with NVIDIA's Arm Grace processor, Ampere Next next-generation GPU, Atos' Exascale Interconnect BXI solution and NVIDIA's Quantum-2 Infiniband networking platform, which enables predictive performance as if the application was running on a dedicated system.
Predicting and Combating Global Warming
Atos and Nvidia will be using the Atos JUWELLS supercomputer, which has benefited from NVIDIA's GPU architectures to help combat climate change. Commenting on this, Agnès Boudot, Senior Vice President, Director of HPC & Quantum activities at Atos said:"This year at SC'20 we are highlighting how Atos is enabling faster, smarter and greener simulation - and the JUWELS supercomputing system is a great example of this. It has been fantastic to work with Jülich and its partners to deliver this increased performance and power to drive unprecedented research and innovation in Europe."For his part, Ian Buck, general manager and vice president of accelerated computing at Nvidia said:
"Our A100 Tensor Core GPUs featured in this 'booster' module are based on NVIDIA's Ampere architecture, a revolutionary technical achievement that delivers a massive performance improvement over the previous generation for AI learning and inference. This datacenter-scale platform with GPU computing units and the Mellanox HDR InfiniBand network enables Jülich researchers to embark on the development and deployment of accelerated data analytics, HPC and AI to meet the challenges of the exascale AI era."As a result, Atos and NVIDIA researchers will run new AI and deep learning models on the Juwell supercomputer based at the Jülich Supercomputing Centre in Germany. These large-scale models can be used to predict the evolution of extreme weather events such as heat waves, hurricanes, etc. Atos and NVIDIA hope to gain a better understanding of the changing behaviour of these weather phenomena due to global warming through exascale computing. Andy Grant, vice president of global sales, HPC,AI and Quantomm at Atos said:
"Atos is firmly committed to its decarbonisation ambitions of offsetting all our residual emissions by 2028 to achieve "net zero emissions" and achieving the SBTi goal of reducing our global carbon emissions under our control and influence by 50% by 2025, Many leading climate modelling centres, such as Météo France, DKRZ, KNMI and AEMet, use our BullSequana supercomputers to run their large scale weather and climate models. The announcement of the EXAIL lab illustrates our continued commitment to this area, one year after the creation of our centre of excellence in weather and climate modelling with ECMWF."At Nvidia, Ian Buck, Vice President and General Manager, Accelerated Computing, commented:
"As climate change intensifies and increases the frequency of extreme weather events that disrupt entire regions, costing governments and economies hundreds of billions of euros each year, EXAIL's goal is to advance critical research to address the urgent global challenges of climate change."