Cerebras Systems CS-2 selected by TotalEnergies for multi-energy research
SUNNYVALE, Calif., March 2, 2022 – Cerebras Systems, the pioneer in high-performance artificial intelligence (AI) computing, today announced that TotalEnergies Research & Technology USA has selected the world’s fastest AI computer, the Cerebras CS-2 system, to accelerate its multi-energy research. This continues the rapid adoption of Cerebras by leading enterprises around the world and marks the first publicly announced deployment of the Cerebras CS-2 in the energy sector.
“TotalEnergies’ roadmap is clear: more energy, fewer emissions. To achieve this, we must combine our strengths with those that allow us to go faster, higher and… stronger,” joked Dr. Vincent Saubestre, CEO and President of TotalEnergies Research & Technology USA with a thinly veiled reference. to the Olympic motto. Citius, Altius, Fortius. “Cerebras Systems offers one of the most powerful AI accelerators. We rely on the CS-2 system to energize our multi-energy research and give our research “athletes” that extra competitive edge.
Thanks to the CS-2’s state-of-the-art AI computation, advanced modeling and analysis will enable fast and accurate simulations on a wide range of problems addressed by TotalEnergies: from batteries to biofuels, wind flows, drillings and CO2 storage.
“We are thrilled to partner with TotalEnergies and bring our industry-leading AI performance to the multi-energy market,” said Andrew Feldman, CEO and co-founder of Cerebras Systems. “The energy industry has long been at the forefront of using computation to generate insights. AI and the integration of AI with simulation can accelerate TotalEnergies’ mission to provide affordable, cleaner and more reliable energy access. We are proud to be part of this important undertaking.
Predictive modeling requires massive computing resources and high bandwidth data communication. Using traditional general-purpose hardware for this job typically requires large clusters of GPUs or CPUs and frequent data movement between individual processors. Limited chip-to-chip bandwidth causes a communications bottleneck, which slows down the modeling workload and delays the time to get insights.
This challenge can be met with a single Cerebras CS-2 system, the fastest AI computer available. A single CS-2 not only offers cluster-scale computing power, but also orders of magnitude greater communication and memory bandwidth than traditional clusters. This translates to extraordinary performance on workloads such as predictive modeling that are critical to efficient energy development and production.
In recent work with TotalEnergies, Cerebras demonstrated over 100x improvement on a finite difference benchmark for seismic modeling compared to traditional architectures. Total and Cerebras engineers wrote the reference code using the new Cerebras software language (CSL). The CSL is part of the Cerebras SDK, which allows developers to take advantage of the strengths of the CS-2 system.
With customers and partners in North America, Asia, Europe and the Middle East, Cerebras provides cutting-edge AI solutions to a growing list of customers in the enterprise, military and high-performance computing, including Argonne National Laboratory, Lawrence Livermore National Laboratory, Pittsburgh Supercomputing Center, EPCC, Tokyo Electron Devices, and GlaxoSmithKline.
For more information on the Cerebras CS-2 system and its application in energy, please visit https://cerebras.net/industries/energy.
About Cerebras Systems
Cerebras Systems is a team of pioneering computer architects, computer scientists, deep learning researchers, and engineers of all types. We have come together to build a new class of computing systems, designed with the sole purpose of accelerating AI and changing the future of AI work forever. Our flagship product, the CS-2 system, is powered by the world’s largest processor – the 850,000-core Cerebras WSE-2, enabling customers to accelerate their deep learning work by orders of magnitude over general purpose computing.
Source: Cerebras Systems