Lawrence Livermore National Laboratory added SambaNova Systems’ spatial data flow accelerator to the lab’s Livermore Computing Center as part of an effort to see if artificial intelligence can aid the nuclear-weapons program.
Livermore will integrate the new hardware to further investigate cognitive simulation approaches combining artificial intelligence (AI) with high-performance computing to learn if deep neural network hardware architectures can accelerate traditional physics-based simulations, according to a statement from Livermore published on Tuesday.
The lab is expected to use the SambaNova AI systems to improve the fidelity of models and manage the growing volumes of data to improve overall speed, performance and productivity for stockpile stewardship applications, fusion energy research and other basic science work, Livermore wrote in the statement.
Understanding the cradle-to-grave requirements of modern nuclear weapons will require significantly more powerful computing capacity than the U.S. nuclear enterprise uses or has plans to develop in the future, according to a recent National Academy of Sciences report. Massive computing power is needed to perform modeling and simulation on modern nuclear weapons, which are no longer detonated to study their effects since the U.S. voluntarily ceased underground nuclear explosive testing in 1992.
“Multi-physics simulation is complex,” Brian Van Essen, Livermore’s informatics group leader, said in Tuesday’s statement. “Our inertial confinement fusion experiments generate huge volumes of data. Yet, connecting the underlying physics to the experimental data is an extremely difficult scientific challenge. AI techniques hold the key to teaching existing models to better mirror experimental models and to create an improved feedback loop between the experiments and models. The SambaNova system will help us create these cognitive simulations.”
In 2020, Livermore integrated SambaNova Systems’ DataScale AI accelerator into the National Nuclear Security Administration’s Corona supercomputing cluster, an 11-plus petaFLOP machine, allowing scientists to use AI calculations to more quickly and efficiently conduct research on stockpile stewardship and inertial confinement fusion applications, as well as for COVID-19 related research. The system was funded by the Advanced Simulation and Computing program and is part of an agreement between the Department of Energy and SambaNova Systems to accelerate AI within the DOE national laboratories.
This next stage of collaboration between the lab and SambaNova will allow the system to be more loosely coupled with the Lab’s supercomputing clusters, supporting a wider selection of workloads, Livermore said.
