image of high performance computer at Lawrence Livermore National Laboratory

LLNL Computational Grand Challenge Award for machine-learning

image of high performance computer at Lawrence Livermore National Laboratory
Image of high performance computer at Lawrence Livermore National Laboratory

CMI researchers at Lawrence Livermore National Laboratory conducted the activity for this highlight

Achievement
CMI researchers have been awarded a significant amount of HPC time at LLNL

Significance and impact

  • Artificial Intelligence (AI) and Machine Learning (ML) models are becoming powerful tools for accelerating a broad range of materials research.
  • Training AI/ML models often requires large amounts of computational resources to develop large data sets of simulation results and to train the neural networks. 
  • An application for a large amount of HPC resources, including significant GPU resources was recently awarded to the Machine Learning for Materials Design project within CMI to explore the use of Graph Neural Networks. 

Details and next steps

  • The LLNL Computational Grand Challenge program releases an annual call for HPC resources
  • Current award period 1/1/2022 – 12/31/2022
  • An allocation was received on LLNL’s Lassen supercomputer (30,096 CPU cores + 3,152 GPUs)
  • Represents more than 50M core hours on Lassen, roughly equivalent to $350k