Pacemaker software is revolutionzing the parametrization and deployment of atomic interaction models. Offering a sophisticated and efficient framework, Pacemaker streamlines the process of constructing ACE models, ensuring both accuracy and scalability in large-scale atomistic simulations. With automated features and GPU acceleration, Pacemaker empowers researchers to systematically converge Atomic Cluster Expansions while maintaining a physically meaningful description of atomic interactions. By providing superior or comparable performance to the best Machine Learning potentials available, Pacemaker stands as a cornerstone for advancing computational materials science, offering researchers a reliable solution for tackling diverse bond chemistries and rapid training of advanced atomic interaction models.

Further tools for efficient use of Pacemaker include:

  • reference DFT data collection
  • optimal selection of representative atomic configurations
  • on-the-fly uncertainty indication during molecular dynamics simulations
  • active learning

Want to know more? Contact us!