TEC⁴
The Transferring Exascale Computational Chemistry to Cloud Computing Environment and Emerging Hardware Technologies (TEC4) project’s goal is to accelerate the transition from the basic research associated with developing and implementing electronic structure methods to their widespread use in solving complex challenges in industrial chemical sciences.
Image by Nathan Johnson | Pacific Northwest National Laboratory
The Transferring Exascale Computational Chemistry to Cloud Computing Environment and Emerging Hardware Technologies (TEC4) project is led by Pacific Northwest National Laboratory in partnership with Central Michigan University, Lawrence Berkeley National Laboratory, Louisiana State University, Micron, Microsoft, and the University of Texas at El Paso.
TEC4 is a deeply collaborative project that will integrate scalable computational chemistry software with cost- and energy-efficient computing resources. This will enable the widespread adoption of Department of Energy-developed electronic structure methods by industrial users, providing powerful solutions that bridge fundamental research and applied challenges. The team will develop a new computational infrastructure that transitions the electronic structure methods to the high-performance computing (HPC) cloud.
Electronic structure models developed and used by researchers for fundamental science can provide the high level of accuracy needed to address key challenges in chemistry and materials science. Extending these models to industrially relevant scales can accelerate discovery for a wide range of technologies, from clean energy materials to separations. The TEC4 approach will leverage Microsoft Azure HPC cloud computing and Micron’s advanced memory technologies to facilitate the development of powerful, accurate, and efficient computational chemistry tools that can be delivered to more researchers through computational chemistry as a service (CCaaS).
CCaaS brings together developments in cloud computing, machine learning, hardware, and chemical expertise as a solution for the shortage of computational facilities powerful enough to perform the large numbers of calculations needed to solve complex problems.
TEC4 is organized into three thrusts, which span HPC tools, machine learning, and CCaaS.
Thrust 2: Machine Learning Models
Thrust 3: CCaaS and Applications
TEC4 researchers will work to develop tools that bridge cloud computing systems and Micron technologies. These tools will enable the efficient use of new computational technologies across HPC and other platforms—a necessity for the machine learning models at the heart of CCaaS. The complexity of the different computational systems requires focused effort to find ways to effectively use the full range of computing infrastructures available for computational chemistry.
Machine learning tools will allow the TEC4 team to extend the applicability of popular electronic structure methods relevant to industrial applications. Combining deep neural network models and novel theoretical formulations will allow users to perform simulations and training on a HPC framework. Once fully trained, the models will be incorporated into workflows needed to develop CCaaS.
The final step of the project is creating and implementing CCaaS. The team will emphasize producing customized electronic structure workflows, innovative machine learning tools, and web interfaces for the workflows. CCaaS will help extend the reach of state-of-the-art computational chemistry tools to accelerate innovation across essential chemistry challenges.
TEC4 is part of the Department of Energy’s Accelerate initiative, which was designed to “integrate novel concepts and approaches into fundamental research that could accelerate the innovation process, greatly reducing the time from discovery to product.” The initiative encourages partnerships across national laboratories, academia, and industry.