CHESS Initiative Final Report
I'm happy to share our report covering the achievements of "Cloud, High-Performance Computing (HPC), and Edge for Science and Security" (CHESS), one of the last laboratory initiatives I participated in at Pacific Northwest National Laboratory.
CHESS focused on developing techniques for executing complex scientific workflows across a heterogeneous computing continuum spanning lab instruments, edge sensors, HPC systems, and cloud resources. This is particularly important for research that involves large datasets and complex computations, such as creating digital twins of physical systems or automating experimental facilities.
The team developed several key capabilities, including:
✅ A way for scientists to easily use a combination of cloud and HPC resources for their workflows.
✅ AI-infused services for workflow and data management.
✅ Predictive models to optimize the use of computing resources.
We demonstrated significant improvements in performance and efficiency, including:
📈 Up to 87x speedup in response time for specific tasks.
🗜️ 10-12x speedup in image compression.
🎯 Improved accuracy in image analysis.
These results show the potential of heterogeneous computing for accelerating scientific discovery. I'm proud to have been a part of this effort, which I believe will benefit not only materials and chemical sciences but also a wide range of scientific domains. I am especially grateful for an exceptional team, including Nathan Tallent, Jan Strube, Oceane Bel, Luanzheng Guo, Ph.D., and others.