Making future extreme scale systems a reality requires not only significant efforts in architectures, programming models, and applications, but also substantial advances towards a highly scalable middleware infrastructure. The latter includes program development and code optimization tools and methodologies, support for fault tolerance as well as runtime mechanisms for power-efficient execution. We are a team of researchers in the Center for Applied Scientific Computing at Lawrence Livermore National Laboratory, working closely with researchers in the Development Environment Group at Livermore Computing, that focuses on the challenges in designing, implementing and maintaining elements of this software stack for both current high-end platforms as well as future extreme scale machines.
- Program Development and Performance Analysis/Modeling
- Power Constrained Performance Optimization
- Scalable Resilience
- Interconnect Studies and Communication Optimizations
- Co-Design Tools and Methodologies
Links to our software can be found at: github.com/llnl