Our research focus is on modeling and simulating next generation heterogeneous multicores and manycores to gain insight and propose new techniques to increase energy efficiency and performance.
We are performing research in the following areas:
- Architectural-independent modeling techniques
- Fast, scalable simulation
- Statistical models that capture (and explain) memory behavior
- Advanced software prefetching techniques based on statistical modeling
- Rethinking the on-chip memory system for power efficiency
- Novel on-chip memory system architectures
- Novel coherence
- Techniques to manage energy-efficiency at runtime
- New execution paradigms to enhance power-efficiency
- Novel task-based programming models to expose power as a first-class constrain to the programmer/compiler
- Nano architectures based on novel devices (nanowires, single-electron devices, non-boolean logic devices)
Research in computer architecture is always evolving in pace with the advancement in the state-of-the-art. New research topics are added while old are fading. Our research focus at present is concentrated to three areas we believe will be of great importance in the near future.