Advanced Computer Architecture Research

Under the aegis of the Novel Technology thrust of the CAL Lab, work has been initiated on several foci of interest of DOE with the intent of expanding the work plan for future renewals.  These are generally far-range goals of common interest and, although out-of-scope of the current project tasks, are of sufficient interest to warrant early exploration in keeping with the CAL mission of exploratory investigations of promising ECI directions.

We explicitly offer no deliverables for these areas, but rather lay the groundwork for future conceptual thrusts.   Broadly organized, the three categories below represent devices, circuits, and systems.  These divisions are for convenience, and often span several topics; quantum may include component development, circuit assembly, or full system design, for example.

An additional area is that of integrative or cross-cutting activities which can foster success by leveraging two or more current DOE research targets and collaboratively co-execute for mutual benefit.  For example, a key finding of the X-Stack effort may be ported to a CAL AMM-based hardware platform for experimental evaluation, thus providing beneficial feedback to both groups.



Future Computational Substrates

Intended to explore atypical approaches at the device level, both semiconducting materials and other physics such as quantum devices.  Also included are emerging memory devices (PCM, reRAM, etc.) in the physical domain.


Novel Circuit Approaches

Applying and leveraging CMOS technology but engaging computing principles at the circuit level.  Examples include spatial vs. temporal processing, specialized and FPGA algorithmic accelerators, revised memory hierarchies incorporating stacked silicon and NVRAM.  We are developing a 2.5D silicon substrate methodology to enable further explorations.




Non-von Neumann Architectures

Promising approaches with embrace non-von Neumann techniques such as neuromorphic computing.


Comments