Necessary increases in computing power with orders of magnitude higher energy efficiency do not only affect applications in the field of machine learning. The need for innovative electronic systems generally applies to all application areas of embedded systems. The “Post Silicon Technologies” research focus area therefore aims to research new technologies that are compatible with and alternative to silicon. These include, for example, non-volatile memory (NVM) technologies based on memristors, e.g. MRAM, FeFETs or ReRAMs, as well as innovative data processing techniques that avoid transporting data from memories at all, but instead filter and process it within or at least close to the memory itself (so-called edge and near-memory computing).
Some ESI members are currently carrying out very successful research work, e.g. as project leader in the two current DFG priority programs SPP 2037 (Scalable Data Management for Future Hardware) and SPP 2262 (Memristive Devices Toward Smart Technical Systems) and the new DFG Priority Program launched in September 2022 SPP 2377 (Disruptive Memory Technologies).
The availability of NVM technologies now requires not only the development of completely new processes for the simulation, synthesis and optimization of such systems with persistent memory, but also a revolutionary new approach at the level of operating systems, data structures and applications (especially database systems). The clever combination of the use of hybrid, i.e. non-volatile and volatile memory with different characteristics in terms of energy, access time and longevity also opens up a wealth of previously unexplored optimization questions and techniques.
